title
Artificial Intelligence Tutorial | Artificial Intelligence Full Course | AI Tutorial | Simplilearn
description
🔥 Purdue Post Graduate Program In AI And Machine Learning: https://www.simplilearn.com/pgp-ai-machine-learning-certification-training-course?utm_campaign=AIFCSixHours-8Pyy2d3SZuM&utm_medium=DescriptionFirstFold&utm_source=youtube
🔥Professional Certificate Course In AI And Machine Learning by IIT Kanpur (India Only): https://www.simplilearn.com/iitk-professional-certificate-course-ai-machine-learning?utm_campaign=23AugustTubebuddyExpPCPAIandML&utm_medium=DescriptionFF&utm_source=youtube
🔥AI Engineer Masters Program (Discount Code - YTBE15): https://www.simplilearn.com/masters-in-artificial-intelligence?utm_campaign=SCE-AIMasters&utm_medium=DescriptionFF&utm_source=youtube
🔥AI & Machine Learning Bootcamp(US Only): https://www.simplilearn.com/ai-machine-learning-bootcamp?utm_campaign=AIFCSixHours-8Pyy2d3SZuM&utm_medium=DescriptionFirstFold&utm_source=youtube
This video on the Artificial Intelligence tutorial will make you learn in detail about the different concepts involved in AI. You will understand the basics of AI and get an idea about Machine Learning and Deep Learning with hands-on demo in this Artificial Intelligence full course. You will look at how to become an AI Engineer and see some vital machine learning and deep learning interview questions. Now, let's dive in and learn artificial intelligence in detail.
Below topics are explained in this Artificial Intelligence tutorial:
1. Introduction to Artificial Intelligence (0:33)
2. What is Artificial Intelligence (06:20)
3. Brief history of Artificial Intelligence (07:16)
4. Types of Artificial Intelligence (10:25)
5. Artificial of Artificial Intelligence (13:23)
6. Future of Artificial Intelligence (14:56)
7. Machine Learning vs Deep Learning vs Artificial Intelligence (16:15)
8. Human vs Artificial Intelligence (19:36)
9. What is Machine Learning and Deep Learning? (21:33)
10. Real-life examples (31:31)
11. Types of Artificial Intelligence (33:49)
12. Machine Learning tutorial (42:38)
13. Why Machine Learning (43:12)
14. What is Machine Learning (47:19)
15. Types of Machine Learning (54:01)
16. Supervised Learning (54:13)
17. Reinforcement Learning (56:35)
18. Supervised vs Unsupervised (57:54)
19. Machine Learning Algorithms (59:12)
20. Linear regression (1:01:00)
21. Decision trees (1:08:12)
22. Support Vector Machine (1:16:31)
23. Clustering (1:44:56)
24. K-means clustering (1:45:45)
25. Logistic Regression (2:15:19)
26. Applications of Machine Learning (2:39:40)
27. What is Deep Learning? (2:44:36)
28. What is a Neural Network? (2:46:46)
29. Machine Learning Interview Questions & Answers (0:20:50)
To learn more about Artificial Intelligence, subscribe to our YouTube channel: https://www.youtube.com/user/Simplilearn?sub_confirmation=1
Download the Artificial Intelligence Career Guide and take a sneak peek into the world that awaits you: https://www.simplilearn.com/artificial-intelligence-career-guide-pdf?utm_campaign=Aritificial-Intelligence-Tutorial-8Pyy2d3SZuM&utm_medium=Tutorials&utm_source=youtube
Watch more videos on Artificial Intelligence: https://www.youtube.com/watch?v=ad79nYk2keg&list=PLEiEAq2VkUULyr_ftxpHB6DumOq1Zz2hq
#ArtificialIntelligence #AI #ArtificialIntelligenceTutorial #ArtificialIntelligenceFullCourse #Simplilearn
➡️ About Post Graduate Program In AI And Machine Learning
This AI ML course is designed to enhance your career in AI and ML by demystifying concepts like machine learning, deep learning, NLP, computer vision, reinforcement learning, and more. You'll also have access to 4 live sessions, led by industry experts, covering the latest advancements in AI such as generative modeling, ChatGPT, OpenAI, and chatbots.
âś… Key Features
- Post Graduate Program certificate and Alumni Association membership
- Exclusive hackathons and Ask me Anything sessions by IBM
- 3 Capstones and 25+ Projects with industry data sets from Twitter, Uber, Mercedes Benz, and many more
- Master Classes delivered by Purdue faculty and IBM experts
- Simplilearn's JobAssist helps you get noticed by top hiring companies
- Gain access to 4 live online sessions on latest AI trends such as ChatGPT, generative AI, explainable AI, and more
- Learn about the applications of ChatGPT, OpenAI, Dall-E, Midjourney & other prominent tools
âś… Skills Covered
- ChatGPT
- Generative AI
- Explainable AI
- Generative Modeling
- Statistics
- Python
- Supervised Learning
- Unsupervised Learning
- NLP
- Neural Networks
- Computer Vision
- And Many More…
👉 Learn More At: 🔥 Purdue Post Graduate Program In AI And Machine Learning: https://www.simplilearn.com/pgp-ai-machine-learning-certification-training-course?utm_campaign=AIFCSixHours-8Pyy2d3SZuM&utm_medium=Description&utm_source=youtube
🔥AI & Machine Learning Bootcamp(US Only): https://www.simplilearn.com/ai-machine-learning-bootcamp?utm_campaign=AIFCSixHours-8Pyy2d3SZuM&utm_medium=Description&utm_source=youtube
🔥🔥 Interested in Attending Live Classes? Call Us: IN - 18002127688 / US - +18445327688
detail
{'title': 'Artificial Intelligence Tutorial | Artificial Intelligence Full Course | AI Tutorial | Simplilearn', 'heatmap': [{'end': 1470.275, 'start': 628.758, 'weight': 1}, {'end': 2522.67, 'start': 1679.918, 'weight': 0.705}, {'end': 3572.79, 'start': 3358.432, 'weight': 0.789}, {'end': 3991.897, 'start': 3777.76, 'weight': 0.734}, {'end': 4831.013, 'start': 4615.743, 'weight': 0.713}, {'end': 20992.621, 'start': 20787.894, 'weight': 0.844}], 'summary': 'Tutorial provides an in-depth understanding of artificial intelligence (ai), machine learning (ml), and deep learning, covering their applications, impact on society, practical examples, and future prospects across various industries, with a focus on tensorflow, neural network fundamentals, and reinforcement learning, and emphasizing the importance of ai skills and concepts for career roles and salary insights.', 'chapters': [{'end': 564.371, 'segs': [{'end': 250.113, 'src': 'embed', 'start': 227.625, 'weight': 0, 'content': [{'end': 236.152, 'text': 'You must be wondering, well, how is artificial intelligence different from machine learning and deep learning? We saw what AI is.', 'start': 227.625, 'duration': 8.527}, {'end': 243.663, 'text': 'Machine learning is a technique to achieve AI, and deep learning, in turn, is a subset of machine learning.', 'start': 236.533, 'duration': 7.13}, {'end': 250.113, 'text': 'Machine learning provides a machine with the capability to learn from data and experience through algorithms.', 'start': 244.445, 'duration': 5.668}], 'summary': 'Ai is achieved through machine learning, with deep learning as a subset. machine learning enables machines to learn from data and experience.', 'duration': 22.488, 'max_score': 227.625, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM227625.jpg'}], 'start': 4.662, 'title': "Ai's impact and fundamentals", 'summary': "Delves into ai's pervasive influence in modern society, covering its applications in smartphones, cars, social media, banking, and more, as well as the core functionalities of ai demonstrated through a robot's adaptable performance. it also discusses ai's capabilities, classifications, the relationship between ai, machine learning, and deep learning, and a brief history of ai, highlighting key milestones and technological advancements.", 'chapters': [{'end': 95.036, 'start': 4.662, 'title': 'Artificial intelligence in daily life', 'summary': "Discusses the pervasive influence of artificial intelligence in modern society, covering its definition, applications in various areas such as smartphones, cars, social media, banking, and more, and the core functionalities of ai demonstrated through a robot's adaptable performance in diverse environments.", 'duration': 90.374, 'highlights': ['AI applications in smartphones, cars, social media feeds, video games, banking, and surveillance are mentioned, showcasing its widespread integration in various aspects of daily life.', 'The definition of AI as the intelligence exhibited by machines is explained, highlighting its dominance in human-computer interaction.', "The core functionality of AI is demonstrated through a robot's ability to adapt and perform as expected in varying environmental conditions, emphasizing its practical applications.", 'The use of complex algorithms and mathematical functions to build human-like intelligence in machines is mentioned, providing insight into the technical aspects of AI development.']}, {'end': 564.371, 'start': 95.436, 'title': 'Understanding artificial intelligence', 'summary': 'Discusses the capabilities of ai, its classification into weak ai and strong ai, the relationship between ai, machine learning, and deep learning, and a brief history of ai, highlighting key milestones and technological advancements.', 'duration': 468.935, 'highlights': ['AI provides machines with the capability to adapt, reason, and provide solutions.', 'The robot uses the given input and finds the solution for a problem, demonstrating problem-solving capabilities.', 'The chapter explains the classification of AI into Weak AI and Strong AI, along with their respective characteristics and examples.', 'The relationship between AI, machine learning, and deep learning is detailed, illustrating the technique to achieve AI and the subset of machine learning.', 'The chapter presents a brief history of AI, highlighting key milestones and advancements from 1956 to 2018.']}], 'duration': 559.709, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM4662.jpg', 'highlights': ['AI applications in smartphones, cars, social media, banking, and surveillance are widespread.', 'The definition of AI as the intelligence exhibited by machines is explained.', "The core functionality of AI is demonstrated through a robot's adaptable performance.", 'The use of complex algorithms and mathematical functions to build human-like intelligence in machines is mentioned.', 'AI provides machines with the capability to adapt, reason, and provide solutions.', 'The robot uses the given input and finds the solution for a problem, demonstrating problem-solving capabilities.', 'The chapter explains the classification of AI into Weak AI and Strong AI, along with their respective characteristics and examples.', 'The relationship between AI, machine learning, and deep learning is detailed.', 'The chapter presents a brief history of AI, highlighting key milestones and advancements from 1956 to 2018.']}, {'end': 1459.234, 'segs': [{'end': 589.846, 'src': 'embed', 'start': 564.471, 'weight': 3, 'content': [{'end': 572.918, 'text': 'We have our RPA, we have dancing robots, smart homes, and many more things are just coming out almost monthly in the world of artificial intelligence.', 'start': 564.471, 'duration': 8.447}, {'end': 582.863, 'text': 'So what is artificial intelligence? AI is a form of computer science used to create intelligent machines that can recognize human speech.', 'start': 573.259, 'duration': 9.604}, {'end': 587.245, 'text': 'Objects can learn, plan, and solve problems like humans.', 'start': 583.184, 'duration': 4.061}, {'end': 589.846, 'text': "And I'd like you to focus just on that last one.", 'start': 587.645, 'duration': 2.201}], 'summary': 'Ai is creating intelligent machines that can learn, plan, and solve problems like humans.', 'duration': 25.375, 'max_score': 564.471, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM564471.jpg'}, {'end': 859.472, 'src': 'embed', 'start': 832.427, 'weight': 0, 'content': [{'end': 837.331, 'text': "Siri, Cortana, Alexa, and Google now use voice recognition to follow the user's commands.", 'start': 832.427, 'duration': 4.904}, {'end': 842.616, 'text': 'These are all wonderful examples of current AIs that are in the commercial business world,', 'start': 837.731, 'duration': 4.885}, {'end': 846.239, 'text': 'and these ones in particular have matured over the last half a decade.', 'start': 842.616, 'duration': 3.623}, {'end': 847.781, 'text': 'For instance,', 'start': 846.7, 'duration': 1.081}, {'end': 856.509, 'text': "very few large banks in today's world would not use banking for fraud detection or for deciding whether it's a good person to give a loan to or not,", 'start': 847.781, 'duration': 8.728}, {'end': 859.472, 'text': "based on their credit scores and where they're from and their income.", 'start': 856.509, 'duration': 2.963}], 'summary': 'Voice recognition ais like siri, cortana, alexa, and google have matured over the last half decade and are widely used in commercial business, including by very few large banks for fraud detection and loan decisions.', 'duration': 27.045, 'max_score': 832.427, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM832427.jpg'}, {'end': 1392.236, 'src': 'embed', 'start': 1363.61, 'weight': 5, 'content': [{'end': 1366.392, 'text': 'Someone sits there and figures out what a tire looks like.', 'start': 1363.61, 'duration': 2.782}, {'end': 1367.772, 'text': 'It takes a lot of work.', 'start': 1366.412, 'duration': 1.36}, {'end': 1372.495, 'text': 'If you try to figure the difference between a car tire, a bicycle tire, a motorcycle tire.', 'start': 1367.813, 'duration': 4.682}, {'end': 1382.762, 'text': "So in the machine learning field, this could take a long time if you're going to do each individual aspect of a car and try to get a result on there.", 'start': 1373.536, 'duration': 9.226}, {'end': 1384.123, 'text': "And that's what they did do.", 'start': 1383.182, 'duration': 0.941}, {'end': 1387.505, 'text': 'This is still used on smaller amounts of data.', 'start': 1384.143, 'duration': 3.362}, {'end': 1390.674, 'text': 'where you figure out what those features are and then you label them.', 'start': 1388.012, 'duration': 2.662}, {'end': 1392.236, 'text': 'Deep learning.', 'start': 1391.475, 'duration': 0.761}], 'summary': 'Machine learning used to identify tire types, reducing time and effort.', 'duration': 28.626, 'max_score': 1363.61, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM1363610.jpg'}, {'end': 1449.772, 'src': 'embed', 'start': 1422.703, 'weight': 1, 'content': [{'end': 1429.39, 'text': "And then it changes all those weights going backward, they call it back propagation, and let it know it's a bicycle, and that's how it learns.", 'start': 1422.703, 'duration': 6.687}, {'end': 1432.994, 'text': "Once you've trained the neural network.", 'start': 1430.491, 'duration': 2.503}, {'end': 1436.157, 'text': 'you then put the new data in and they call this testing the model.', 'start': 1432.994, 'duration': 3.163}, {'end': 1437.218, 'text': 'so you need to have some data.', 'start': 1436.157, 'duration': 1.061}, {'end': 1439.22, 'text': "you've kept off to the side where you know the answer to.", 'start': 1437.218, 'duration': 2.002}, {'end': 1444.79, 'text': 'And you take that and you provide the required output and you say okay, is this neural network working correctly??', 'start': 1439.929, 'duration': 4.861}, {'end': 1449.772, 'text': 'Did it identify a bike as a bike, a truck as a truck, a motorcycle as a motorcycle?', 'start': 1444.83, 'duration': 4.942}], 'summary': 'Neural network training involves back propagation and testing with new data to ensure correct identification of objects.', 'duration': 27.069, 'max_score': 1422.703, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM1422703.jpg'}], 'start': 564.471, 'title': "Ai's applications and future prospects", 'summary': "Covers the concepts of ai, including its types, applications, and future prospects such as commercial uses in banking, online customer support, cybersecurity, and virtual assistants, as well as the potential impact on society, including automated transportation, smart cities, and robotic assistance for the elderly. it also discusses practical applications of ai, machine learning, and deep learning in everyday life, highlighting examples such as amazon echo's voice commands processing and google's personalized search results.", 'chapters': [{'end': 1002.645, 'start': 564.471, 'title': 'Understanding artificial intelligence and its future', 'summary': 'Explores the concepts of artificial intelligence, including types, applications, and future prospects, such as commercial uses in banking, online customer support, cybersecurity, and virtual assistants, as well as the potential impact on society, such as automated transportation, smart cities, and robotic assistance for the elderly.', 'duration': 438.174, 'highlights': ['The chapter explores the concepts of artificial intelligence, including types, applications, and future prospects, such as commercial uses in banking, online customer support, cybersecurity, and virtual assistants.', 'The potential impact of AI on society, such as automated transportation, smart cities, and robotic assistance for the elderly, is discussed.', 'The use of AI in banking for fraud detection, online customer support, cybersecurity, and virtual assistants is detailed, with examples of how AI has matured and is widely used in these areas.']}, {'end': 1459.234, 'start': 1003.446, 'title': 'Ai in everyday life', 'summary': 'Discusses the practical applications of artificial intelligence, machine learning, and deep learning, highlighting how amazon echo processes voice commands, google uses machine learning to personalize search results, and deep learning is used to identify and color objects in images.', 'duration': 455.788, 'highlights': ['Amazon Echo uses AI to process voice commands and provide information, exemplifying practical AI applications.', 'Google utilizes machine learning to personalize search results based on user behavior, improving user experience.', 'Deep learning is employed to identify and color objects in images, demonstrating the practical application of AI in image recognition and processing.']}], 'duration': 894.763, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM564471.jpg', 'highlights': ['The use of AI in banking for fraud detection, online customer support, cybersecurity, and virtual assistants is detailed, with examples of how AI has matured and is widely used in these areas.', 'The potential impact of AI on society, such as automated transportation, smart cities, and robotic assistance for the elderly, is discussed.', 'Google utilizes machine learning to personalize search results based on user behavior, improving user experience.', 'Amazon Echo uses AI to process voice commands and provide information, exemplifying practical AI applications.', 'The chapter explores the concepts of artificial intelligence, including types, applications, and future prospects, such as commercial uses in banking, online customer support, cybersecurity, and virtual assistants.', 'Deep learning is employed to identify and color objects in images, demonstrating the practical application of AI in image recognition and processing.']}, {'end': 2881.094, 'segs': [{'end': 2522.67, 'src': 'heatmap', 'start': 1679.918, 'weight': 0.705, 'content': [{'end': 1681.86, 'text': 'They kind of put it on a bullet on the side here.', 'start': 1679.918, 'duration': 1.942}, {'end': 1685.062, 'text': "It's a subset of machine learning.", 'start': 1682.36, 'duration': 2.702}, {'end': 1686.063, 'text': 'This is important.', 'start': 1685.262, 'duration': 0.801}, {'end': 1690.054, 'text': 'When we talk about deep learning, it is a form of machine learning.', 'start': 1686.693, 'duration': 3.361}, {'end': 1693.555, 'text': "There's lots of other forms of machine learning data analysis,", 'start': 1690.154, 'duration': 3.401}, {'end': 1701.157, 'text': 'but this is the newest and biggest thing that they apply to a lot of different packages and they use all the other machine learning tools available to work with it.', 'start': 1693.555, 'duration': 7.602}, {'end': 1703.018, 'text': "And it's very fast to test.", 'start': 1701.777, 'duration': 1.241}, {'end': 1711.12, 'text': 'You put in your information, you then have your group of tests, and then you held some aside, and you see how does it do.', 'start': 1703.698, 'duration': 7.422}, {'end': 1714.681, 'text': "It's very quick to test it and see what's going on with your deep learning and your neural network.", 'start': 1711.16, 'duration': 3.521}, {'end': 1721.358, 'text': 'Are they really all that different? AI versus machine learning versus deep learning.', 'start': 1715.647, 'duration': 5.711}, {'end': 1723.279, 'text': 'Concepts of AI.', 'start': 1721.978, 'duration': 1.301}, {'end': 1725.82, 'text': 'So we have concepts of AI.', 'start': 1724.099, 'duration': 1.721}, {'end': 1731.542, 'text': "You'll see natural language processing, machine learning, an approach to create artificial intelligence.", 'start': 1726.1, 'duration': 5.442}, {'end': 1734.303, 'text': "So it's one of the subsets of artificial intelligence.", 'start': 1732.062, 'duration': 2.241}, {'end': 1739.785, 'text': 'Knowledge representation, automated reasoning, computer vision, robotics.', 'start': 1735.043, 'duration': 4.742}, {'end': 1745.887, 'text': 'Machine learning versus AI versus deep learning or AI and machine learning and deep learning.', 'start': 1740.585, 'duration': 5.302}, {'end': 1751.168, 'text': 'So when we look at this, we have AI with machine learning and deep learning.', 'start': 1747.304, 'duration': 3.864}, {'end': 1753.11, 'text': "And so we're going to put them all together.", 'start': 1751.908, 'duration': 1.202}, {'end': 1755.071, 'text': 'We find out that AI is the big picture.', 'start': 1753.17, 'duration': 1.901}, {'end': 1756.493, 'text': 'We have a collection of books.', 'start': 1755.372, 'duration': 1.121}, {'end': 1758.555, 'text': 'It goes through some deep learning.', 'start': 1757.093, 'duration': 1.462}, {'end': 1760.897, 'text': 'The digital data is analyzed.', 'start': 1759.235, 'duration': 1.662}, {'end': 1762.939, 'text': 'Text mining comes through.', 'start': 1761.057, 'duration': 1.882}, {'end': 1768.464, 'text': "The particular book you're looking for, maybe it's a genre of books, is identified.", 'start': 1762.959, 'duration': 5.505}, {'end': 1772.588, 'text': 'And in this case, we have a robot that goes and gives a book to the patron.', 'start': 1769.085, 'duration': 3.503}, {'end': 1777.37, 'text': 'I have yet to be at a library that has a robot bring me a book, but that will be cool when it happens.', 'start': 1773.106, 'duration': 4.264}, {'end': 1779.652, 'text': "So we'll look at some of the pieces here.", 'start': 1778.291, 'duration': 1.361}, {'end': 1788.921, 'text': 'This information goes into, as far as this example, the translation of the handwritten printed data to digital form.', 'start': 1779.672, 'duration': 9.249}, {'end': 1791.102, 'text': "That's pretty hard to do.", 'start': 1789.941, 'duration': 1.161}, {'end': 1799.849, 'text': "That's pretty hard to go in there and translate hundreds and hundreds of books and understand what they're trying to say if you've never read them.", 'start': 1791.122, 'duration': 8.727}, {'end': 1806.634, 'text': "So in this case, we used a deep learning because you can already use examples where they've already classified a lot of books.", 'start': 1800.449, 'duration': 6.185}, {'end': 1811.837, 'text': 'And then they can compare those texts and say, oh, okay, this is a book on automotive repair.', 'start': 1806.974, 'duration': 4.863}, {'end': 1814.519, 'text': 'This is a book on robotic building.', 'start': 1812.017, 'duration': 2.502}, {'end': 1816.66, 'text': 'The digital data is analyzed.', 'start': 1814.719, 'duration': 1.941}, {'end': 1820.022, 'text': 'Then we have more text mining using machine learning.', 'start': 1817.12, 'duration': 2.902}, {'end': 1829.408, 'text': "So maybe we'd use a different program to do a basic classify what you're looking for and say, oh, you're looking for auto repair and computers.", 'start': 1820.503, 'duration': 8.905}, {'end': 1830.849, 'text': "So you're looking for automated cars.", 'start': 1829.468, 'duration': 1.381}, {'end': 1833.491, 'text': "Once it's identified, then, of course, it brings you the book.", 'start': 1831.369, 'duration': 2.122}, {'end': 1840.506, 'text': "So here's a nice summation of what we were just talking about, AI with machine learning and deep learning.", 'start': 1835.342, 'duration': 5.164}, {'end': 1845.649, 'text': 'Deep learning is a subset of machine learning, which is a subset of artificial intelligence.', 'start': 1840.826, 'duration': 4.823}, {'end': 1848.992, 'text': 'So you can look at artificial intelligence as the big picture.', 'start': 1846.49, 'duration': 2.502}, {'end': 1852.675, 'text': 'How does this compare to the human experience?', 'start': 1849.252, 'duration': 3.423}, {'end': 1858.659, 'text': 'in either doing the same thing as a human we do, or does it better than us?', 'start': 1852.675, 'duration': 5.984}, {'end': 1866.302, 'text': 'And machine learning, which has a lot of tools, is something that learns from data, past experiences.', 'start': 1859.338, 'duration': 6.964}, {'end': 1867.262, 'text': "It's programmed.", 'start': 1866.402, 'duration': 0.86}, {'end': 1871.785, 'text': 'It comes in there and it says, hey, we already had these five things happen.', 'start': 1867.483, 'duration': 4.302}, {'end': 1873.666, 'text': 'The sixth one should be about the same.', 'start': 1872.085, 'duration': 1.581}, {'end': 1879.789, 'text': "And then there's a lot of tools in machine learning, but deep learning then is a very specific tool in machine learning.", 'start': 1874.226, 'duration': 5.563}, {'end': 1888.061, 'text': "It's the artificial neural network which handles large amounts of data and is able to take huge pools of of experiences,", 'start': 1880.25, 'duration': 7.811}, {'end': 1890.344, 'text': 'pictures and ideas and bring them together.', 'start': 1888.061, 'duration': 2.283}, {'end': 1892.746, 'text': 'Real life examples.', 'start': 1891.365, 'duration': 1.381}, {'end': 1894.989, 'text': 'Artificial intelligence.', 'start': 1893.868, 'duration': 1.121}, {'end': 1896.27, 'text': 'News generation.', 'start': 1895.449, 'duration': 0.821}, {'end': 1903.719, 'text': 'Very common nowadays as it goes through there and finds the news articles or generates the news based upon the news feeds.', 'start': 1896.431, 'duration': 7.288}, {'end': 1909.412, 'text': "or the back end coming in and says, okay, let's give you the actual news based on this.", 'start': 1904.551, 'duration': 4.861}, {'end': 1914.914, 'text': "There's all the different things, Amazon Echo, they have a number of different Prime Music on there.", 'start': 1909.932, 'duration': 4.982}, {'end': 1920.155, 'text': "Of course, there's also the Google Command, and there's also Cortana.", 'start': 1914.934, 'duration': 5.221}, {'end': 1927.377, 'text': "There's tons of smart home devices now where we can ask it to turn the TV on or play music for us.", 'start': 1920.195, 'duration': 7.182}, {'end': 1929.248, 'text': "That's all artificial intelligence.", 'start': 1927.907, 'duration': 1.341}, {'end': 1939.037, 'text': "From front to back, you're having a human experience with these computers and these objects that are connected to the processing.", 'start': 1929.328, 'duration': 9.709}, {'end': 1940.098, 'text': 'Machine learning.', 'start': 1939.418, 'duration': 0.68}, {'end': 1941.94, 'text': 'Spam detection.', 'start': 1941.019, 'duration': 0.921}, {'end': 1942.601, 'text': 'Very common.', 'start': 1942.04, 'duration': 0.561}, {'end': 1946.785, 'text': "Machine learning doesn't really have the human interaction part.", 'start': 1942.641, 'duration': 4.144}, {'end': 1951.769, 'text': "So this is the part where it goes and says, okay, that's a spam, that's not a spam.", 'start': 1947.725, 'duration': 4.044}, {'end': 1953.391, 'text': 'And it puts it in your spam folder.', 'start': 1951.989, 'duration': 1.402}, {'end': 1956.272, 'text': 'Search engine result refining.', 'start': 1954.651, 'duration': 1.621}, {'end': 1965.054, 'text': 'Another example of machine learning, whereas it looks at your different results and it is able to categorize them.', 'start': 1956.972, 'duration': 8.082}, {'end': 1968.435, 'text': 'as far as this had the most hits, this is the least viewed.', 'start': 1965.054, 'duration': 3.381}, {'end': 1970.416, 'text': 'this has five stars, you know.', 'start': 1968.435, 'duration': 1.981}, {'end': 1971.376, 'text': 'however, they want to weight it.', 'start': 1970.416, 'duration': 0.96}, {'end': 1973.977, 'text': 'All good examples of machine learning.', 'start': 1972.116, 'duration': 1.861}, {'end': 1975.598, 'text': 'And then the deep learning.', 'start': 1974.497, 'duration': 1.101}, {'end': 1982.98, 'text': "Deep learning, another example is as you have like an exit sign, in this case it's translating it into French.", 'start': 1976.298, 'duration': 6.682}, {'end': 1985.819, 'text': 'Sortie I hope I said that right.', 'start': 1984.239, 'duration': 1.58}, {'end': 1994.502, 'text': "The neural network has been programmed with all these different words and images, and so it's able to look at the exit in the middle and it goes okay,", 'start': 1987.2, 'duration': 7.302}, {'end': 2000.403, 'text': "we want to know what that is in French and it's able to push that out in French and learn how to do that.", 'start': 1994.502, 'duration': 5.901}, {'end': 2003.004, 'text': 'And then we have chatbots.', 'start': 2001.843, 'duration': 1.161}, {'end': 2006.264, 'text': 'I remember when Microsoft first had their little paperclip.', 'start': 2003.804, 'duration': 2.46}, {'end': 2009.245, 'text': 'Boy, that was like a long time ago.', 'start': 2007.345, 'duration': 1.9}, {'end': 2012.246, 'text': 'They came up, and you would type in there and chat with it.', 'start': 2009.685, 'duration': 2.561}, {'end': 2013.699, 'text': 'These are growing.', 'start': 2012.998, 'duration': 0.701}, {'end': 2017.882, 'text': "You know, it's nice to just be able to ask a question and it comes up and gives you the answer.", 'start': 2013.719, 'duration': 4.163}, {'end': 2023.326, 'text': "And instead of it being where you're just doing a search on certain words,", 'start': 2018.602, 'duration': 4.724}, {'end': 2027.869, 'text': "it's now able to start linking those words together and form a sentence in that chat box.", 'start': 2023.326, 'duration': 4.543}, {'end': 2030.611, 'text': 'Types of AI and machine learning.', 'start': 2028.669, 'duration': 1.942}, {'end': 2033.449, 'text': 'Types of artificial intelligence.', 'start': 2031.909, 'duration': 1.54}, {'end': 2036.05, 'text': 'This and the next few slides are really important.', 'start': 2034.009, 'duration': 2.041}, {'end': 2039.891, 'text': 'So one of the types of artificial intelligence is reactive machines.', 'start': 2036.45, 'duration': 3.441}, {'end': 2041.911, 'text': 'Systems that only react.', 'start': 2040.471, 'duration': 1.44}, {'end': 2043.011, 'text': "They don't form memories.", 'start': 2041.991, 'duration': 1.02}, {'end': 2044.431, 'text': "They don't have past experiences.", 'start': 2043.051, 'duration': 1.38}, {'end': 2046.912, 'text': 'They have something that happens to them and they react to it.', 'start': 2044.591, 'duration': 2.321}, {'end': 2049.891, 'text': 'My washing machine is one of those.', 'start': 2047.732, 'duration': 2.159}, {'end': 2058.074, 'text': 'If I put a ton of clothes in it and they get all clumped on one side, it automatically adds a weight to re-center it.', 'start': 2050.612, 'duration': 7.462}, {'end': 2063.752, 'text': 'So my washing machine is actually a reactive machine working with whatever the load is.', 'start': 2058.668, 'duration': 5.084}, {'end': 2067.893, 'text': "It keeps it nice and so when it spins it doesn't go thumping against the side.", 'start': 2064.032, 'duration': 3.861}, {'end': 2069.195, 'text': 'Limited memory.', 'start': 2068.514, 'duration': 0.681}, {'end': 2071.016, 'text': 'Another form of artificial intelligence.', 'start': 2069.495, 'duration': 1.521}, {'end': 2073.078, 'text': 'Systems look into the past.', 'start': 2071.677, 'duration': 1.401}, {'end': 2078.04, 'text': 'Information is added over a period of time and information is short-lived.', 'start': 2073.838, 'duration': 4.202}, {'end': 2087.025, 'text': "When we're talking about this and you look at like a neural network that's been programmed to identify cars, it doesn't remember all those pictures.", 'start': 2078.061, 'duration': 8.964}, {'end': 2091.552, 'text': 'It has no memory as far as the hundreds of pictures you process through it.', 'start': 2087.791, 'duration': 3.761}, {'end': 2097.953, 'text': 'All it has is, this is the pattern I used to identify cars as the final output for that neural network we looked at.', 'start': 2092.132, 'duration': 5.821}, {'end': 2101.154, 'text': "So when they talk about limited memory, this is what they're talking about.", 'start': 2098.813, 'duration': 2.341}, {'end': 2106.075, 'text': "They're talking about, I've created this based on all these things, but I'm not going to remember any one specifically.", 'start': 2101.174, 'duration': 4.901}, {'end': 2107.815, 'text': 'Theory of mind.', 'start': 2107.015, 'duration': 0.8}, {'end': 2115.617, 'text': 'Systems being able to understand human emotions and how they affect decision making to adjust their behaviors according to their human understandings.', 'start': 2108.255, 'duration': 7.362}, {'end': 2118.471, 'text': 'This is important because this is our page mark.', 'start': 2116.55, 'duration': 1.921}, {'end': 2122.612, 'text': 'This is how we know whether it is an artificial intelligence or not.', 'start': 2118.511, 'duration': 4.101}, {'end': 2131.014, 'text': "Is it interacting with humans in a way that we can understand? Without that interaction, it's just an object.", 'start': 2123.372, 'duration': 7.642}, {'end': 2135.896, 'text': 'So when we talk about theory of mind, we really understand how it interfaces that whole.', 'start': 2131.515, 'duration': 4.381}, {'end': 2139.637, 'text': "if you're in web development, user experience would be the term I would put in there.", 'start': 2135.896, 'duration': 3.741}, {'end': 2142.218, 'text': 'So the theory of mind would be user experience.', 'start': 2140.237, 'duration': 1.981}, {'end': 2144.319, 'text': 'How is the whole UI connected together?', 'start': 2142.258, 'duration': 2.061}, {'end': 2150.576, 'text': 'And one of the final things is as we get into artificial intelligence, is systems being aware of themselves,', 'start': 2144.759, 'duration': 5.817}, {'end': 2155.798, 'text': "understanding their internal states and predicting other people's feelings and act appropriately?", 'start': 2150.576, 'duration': 5.222}, {'end': 2162.92, 'text': 'So, as artificial intelligence continues to progress, we see ones that are trying to understand well what makes people happy?', 'start': 2156.318, 'duration': 6.602}, {'end': 2164.501, 'text': 'How would they increase our happiness?', 'start': 2163.18, 'duration': 1.321}, {'end': 2168.902, 'text': "How would they keep themselves from breaking down if something's broken inside?", 'start': 2165.481, 'duration': 3.421}, {'end': 2170.903, 'text': 'They have that self-awareness to be able to fix it.', 'start': 2168.942, 'duration': 1.961}, {'end': 2177.365, 'text': 'And just based on all that information, predicting which action would work the best.', 'start': 2171.483, 'duration': 5.882}, {'end': 2178.786, 'text': 'What would help people?', 'start': 2177.606, 'duration': 1.18}, {'end': 2184.869, 'text': "If I know that you're having a cup of coffee first thing in the morning is what makes you happy as a robot.", 'start': 2179.847, 'duration': 5.022}, {'end': 2190.332, 'text': 'I might make you a cup of coffee every morning at the same time to help your life and help you grow.', 'start': 2184.869, 'duration': 5.463}, {'end': 2193.414, 'text': "That'd be the self-awareness is being able to know all those different things.", 'start': 2190.352, 'duration': 3.062}, {'end': 2198.363, 'text': 'Types of machine learning, and like I said on the last slide, this is very important.', 'start': 2194.281, 'duration': 4.082}, {'end': 2199.464, 'text': 'This is very important.', 'start': 2198.564, 'duration': 0.9}, {'end': 2207.429, 'text': 'If you decide to go in and get certified in machine learning or know more about it, these are the three primary types of machine learning.', 'start': 2199.844, 'duration': 7.585}, {'end': 2210.27, 'text': 'The first one is supervised learning.', 'start': 2208.149, 'duration': 2.121}, {'end': 2214.573, 'text': 'Systems are able to predict future outcome based on past data.', 'start': 2210.631, 'duration': 3.942}, {'end': 2219.036, 'text': 'Requires both an input and an output to be given to the model for it to be trained.', 'start': 2215.113, 'duration': 3.923}, {'end': 2224.722, 'text': "So in this case, we're looking at anything where you have 100 images of a bicycle.", 'start': 2220.037, 'duration': 4.685}, {'end': 2228.505, 'text': 'And those 100 images you know are bicycle.', 'start': 2225.723, 'duration': 2.782}, {'end': 2229.706, 'text': "So they're presets.", 'start': 2228.645, 'duration': 1.061}, {'end': 2232.609, 'text': 'Someone already looked at all 100 images and said, these are pictures of bicycles.', 'start': 2229.746, 'duration': 2.863}, {'end': 2234.731, 'text': 'And so the computer learns from those.', 'start': 2233.31, 'duration': 1.421}, {'end': 2236.893, 'text': "And then it's given another picture.", 'start': 2235.112, 'duration': 1.781}, {'end': 2243.878, 'text': "And maybe the next picture is a bicycle and it says, oh, that resembles all these other bicycles, so it's a bicycle.", 'start': 2237.894, 'duration': 5.984}, {'end': 2246.699, 'text': "And the next one's a car and it says, eh, it's not a bicycle.", 'start': 2244.418, 'duration': 2.281}, {'end': 2250.081, 'text': 'That would be supervised learning because we had to train it.', 'start': 2247.34, 'duration': 2.741}, {'end': 2251.122, 'text': 'We had to supervise it.', 'start': 2250.181, 'duration': 0.941}, {'end': 2252.643, 'text': 'Unsupervised learning.', 'start': 2251.602, 'duration': 1.041}, {'end': 2256.505, 'text': 'Systems are able to identify hidden patterns from the input data provided.', 'start': 2253.003, 'duration': 3.502}, {'end': 2262.847, 'text': 'By making the data more readable and organized, the patterns, similarities, or anomalies become more evident.', 'start': 2257.185, 'duration': 5.662}, {'end': 2264.967, 'text': "You'll heard the term cluster.", 'start': 2263.407, 'duration': 1.56}, {'end': 2268.948, 'text': "How do you cluster things together? Some of these things go together, some of these don't.", 'start': 2265.327, 'duration': 3.621}, {'end': 2276.611, 'text': "This is unsupervised where it can look at an image and start pulling the different pieces of the image out because they aren't the same.", 'start': 2269.349, 'duration': 7.262}, {'end': 2284.253, 'text': "All the parts of the human are not the same as a fuzzy tree behind them because it's slightly out of focus, which is not the same as the beach ball.", 'start': 2277.471, 'duration': 6.782}, {'end': 2290.554, 'text': "It's unsupervised because we never told it what a beach ball was, we never told it what the human was, and we never told it that those were trees.", 'start': 2284.873, 'duration': 5.681}, {'end': 2297.795, 'text': "All we told it was, hey, separate this picture by things that don't match and things that do match and come together.", 'start': 2291.254, 'duration': 6.541}, {'end': 2300.436, 'text': "And finally, there's reinforcement learning.", 'start': 2298.715, 'duration': 1.721}, {'end': 2302.376, 'text': 'Systems are given no training.', 'start': 2300.776, 'duration': 1.6}, {'end': 2307.277, 'text': 'It learns on the basis of the reward punishment it received for performing its last action.', 'start': 2302.596, 'duration': 4.681}, {'end': 2310.997, 'text': 'It helps increase the efficiency of a tool function or a program.', 'start': 2307.817, 'duration': 3.18}, {'end': 2316.119, 'text': 'Reinforced learning, or reinforcement learning, is kind of you give it a yes or no.', 'start': 2311.177, 'duration': 4.942}, {'end': 2317.66, 'text': 'Yes, you gave me the right response.', 'start': 2316.319, 'duration': 1.341}, {'end': 2318.421, 'text': "No, you didn't.", 'start': 2317.88, 'duration': 0.541}, {'end': 2325.505, 'text': 'And then it looks at that and says oh okay, so, based on this data coming in, what I gave you was a wrong response,', 'start': 2319.061, 'duration': 6.444}, {'end': 2326.946, 'text': "so next time I'll give you a different one.", 'start': 2325.505, 'duration': 1.441}, {'end': 2329.982, 'text': 'Comparing machine learning and deep learning.', 'start': 2327.901, 'duration': 2.081}, {'end': 2333.465, 'text': 'So remember that deep learning is a subcategory of machine learning.', 'start': 2330.122, 'duration': 3.343}, {'end': 2334.946, 'text': "So it's one of the many tools.", 'start': 2333.645, 'duration': 1.301}, {'end': 2339.048, 'text': 'And so they were grouping a ton of machine learning tools all together.', 'start': 2335.566, 'duration': 3.482}, {'end': 2345.312, 'text': "Linear regression, k-means clustering, there's all kinds of cool tools out there you can use in machine learning.", 'start': 2339.308, 'duration': 6.004}, {'end': 2350.836, 'text': 'Enables machines to take decisions, to make decisions on their own based on past data.', 'start': 2345.792, 'duration': 5.044}, {'end': 2355.579, 'text': 'enables machines to make decisions with the help of artificial neural networks.', 'start': 2351.616, 'duration': 3.963}, {'end': 2362.103, 'text': "So it's doing the same thing, but we're using an artificial neural network as opposed to one of the more traditional machine learning tools.", 'start': 2356.099, 'duration': 6.004}, {'end': 2364.825, 'text': 'Needs only a small amount of training data.', 'start': 2362.844, 'duration': 1.981}, {'end': 2367.687, 'text': "This is very important when you're talking about machine learning.", 'start': 2365.345, 'duration': 2.342}, {'end': 2370.712, 'text': "They're usually not talking about huge amounts of data.", 'start': 2368.53, 'duration': 2.182}, {'end': 2375.014, 'text': "We're talking about maybe your spreadsheet from your business and your totals for the end of the year.", 'start': 2370.812, 'duration': 4.202}, {'end': 2379.818, 'text': "When you're talking about neural networks, you usually need a large amount of data to train the data.", 'start': 2375.034, 'duration': 4.784}, {'end': 2381.859, 'text': "So there's a lot of training involved.", 'start': 2380.238, 'duration': 1.621}, {'end': 2387.303, 'text': "If you have under 500 points of data, that's probably not going to go into machine learning.", 'start': 2381.979, 'duration': 5.324}, {'end': 2393.667, 'text': 'Or maybe you have like the case of one of the things, 500 points of data and 30 different fields.', 'start': 2387.823, 'duration': 5.844}, {'end': 2398.21, 'text': 'It starts getting really confusing there in artificial intelligence or machine learning.', 'start': 2393.687, 'duration': 4.523}, {'end': 2403.534, 'text': "And the deep learning aspect really shines when you get to that larger data that's really complex.", 'start': 2398.631, 'duration': 4.903}, {'end': 2406.336, 'text': 'Works well on a low end systems.', 'start': 2404.555, 'duration': 1.781}, {'end': 2413.461, 'text': 'So a lot of the machine learning tools out there you can run on your laptop with no problem and do the calculations there.', 'start': 2407.236, 'duration': 6.225}, {'end': 2417.742, 'text': 'Where with the machine learning, it usually needs a higher end system to work.', 'start': 2414.4, 'duration': 3.342}, {'end': 2422.385, 'text': 'It takes a lot more processing power to build those neural networks and to train them.', 'start': 2417.782, 'duration': 4.603}, {'end': 2423.726, 'text': 'It goes through a lot of data.', 'start': 2422.465, 'duration': 1.261}, {'end': 2426.508, 'text': "We're talking about the general machine learning tools.", 'start': 2424.547, 'duration': 1.961}, {'end': 2430.11, 'text': 'Most features need to be identified in advanced and manually coded.', 'start': 2426.768, 'duration': 3.342}, {'end': 2432.352, 'text': "So there's a lot of human work on here.", 'start': 2430.13, 'duration': 2.222}, {'end': 2436.014, 'text': 'The machine learns the features from the data it is provided.', 'start': 2432.992, 'duration': 3.022}, {'end': 2437.895, 'text': "So again, it's like a magic box.", 'start': 2436.314, 'duration': 1.581}, {'end': 2440.057, 'text': "You don't have to know what a tire is.", 'start': 2438.395, 'duration': 1.662}, {'end': 2441.858, 'text': 'It figures it out for you.', 'start': 2440.677, 'duration': 1.181}, {'end': 2446.852, 'text': 'The problem is divided into parts and solved individually and then combined.', 'start': 2442.95, 'duration': 3.902}, {'end': 2451.295, 'text': 'So in machine learning, you usually have all these different tools and use different tools for different parts.', 'start': 2447.333, 'duration': 3.962}, {'end': 2454.517, 'text': 'And the problem is solved in an end-to-end manner.', 'start': 2452.356, 'duration': 2.161}, {'end': 2460, 'text': 'So you only have one neural network or two neural networks that is bringing the data in and putting it out.', 'start': 2454.537, 'duration': 5.463}, {'end': 2463.643, 'text': "It's not going through a lot of different processes to get there.", 'start': 2460.581, 'duration': 3.062}, {'end': 2466.624, 'text': 'And remember, you can put machine learning and deep learning together.', 'start': 2464.203, 'duration': 2.421}, {'end': 2470.667, 'text': "So you don't always have just the deep learning solving the problem.", 'start': 2466.865, 'duration': 3.802}, {'end': 2472.368, 'text': 'You might have it solving one piece of the puzzle.', 'start': 2470.687, 'duration': 1.681}, {'end': 2479.457, 'text': 'With regular machine learning and most machine learning tools out there, they take longer to test and understand how they work.', 'start': 2473.553, 'duration': 5.904}, {'end': 2481.879, 'text': "And with the deep learning, it's pretty quick.", 'start': 2480.278, 'duration': 1.601}, {'end': 2484.461, 'text': 'Once you build that neural network, you test it and you know.', 'start': 2482.179, 'duration': 2.282}, {'end': 2488.721, 'text': "So we're dealing with very crisp rules, limited resources.", 'start': 2485.799, 'duration': 2.922}, {'end': 2493.565, 'text': 'You have to really explain how the decision was made when you use most machine learning tools.', 'start': 2489.542, 'duration': 4.023}, {'end': 2500.731, 'text': 'But when you use the deep learning tool inside the machine learning tools, the system takes care of it based on its own logic and reasoning.', 'start': 2494.066, 'duration': 6.665}, {'end': 2502.673, 'text': "And again, it's like a magic black box.", 'start': 2500.991, 'duration': 1.682}, {'end': 2505.775, 'text': "You really don't know how it came up with the answer.", 'start': 2503.333, 'duration': 2.442}, {'end': 2507.757, 'text': 'You just know it came up with the right answer.', 'start': 2505.875, 'duration': 1.882}, {'end': 2509.678, 'text': 'A glimpse into the future.', 'start': 2508.437, 'duration': 1.241}, {'end': 2511.92, 'text': 'So a quick glimpse into the future.', 'start': 2510.439, 'duration': 1.481}, {'end': 2513.622, 'text': 'Artificial intelligence.', 'start': 2512.421, 'duration': 1.201}, {'end': 2516.605, 'text': 'Be using it to detecting crimes before they happen.', 'start': 2514.122, 'duration': 2.483}, {'end': 2520.048, 'text': 'Humanoid AI helpers, which we already have a lot of.', 'start': 2517.285, 'duration': 2.763}, {'end': 2521.089, 'text': "There'll be more and more.", 'start': 2520.088, 'duration': 1.001}, {'end': 2522.67, 'text': "Maybe it'll actually be androids.", 'start': 2521.409, 'duration': 1.261}], 'summary': 'Deep learning is a subset of machine learning, a part of artificial intelligence, which is rapidly advancing and has diverse applications.', 'duration': 842.752, 'max_score': 1679.918, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM1679918.jpg'}, {'end': 1731.542, 'src': 'embed', 'start': 1703.698, 'weight': 3, 'content': [{'end': 1711.12, 'text': 'You put in your information, you then have your group of tests, and then you held some aside, and you see how does it do.', 'start': 1703.698, 'duration': 7.422}, {'end': 1714.681, 'text': "It's very quick to test it and see what's going on with your deep learning and your neural network.", 'start': 1711.16, 'duration': 3.521}, {'end': 1721.358, 'text': 'Are they really all that different? AI versus machine learning versus deep learning.', 'start': 1715.647, 'duration': 5.711}, {'end': 1723.279, 'text': 'Concepts of AI.', 'start': 1721.978, 'duration': 1.301}, {'end': 1725.82, 'text': 'So we have concepts of AI.', 'start': 1724.099, 'duration': 1.721}, {'end': 1731.542, 'text': "You'll see natural language processing, machine learning, an approach to create artificial intelligence.", 'start': 1726.1, 'duration': 5.442}], 'summary': 'Testing and comparing ai concepts like natural language processing and machine learning for creating artificial intelligence.', 'duration': 27.844, 'max_score': 1703.698, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM1703698.jpg'}, {'end': 2413.461, 'src': 'embed', 'start': 2387.823, 'weight': 4, 'content': [{'end': 2393.667, 'text': 'Or maybe you have like the case of one of the things, 500 points of data and 30 different fields.', 'start': 2387.823, 'duration': 5.844}, {'end': 2398.21, 'text': 'It starts getting really confusing there in artificial intelligence or machine learning.', 'start': 2393.687, 'duration': 4.523}, {'end': 2403.534, 'text': "And the deep learning aspect really shines when you get to that larger data that's really complex.", 'start': 2398.631, 'duration': 4.903}, {'end': 2406.336, 'text': 'Works well on a low end systems.', 'start': 2404.555, 'duration': 1.781}, {'end': 2413.461, 'text': 'So a lot of the machine learning tools out there you can run on your laptop with no problem and do the calculations there.', 'start': 2407.236, 'duration': 6.225}], 'summary': 'Deep learning excels with large, complex data, performing well on low-end systems.', 'duration': 25.638, 'max_score': 2387.823, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM2387823.jpg'}, {'end': 2705.269, 'src': 'embed', 'start': 2678.596, 'weight': 8, 'content': [{'end': 2685.659, 'text': 'Well, this is a huge thing in a social media when people are getting spammed, and so this tactic, known as engagement bait,', 'start': 2678.596, 'duration': 7.063}, {'end': 2691.321, 'text': "takes advantage of Facebook's newsfeed algorithm by choosing engagement in order to get the greater reach.", 'start': 2685.659, 'duration': 5.662}, {'end': 2693.842, 'text': 'To eliminate engagement bait.', 'start': 2691.841, 'duration': 2.001}, {'end': 2700.665, 'text': 'the company reviewed and categorized hundreds of thousands of posts to train a machine learning model that detects different types of engagement bait.', 'start': 2693.842, 'duration': 6.823}, {'end': 2705.269, 'text': "So in this case, we're using Facebook, but this is, of course, across all the different social media.", 'start': 2701.005, 'duration': 4.264}], 'summary': 'Facebook tackles engagement bait by training ml model to detect and eliminate it.', 'duration': 26.673, 'max_score': 2678.596, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM2678596.jpg'}, {'end': 2763.662, 'src': 'embed', 'start': 2712.294, 'weight': 0, 'content': [{'end': 2716.137, 'text': "It notices that there's a certain setup with Facebook and it's able to replace it.", 'start': 2712.294, 'duration': 3.843}, {'end': 2720.881, 'text': 'And they have like vote baiting, react baiting, share baiting.', 'start': 2716.678, 'duration': 4.203}, {'end': 2722.002, 'text': 'They have all these different.', 'start': 2721.141, 'duration': 0.861}, {'end': 2727.066, 'text': 'these are kind of general titles, but there certainly are a lot of way of baiting you to go in there and click on something.', 'start': 2722.002, 'duration': 5.064}, {'end': 2728.507, 'text': 'So they fed all this.', 'start': 2727.606, 'duration': 0.901}, {'end': 2730.188, 'text': 'This data was fed into the machine.', 'start': 2728.687, 'duration': 1.501}, {'end': 2731.809, 'text': 'And then they have the new post.', 'start': 2730.508, 'duration': 1.301}, {'end': 2735.231, 'text': 'The new post comes up that takes over part of the Facebook setup.', 'start': 2731.829, 'duration': 3.402}, {'end': 2736.332, 'text': "And that's what you're looking at.", 'start': 2735.451, 'duration': 0.881}, {'end': 2739.614, 'text': "You're looking at this new post that's replaced, like a virus has replaced that.", 'start': 2736.352, 'duration': 3.262}, {'end': 2746.838, 'text': 'So what Facebook did to eliminate this is they start scanning for keywords and like this and checks the click-through rate,', 'start': 2739.794, 'duration': 7.044}, {'end': 2751.919, 'text': 'so starts looking for people who are clicking through it without even looking at it or clicking through it.', 'start': 2746.838, 'duration': 5.081}, {'end': 2754.159, 'text': "it's not something that normally would be clicked through.", 'start': 2751.919, 'duration': 2.24}, {'end': 2762.321, 'text': 'once Facebook has scanned for these keywords and phrases, it is now able to identify the spam coming in, and this makes your life easier.', 'start': 2754.159, 'duration': 8.162}, {'end': 2763.662, 'text': "so you're not getting spammed.", 'start': 2762.321, 'duration': 1.341}], 'summary': 'Facebook uses scanning for keywords to identify and eliminate spam, improving user experience.', 'duration': 51.368, 'max_score': 2712.294, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM2712294.jpg'}], 'start': 1461.533, 'title': 'Machine learning and ai', 'summary': 'Explores deep learning, machine learning, and ai, emphasizing their potential to revolutionize industries, discussing types of ai and machine learning, their tools and potential applications, and advancements in various sectors including healthcare, social media, and gaming.', 'chapters': [{'end': 1745.887, 'start': 1461.533, 'title': 'Understanding deep learning and machine learning', 'summary': 'Explores the concept of deep learning and machine learning, highlighting their ability to process large amounts of data quickly and accurately, with deep learning using artificial neural networks to improve performance with more data and machine learning making predictions based on past data. it also emphasizes the affordability and scalability of machine learning, showcasing its potential to revolutionize businesses and industries.', 'duration': 284.354, 'highlights': ['Deep learning uses artificial neural networks to improve performance with more data', 'Machine learning makes predictions based on past data', 'Machine learning is quick and accurate, with the capability to process large amounts of data', 'Machine learning is affordable and scalable', 'Deep learning and machine learning are subsets of artificial intelligence']}, {'end': 2013.699, 'start': 1747.304, 'title': 'Ai with machine learning and deep learning', 'summary': 'Discusses the role of ai as the big picture, with machine learning and deep learning as subsets, exploring examples such as text mining, book classification, and real-life applications of artificial intelligence and machine learning.', 'duration': 266.395, 'highlights': ['AI is the big picture, with machine learning and deep learning as subsets', 'Examples of text mining, book classification, and real-life applications', 'Role of deep learning in language translation and chatbots']}, {'end': 2387.303, 'start': 2013.719, 'title': 'Types of ai & machine learning', 'summary': 'Discusses the types of artificial intelligence including reactive machines, limited memory systems, theory of mind, and self-aware systems. it also covers the three primary types of machine learning: supervised learning, unsupervised learning, and reinforcement learning.', 'duration': 373.584, 'highlights': ['Reactive machines are one of the types of artificial intelligence, which react to events without forming memories, such as a washing machine automatically adjusting the load to prevent thumping.', 'Limited memory systems in AI look into the past but have short-lived information, exemplified by neural networks not remembering past pictures when identifying patterns like cars.', 'Theory of mind in AI involves systems understanding human emotions to adjust behaviors, essential for human interaction and user experience.', 'Self-aware systems in AI understand their internal states, predict human feelings, and act appropriately, aiming to enhance human happiness and functionality.', 'Supervised learning in machine learning involves predicting future outcomes based on past data, requiring both input and output for training.', 'Unsupervised learning in machine learning identifies hidden patterns from input data and organizes it to reveal similarities or anomalies.', 'Reinforcement learning in machine learning involves systems learning on the basis of the reward or punishment received for performing actions, improving the efficiency of functions or programs.']}, {'end': 2620.646, 'start': 2387.823, 'title': 'Machine learning: tools and potential', 'summary': 'Discusses the differences between machine learning and deep learning, their capabilities, and potential applications, including the use of machine learning in healthcare and the future prospects of ai. it also emphasizes the increasing role of deep learning in personalization and the development of hyper-intelligent personal assistants.', 'duration': 232.823, 'highlights': ['The deep learning aspect shines with larger and complex data, working well even on low-end systems.', 'Machine learning typically requires a higher-end system for building and training neural networks, demanding more processing power.', 'Deep learning enables quicker testing and understanding of models compared to most machine learning tools.', 'The potential applications of artificial intelligence and machine learning include the detection of crimes before they happen, increased efficiency in healthcare, and advancements in marketing techniques.', 'The increasing role of deep learning in personalization and the development of hyper-intelligent personal assistants is highlighted, offering a glimpse into the future of AI.']}, {'end': 2881.094, 'start': 2620.866, 'title': 'Advancements in machine learning', 'summary': 'Discusses the impact of machine learning in various sectors, including healthcare, social media, and gaming, highlighting the use of technology to pre-diagnose illnesses, combat engagement bait on social media, and achieve breakthroughs in game-playing ai, showcasing the potential and challenges of an increasingly automated world.', 'duration': 260.228, 'highlights': ["Facebook's effort to combat engagement bait on social media", 'Advancements in healthcare through machine learning and pre-diagnosis technology', "Google's DeepMind project AlphaGo's victory over the world's top Go player", 'Explanation of machine learning and its predictive capabilities', 'Challenges and potential risks in the development of machine learning and AI']}], 'duration': 1419.561, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM1461533.jpg', 'highlights': ['Deep learning and machine learning are subsets of artificial intelligence', 'The potential applications of artificial intelligence and machine learning include the detection of crimes before they happen, increased efficiency in healthcare, and advancements in marketing techniques', 'The increasing role of deep learning in personalization and the development of hyper-intelligent personal assistants is highlighted, offering a glimpse into the future of AI', 'Reinforcement learning in machine learning involves systems learning on the basis of the reward or punishment received for performing actions, improving the efficiency of functions or programs', 'AI is the big picture, with machine learning and deep learning as subsets', 'Deep learning uses artificial neural networks to improve performance with more data', 'Machine learning makes predictions based on past data', 'Machine learning is quick and accurate, with the capability to process large amounts of data', 'Machine learning is affordable and scalable', 'The deep learning aspect shines with larger and complex data, working well even on low-end systems']}, {'end': 3770.114, 'segs': [{'end': 3081.282, 'src': 'embed', 'start': 3053.963, 'weight': 0, 'content': [{'end': 3057.205, 'text': "You want to know when someone's making a withdrawal that might not be their own account.", 'start': 3053.963, 'duration': 3.242}, {'end': 3060.367, 'text': "We've actually brought this up because this is really big right now.", 'start': 3057.525, 'duration': 2.842}, {'end': 3067.552, 'text': "If you're predicting the stock, whether to buy stock or not, you want to be able to know if what's going on in the stock market is an anomaly.", 'start': 3060.647, 'duration': 6.905}, {'end': 3070.354, 'text': "It's a different prediction model because something else is going on.", 'start': 3067.792, 'duration': 2.562}, {'end': 3072.315, 'text': "You've got to pull out new information in there.", 'start': 3070.374, 'duration': 1.941}, {'end': 3077.059, 'text': "Or is this just the norm? I'm going to get my normal return on my money invested.", 'start': 3072.795, 'duration': 4.264}, {'end': 3081.282, 'text': 'So being able to detect anomalies is very big in data science these days.', 'start': 3077.419, 'duration': 3.863}], 'summary': 'Detecting anomalies in stock market data is crucial for predicting stock outcomes and making informed investment decisions.', 'duration': 27.319, 'max_score': 3053.963, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM3053963.jpg'}, {'end': 3572.79, 'src': 'heatmap', 'start': 3358.432, 'weight': 0.789, 'content': [{'end': 3360.974, 'text': "And it's able to filter those through and group them together.", 'start': 3358.432, 'duration': 2.542}, {'end': 3364.955, 'text': 'We talked about that earlier with looking at a group of people who are out shopping.', 'start': 3361.234, 'duration': 3.721}, {'end': 3368.177, 'text': 'We want to group them together to find out what they have in common.', 'start': 3365.215, 'duration': 2.962}, {'end': 3374.521, 'text': "And, of course, once you understand what people have in common, Maybe you have one of them who's a customer at your store,", 'start': 3368.537, 'duration': 5.984}, {'end': 3380.448, 'text': 'or you have five of them who are a customer at your store and they have a lot in common with five others who are not customers at your store.', 'start': 3374.521, 'duration': 5.927}, {'end': 3383.311, 'text': "How do you market to those five who aren't customers at your store yet?", 'start': 3380.688, 'duration': 2.623}, {'end': 3388.237, 'text': "They fit the demographics of who's going to shop there, and you'd like them to shop at your store, not the one next door.", 'start': 3383.471, 'duration': 4.766}, {'end': 3390.158, 'text': 'Of course, this is a simplified version.', 'start': 3388.677, 'duration': 1.481}, {'end': 3394.922, 'text': 'You can see very easily the difference between a triangle and a circle, which might not be so easy in marketing.', 'start': 3390.198, 'duration': 4.724}, {'end': 3396.484, 'text': 'Reinforcement learning.', 'start': 3395.343, 'duration': 1.141}, {'end': 3404.671, 'text': 'Reinforcement. learning is an important type of machine learning, where an agent learns how to behave in an environment by performing actions and seeing the result.', 'start': 3396.684, 'duration': 7.987}, {'end': 3407.413, 'text': 'We have here, in this case, a baby.', 'start': 3405.191, 'duration': 2.222}, {'end': 3414.118, 'text': "It's actually great that they used an infant for this slide because the reinforcement learning is very much in its infant stages.", 'start': 3407.753, 'duration': 6.365}, {'end': 3425.166, 'text': "But it's also probably the biggest machine learning demand out there right now or in the future it's going to be coming up over the next few years is reinforcement learning and how to make that work for us.", 'start': 3414.618, 'duration': 10.548}, {'end': 3427.648, 'text': 'And you can see here where we have our action.', 'start': 3425.746, 'duration': 1.902}, {'end': 3430.51, 'text': 'In the action in this one, It goes into the fire.', 'start': 3427.888, 'duration': 2.622}, {'end': 3434.591, 'text': "Hopefully the baby didn't, it was just a little candle, not a giant fire pit like it looks like here.", 'start': 3430.75, 'duration': 3.841}, {'end': 3440.054, 'text': 'When the baby comes out and the new state is the baby is sad and crying because they got burned on the fire.', 'start': 3434.872, 'duration': 5.182}, {'end': 3441.935, 'text': 'And then maybe they take another action.', 'start': 3440.454, 'duration': 1.481}, {'end': 3445.156, 'text': "The baby's called the agent because it's the one taking the actions.", 'start': 3442.275, 'duration': 2.881}, {'end': 3447.317, 'text': "And in this case, they didn't go into the fire.", 'start': 3445.577, 'duration': 1.74}, {'end': 3450.579, 'text': "They went a different direction and now the baby's happy and laughing and playing.", 'start': 3447.357, 'duration': 3.222}, {'end': 3456.544, 'text': "Reinforcement learning is very easy to understand because that's how, as humans, that's one of the ways we learn.", 'start': 3451.039, 'duration': 5.505}, {'end': 3461.529, 'text': "We learn whether it is you burn yourself on the stove, don't do that anymore, don't touch the stove.", 'start': 3456.684, 'duration': 4.845}, {'end': 3467.935, 'text': 'In the big picture, being able to have a machine learning program or an AI be able to do this is huge,', 'start': 3462.009, 'duration': 5.926}, {'end': 3470.937, 'text': "because now we're starting to learn how to learn.", 'start': 3467.935, 'duration': 3.002}, {'end': 3474.661, 'text': "That's a big jump in the world of computer and machine learning.", 'start': 3471.198, 'duration': 3.463}, {'end': 3479.906, 'text': "And we're going to go back and just kind of go back over supervised versus unsupervised learning.", 'start': 3474.941, 'duration': 4.965}, {'end': 3484.69, 'text': "Understanding this is huge because this is going to come up in any project you're working on.", 'start': 3480.166, 'duration': 4.524}, {'end': 3488.494, 'text': 'We have in supervised learning, we have labeled data.', 'start': 3485.251, 'duration': 3.243}, {'end': 3490.155, 'text': 'We have direct feedback.', 'start': 3488.794, 'duration': 1.361}, {'end': 3493.338, 'text': "So someone's already gone in there and said, yes, that's a triangle.", 'start': 3490.455, 'duration': 2.883}, {'end': 3494.659, 'text': "No, that's not a triangle.", 'start': 3493.558, 'duration': 1.101}, {'end': 3496.121, 'text': 'And then you predict an outcome.', 'start': 3494.92, 'duration': 1.201}, {'end': 3497.342, 'text': 'So you have a nice prediction.', 'start': 3496.301, 'duration': 1.041}, {'end': 3500.705, 'text': "This new set of data is coming in and we know what it's going to be.", 'start': 3497.382, 'duration': 3.323}, {'end': 3506.528, 'text': "and then with unsupervised training, it's not labeled, so we really don't know what it is.", 'start': 3501.005, 'duration': 5.523}, {'end': 3510.051, 'text': "there's no feedback, so we're not telling it whether it's right or wrong.", 'start': 3506.528, 'duration': 3.523}, {'end': 3512.972, 'text': "we're not telling it whether it's a triangle or a square.", 'start': 3510.051, 'duration': 2.921}, {'end': 3515.314, 'text': "we're not telling it to go left or right.", 'start': 3512.972, 'duration': 2.342}, {'end': 3522.098, 'text': "all we do is we're finding hidden structure in the data, grouping the data together to find out what connects to each other,", 'start': 3515.314, 'duration': 6.784}, {'end': 3524.58, 'text': 'and then you can use these together.', 'start': 3522.598, 'duration': 1.982}, {'end': 3528.884, 'text': "so imagine you have an image and you're not sure what you're looking for.", 'start': 3524.58, 'duration': 4.304}, {'end': 3537.033, 'text': 'so you go in and you have the unstructured data, find all these things that are connected together and then somebody looks at those and labels them.', 'start': 3528.884, 'duration': 8.149}, {'end': 3541.897, 'text': "now you can take that label data and program something to predict what's in the picture,", 'start': 3537.033, 'duration': 4.864}, {'end': 3547.642, 'text': 'so you can see how they go back and forth and you can start connecting all these different tools together to make a bigger picture.', 'start': 3541.897, 'duration': 5.745}, {'end': 3551.123, 'text': 'There are many interesting machine learning algorithms.', 'start': 3548.122, 'duration': 3.001}, {'end': 3552.644, 'text': "Let's have a look at a few of them.", 'start': 3551.343, 'duration': 1.301}, {'end': 3555.005, 'text': "Hopefully this gives you a little flavor of what's out there.", 'start': 3552.904, 'duration': 2.101}, {'end': 3558.467, 'text': 'And these are some of the most important ones that are currently being used.', 'start': 3555.245, 'duration': 3.222}, {'end': 3563.569, 'text': "We'll take a look at linear regression, decision tree, and the support vector machine.", 'start': 3558.787, 'duration': 4.782}, {'end': 3566.831, 'text': "Let's start with a closer look at linear regression.", 'start': 3564.15, 'duration': 2.681}, {'end': 3572.79, 'text': 'Linear regression is perhaps one of the most well-known and well-understood algorithms in statistics and machine learning.', 'start': 3567.305, 'duration': 5.485}], 'summary': 'Understanding reinforcement learning and machine learning algorithms is crucial for ai development.', 'duration': 214.358, 'max_score': 3358.432, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM3358432.jpg'}, {'end': 3425.166, 'src': 'embed', 'start': 3396.684, 'weight': 9, 'content': [{'end': 3404.671, 'text': 'Reinforcement. learning is an important type of machine learning, where an agent learns how to behave in an environment by performing actions and seeing the result.', 'start': 3396.684, 'duration': 7.987}, {'end': 3407.413, 'text': 'We have here, in this case, a baby.', 'start': 3405.191, 'duration': 2.222}, {'end': 3414.118, 'text': "It's actually great that they used an infant for this slide because the reinforcement learning is very much in its infant stages.", 'start': 3407.753, 'duration': 6.365}, {'end': 3425.166, 'text': "But it's also probably the biggest machine learning demand out there right now or in the future it's going to be coming up over the next few years is reinforcement learning and how to make that work for us.", 'start': 3414.618, 'duration': 10.548}], 'summary': 'Reinforcement learning is in its infant stages, but is a growing demand in machine learning.', 'duration': 28.482, 'max_score': 3396.684, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM3396684.jpg'}, {'end': 3537.033, 'src': 'embed', 'start': 3497.382, 'weight': 4, 'content': [{'end': 3500.705, 'text': "This new set of data is coming in and we know what it's going to be.", 'start': 3497.382, 'duration': 3.323}, {'end': 3506.528, 'text': "and then with unsupervised training, it's not labeled, so we really don't know what it is.", 'start': 3501.005, 'duration': 5.523}, {'end': 3510.051, 'text': "there's no feedback, so we're not telling it whether it's right or wrong.", 'start': 3506.528, 'duration': 3.523}, {'end': 3512.972, 'text': "we're not telling it whether it's a triangle or a square.", 'start': 3510.051, 'duration': 2.921}, {'end': 3515.314, 'text': "we're not telling it to go left or right.", 'start': 3512.972, 'duration': 2.342}, {'end': 3522.098, 'text': "all we do is we're finding hidden structure in the data, grouping the data together to find out what connects to each other,", 'start': 3515.314, 'duration': 6.784}, {'end': 3524.58, 'text': 'and then you can use these together.', 'start': 3522.598, 'duration': 1.982}, {'end': 3528.884, 'text': "so imagine you have an image and you're not sure what you're looking for.", 'start': 3524.58, 'duration': 4.304}, {'end': 3537.033, 'text': 'so you go in and you have the unstructured data, find all these things that are connected together and then somebody looks at those and labels them.', 'start': 3528.884, 'duration': 8.149}], 'summary': 'Unsupervised training finds hidden data structure, grouping, and connecting data, then labeling for interpretation.', 'duration': 39.651, 'max_score': 3497.382, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM3497382.jpg'}, {'end': 3644.65, 'src': 'embed', 'start': 3617.686, 'weight': 2, 'content': [{'end': 3621.73, 'text': 'And so over a certain amount of time, his distance equals 36 kilometers.', 'start': 3617.686, 'duration': 4.044}, {'end': 3626.614, 'text': "We have a second bicyclist who's going twice the speed, or 20 meters per second.", 'start': 3622.331, 'duration': 4.283}, {'end': 3632.499, 'text': "And you can guess if he's going twice the speed and time is a constant, then he's going to go twice the distance.", 'start': 3627.215, 'duration': 5.284}, {'end': 3634.301, 'text': "And that's easily to compute.", 'start': 3633.16, 'duration': 1.141}, {'end': 3638.004, 'text': '36 times 2, you get 72 kilometers.', 'start': 3634.641, 'duration': 3.363}, {'end': 3644.65, 'text': 'And so if you had the question of how fast would somebody is going three times that speed, or 30 meters per second is,', 'start': 3638.445, 'duration': 6.205}], 'summary': 'Bicyclist covers 36 km, second at 72 km, and third at 108 km.', 'duration': 26.964, 'max_score': 3617.686, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM3617686.jpg'}, {'end': 3742.209, 'src': 'embed', 'start': 3713.848, 'weight': 3, 'content': [{'end': 3719.251, 'text': "Since he's going twice the speed, we can guess that he'll cover the distance in about half the time, 50 seconds.", 'start': 3713.848, 'duration': 5.403}, {'end': 3725.695, 'text': "And of course, you could probably guess on the third one, 100 divided by 30, since he's going three times the speed,", 'start': 3719.832, 'duration': 5.863}, {'end': 3727.416, 'text': 'you could easily guess that this is 33.333 seconds time.', 'start': 3725.695, 'duration': 1.721}, {'end': 3732.721, 'text': 'We put that into a linear regression model or a graph.', 'start': 3730.098, 'duration': 2.623}, {'end': 3737.205, 'text': "If the distance is assumed to be constant, let's see the relationship between speed and time.", 'start': 3733.161, 'duration': 4.044}, {'end': 3742.209, 'text': 'And as time goes up, the amount of speed to go that same distance goes down.', 'start': 3737.505, 'duration': 4.704}], 'summary': 'Analyzing the relationship between speed and time using linear regression model, with examples indicating 50 and 33.333 seconds for different speeds.', 'duration': 28.361, 'max_score': 3713.848, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM3713848.jpg'}], 'start': 2881.274, 'title': 'Steps and overview of ml', 'summary': 'Covers the general steps in the machine learning process including defining objectives, collecting and preparing data, selecting and training algorithms, testing and running predictions, and deploying the model. it also provides an overview of machine learning models, including classification, regression, anomaly detection, clustering, supervised, unsupervised, and reinforcement learning, with a focus on linear regression.', 'chapters': [{'end': 2975.365, 'start': 2881.274, 'title': 'Steps in machine learning process', 'summary': 'Discusses the general steps in the machine learning process, emphasizing the importance of defining objectives, collecting and preparing data, selecting and training algorithms, testing and running predictions, and deploying the model, while highlighting the domain specificity of these steps.', 'duration': 94.091, 'highlights': ['The importance of defining objectives before collecting data and selecting algorithms, as it is crucial to know what is being predicted (e.g., Photo identification, language, medical, physics).', "The emphasis on the collection and preparation of clean data, recognizing the impact of 'bad data in, bad answer out' and the significance of spending time in data science on this task.", 'The selection and training of algorithms, such as the mention of working with SVM (support vector machine) for training, a crucial step in the machine learning process.', "The testing and validation of the model's accuracy, highlighting the importance of ensuring that the model works effectively for the intended purpose before running predictions.", 'The deployment of the model, with a reminder of the domain-specific nature of the entire process, indicating that certain steps may vary based on the specific domain being addressed.']}, {'end': 3770.114, 'start': 2975.405, 'title': 'Overview of machine learning models', 'summary': 'Provides an overview of machine learning models, including the importance of domain-specific factors, the major divisions of machine learning (classification and regression), anomaly detection, clustering, supervised, unsupervised, and reinforcement learning, and a closer look at linear regression.', 'duration': 794.709, 'highlights': ['The importance of domain-specific factors in machine learning models', 'The major divisions of machine learning: classification and regression', 'The significance of anomaly detection in data science', 'The concept of clustering in machine learning', 'Explanation of supervised, unsupervised, and reinforcement learning', 'Detailed overview of linear regression']}], 'duration': 888.84, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM2881274.jpg', 'highlights': ['The importance of defining objectives before collecting data and selecting algorithms, as it is crucial to know what is being predicted (e.g., Photo identification, language, medical, physics).', "The emphasis on the collection and preparation of clean data, recognizing the impact of 'bad data in, bad answer out' and the significance of spending time in data science on this task.", 'The selection and training of algorithms, such as the mention of working with SVM (support vector machine) for training, a crucial step in the machine learning process.', "The testing and validation of the model's accuracy, highlighting the importance of ensuring that the model works effectively for the intended purpose before running predictions.", 'The deployment of the model, with a reminder of the domain-specific nature of the entire process, indicating that certain steps may vary based on the specific domain being addressed.', 'The major divisions of machine learning: classification and regression', 'The concept of clustering in machine learning', 'The significance of anomaly detection in data science', 'Explanation of supervised, unsupervised, and reinforcement learning', 'Detailed overview of linear regression', 'The importance of domain-specific factors in machine learning models']}, {'end': 5783.107, 'segs': [{'end': 3991.897, 'src': 'heatmap', 'start': 3777.76, 'weight': 0.734, 'content': [{'end': 3781.623, 'text': 'When we take that and we go ahead and plot these points on a graph,', 'start': 3777.76, 'duration': 3.863}, {'end': 3786.287, 'text': "you can see there's kind of a nice scattering and you could probably eyeball a line through the middle of it.", 'start': 3781.623, 'duration': 4.664}, {'end': 3789.53, 'text': "But we're going to calculate that exact line for linear regression.", 'start': 3786.748, 'duration': 2.782}, {'end': 3794.214, 'text': 'And the first thing we do is we come up here and we have the mean of xi.', 'start': 3789.81, 'duration': 4.404}, {'end': 3796.476, 'text': 'And remember, mean is basically the average.', 'start': 3794.494, 'duration': 1.982}, {'end': 3800.64, 'text': 'So we added 5 plus 4 plus 3 plus 2 plus 1 and divide by 5.', 'start': 3796.676, 'duration': 3.964}, {'end': 3803.632, 'text': 'And that simply comes out as 3.', 'start': 3800.64, 'duration': 2.992}, {'end': 3805.372, 'text': "And then we'll do the same for y.", 'start': 3803.632, 'duration': 1.74}, {'end': 3811.894, 'text': "We'll go ahead and add up all those numbers and divide by 5, and we end up with a mean value of y of i equals 2.8,,", 'start': 3805.372, 'duration': 6.522}, {'end': 3819.276, 'text': 'where the xi references as an average or means value and the yi also equals a means value of y.', 'start': 3811.894, 'duration': 7.382}, {'end': 3825.918, 'text': "And when we plot that, you'll see that we can put in the y equals 2.8 and the x equals 3 in there on our graph.", 'start': 3819.276, 'duration': 6.642}, {'end': 3829.719, 'text': 'We kind of gave it a little different color so you could sort it out with the dashed lines on it.', 'start': 3826.038, 'duration': 3.681}, {'end': 3836.626, 'text': "And it's important to note that when we do the linear regression, the linear regression model should go through that dot.", 'start': 3830.299, 'duration': 6.327}, {'end': 3840.15, 'text': "Now let's find our regression equation to find the best fit line.", 'start': 3837.227, 'duration': 2.923}, {'end': 3844.655, 'text': "Remember we go ahead and take our y equals mx plus c, so we're looking for m and c.", 'start': 3840.17, 'duration': 4.485}, {'end': 3850.622, 'text': 'So to find this equation for our data, we need to find our slope of m and our coefficient of c.', 'start': 3844.655, 'duration': 5.967}, {'end': 3864.248, 'text': 'And we have y equals mx plus c, where m equals the sum of x minus x average times, y minus y average or y means and x means over the sum of x minus.', 'start': 3851.363, 'duration': 12.885}, {'end': 3865.688, 'text': 'x means squared.', 'start': 3864.248, 'duration': 1.44}, {'end': 3867.989, 'text': "That's how we get the slope of the value of the line.", 'start': 3865.988, 'duration': 2.001}, {'end': 3870.93, 'text': 'And we can easily do that by creating some columns here.', 'start': 3868.389, 'duration': 2.541}, {'end': 3872.331, 'text': 'We have x, y.', 'start': 3870.99, 'duration': 1.341}, {'end': 3878.553, 'text': 'Computers are really good about iterating through data, and so we can easily compute this and fill in a graph of data.', 'start': 3872.331, 'duration': 6.222}, {'end': 3887.275, 'text': 'And in our graph you can easily see that if we have our x value of 1, and if you remember the xi or the means, value is 3,', 'start': 3879.133, 'duration': 8.142}, {'end': 3894.277, 'text': '1 minus 3 equals a negative 2, and 2 minus 3 equals a negative 1, so on and so forth.', 'start': 3887.275, 'duration': 7.002}, {'end': 3906.644, 'text': 'And we can easily fill in the column of x, minus xi y minus yi. And then from those we can compute x minus xi squared and x minus xi times y minus yi.', 'start': 3894.597, 'duration': 12.047}, {'end': 3911.348, 'text': 'And you can guess it that the next step is to go ahead and sum the different columns for the answers we need.', 'start': 3906.984, 'duration': 4.364}, {'end': 3919.295, 'text': 'So we get a total of 10 for our x minus xi squared and a total of 2 for x minus xi times y minus yi.', 'start': 3911.868, 'duration': 7.427}, {'end': 3923.478, 'text': 'And we plug those in, we get 2 tenths, which equals 0.2.', 'start': 3919.915, 'duration': 3.563}, {'end': 3926.641, 'text': 'So now we know the slope of our line equals 0.2.', 'start': 3923.478, 'duration': 3.163}, {'end': 3928.022, 'text': 'So we can calculate the value of c.', 'start': 3926.641, 'duration': 1.381}, {'end': 3932.071, 'text': "That'd be the next step is we need to know where it crosses the y-axis.", 'start': 3928.769, 'duration': 3.302}, {'end': 3941.057, 'text': 'And if you remember, I mentioned earlier that the linear regression line has to pass through the means value, the one that we showed earlier.', 'start': 3932.592, 'duration': 8.465}, {'end': 3943.179, 'text': 'We can just flip back up there to that graph.', 'start': 3941.338, 'duration': 1.841}, {'end': 3949.937, 'text': "and you can see right here, there's our means value, which is 3, x equals 3 and y equals 2.8.", 'start': 3943.795, 'duration': 6.142}, {'end': 3955.199, 'text': 'and since we know that value, we can simply plug that into our formula.', 'start': 3949.937, 'duration': 5.262}, {'end': 3957.339, 'text': 'y equals 0.2, x plus c.', 'start': 3955.199, 'duration': 2.14}, {'end': 3958.18, 'text': 'so we plug that in.', 'start': 3957.339, 'duration': 0.841}, {'end': 3963.801, 'text': 'we get 2.8, equals 0.2 times 3 plus c, and you can just solve for c.', 'start': 3958.18, 'duration': 5.621}, {'end': 3971.892, 'text': 'so now we know that our coefficient equals 2.2, and once we have all that, We can go ahead and plot our regression line.', 'start': 3963.801, 'duration': 8.091}, {'end': 3974.156, 'text': 'y equals 0.2 times x plus 2.2..', 'start': 3971.892, 'duration': 2.264}, {'end': 3979.083, 'text': 'And then from this equation, we can compute new values.', 'start': 3974.156, 'duration': 4.927}, {'end': 3983.309, 'text': "So let's predict the values of y using x equals 1, 2, 3, 4, 5 and plot the points.", 'start': 3979.323, 'duration': 3.986}, {'end': 3987.934, 'text': 'Remember the 1, 2, 3, 4, 5 was our original x values.', 'start': 3985.111, 'duration': 2.823}, {'end': 3991.897, 'text': "So now we're going to see what y thinks they are, not what they actually are.", 'start': 3988.274, 'duration': 3.623}], 'summary': 'Linear regression calculated a best fit line y = 0.2x + 2.2 through the points, with mean x = 3 and mean y = 2.8.', 'duration': 214.137, 'max_score': 3777.76, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM3777760.jpg'}, {'end': 3957.339, 'src': 'embed', 'start': 3932.592, 'weight': 2, 'content': [{'end': 3941.057, 'text': 'And if you remember, I mentioned earlier that the linear regression line has to pass through the means value, the one that we showed earlier.', 'start': 3932.592, 'duration': 8.465}, {'end': 3943.179, 'text': 'We can just flip back up there to that graph.', 'start': 3941.338, 'duration': 1.841}, {'end': 3949.937, 'text': "and you can see right here, there's our means value, which is 3, x equals 3 and y equals 2.8.", 'start': 3943.795, 'duration': 6.142}, {'end': 3955.199, 'text': 'and since we know that value, we can simply plug that into our formula.', 'start': 3949.937, 'duration': 5.262}, {'end': 3957.339, 'text': 'y equals 0.2, x plus c.', 'start': 3955.199, 'duration': 2.14}], 'summary': 'Linear regression line passes through means value, x=3, y=2.8, y=0.2x+c formula used.', 'duration': 24.747, 'max_score': 3932.592, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM3932592.jpg'}, {'end': 4608.421, 'src': 'embed', 'start': 4582.194, 'weight': 4, 'content': [{'end': 4586.715, 'text': "Now that we've looked at our decision tree, let's look at the third one of our algorithms we're investigating.", 'start': 4582.194, 'duration': 4.521}, {'end': 4588.556, 'text': 'Support Vector Machine.', 'start': 4587.115, 'duration': 1.441}, {'end': 4592.357, 'text': 'Support Vector Machine is a widely used classification algorithm.', 'start': 4589.056, 'duration': 3.301}, {'end': 4594.637, 'text': 'The idea of Support Vector Machine is simple.', 'start': 4592.577, 'duration': 2.06}, {'end': 4599.238, 'text': 'The algorithm creates a separation line which divides the classes in the best possible manner.', 'start': 4594.817, 'duration': 4.421}, {'end': 4602.519, 'text': 'For example, dog or cat, disease or no disease.', 'start': 4599.539, 'duration': 2.98}, {'end': 4608.421, 'text': 'Suppose we have a labeled sample data which tells height and weight of males and females.', 'start': 4603.079, 'duration': 5.342}], 'summary': 'Introducing support vector machine, a widely used classification algorithm creating a separation line for class division.', 'duration': 26.227, 'max_score': 4582.194, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM4582194.jpg'}, {'end': 4831.013, 'src': 'heatmap', 'start': 4615.743, 'weight': 0.713, 'content': [{'end': 4621.424, 'text': 'We draw decision lines, but if we consider decision line one, then we will classify the individual as a male.', 'start': 4615.743, 'duration': 5.681}, {'end': 4625.205, 'text': 'And if we consider decision line two, then it will be a female.', 'start': 4621.885, 'duration': 3.32}, {'end': 4628.166, 'text': 'So you can see this person kind of lies in the middle of the two groups.', 'start': 4625.686, 'duration': 2.48}, {'end': 4630.767, 'text': "It's a little confusing trying to figure out which line they should be under.", 'start': 4628.246, 'duration': 2.521}, {'end': 4634.71, 'text': 'We need to know which line divides the classes correctly, but how?', 'start': 4631.107, 'duration': 3.603}, {'end': 4641.815, 'text': 'The goal is to choose a hyperplane, and that is one of the key words they use when we talk about support vector machines.', 'start': 4635.27, 'duration': 6.545}, {'end': 4648.719, 'text': 'Choose a hyperplane with the greatest possible margin between the decision line and the nearest point within the training set.', 'start': 4642.335, 'duration': 6.384}, {'end': 4656.165, 'text': 'So you can see here we have our support vector, we have the two nearest points to it, and we draw a line between those two points.', 'start': 4649.4, 'duration': 6.765}, {'end': 4662.407, 'text': 'And the distance margin is the distance between the hyperplane and the nearest data point from either set.', 'start': 4656.665, 'duration': 5.742}, {'end': 4669.108, 'text': "So we actually have a value, and it should be equally distant between the two points that we're comparing it to.", 'start': 4662.827, 'duration': 6.281}, {'end': 4673.849, 'text': 'When we draw the hyperplanes we observe that line one has a maximum distance.', 'start': 4669.428, 'duration': 4.421}, {'end': 4677.53, 'text': 'so we observe that line one has a maximum distance margin.', 'start': 4673.849, 'duration': 3.681}, {'end': 4679.67, 'text': "so we'll classify the new data point correctly.", 'start': 4677.53, 'duration': 2.14}, {'end': 4683.431, 'text': 'And our result on this one is going to be that the new data point is mel.', 'start': 4680.11, 'duration': 3.321}, {'end': 4692.077, 'text': "One of the reasons we call it a hyperplane versus a line is that a lot of times we're not looking at just weight and height.", 'start': 4684.071, 'duration': 8.006}, {'end': 4696, 'text': 'We might be looking at 36 different features or dimensions.', 'start': 4692.437, 'duration': 3.563}, {'end': 4703.966, 'text': "And so when we cut it with a hyperplane, it's more of a three-dimensional cut in the data or multi-dimensional that cuts the data a certain way.", 'start': 4696.56, 'duration': 7.406}, {'end': 4708.609, 'text': 'And each plane continues to cut it down until we get the best fit or match.', 'start': 4704.266, 'duration': 4.343}, {'end': 4711.091, 'text': "Let's understand this with the help of an example.", 'start': 4709.149, 'duration': 1.942}, {'end': 4712.132, 'text': 'Problem statement.', 'start': 4711.431, 'duration': 0.701}, {'end': 4714.714, 'text': "You always start with a problem statement when you're going to put some code together.", 'start': 4712.212, 'duration': 2.502}, {'end': 4715.755, 'text': "We're going to do some coding now.", 'start': 4714.734, 'duration': 1.021}, {'end': 4720.398, 'text': 'Classifying muffin and cupcake recipes using support vector machines.', 'start': 4716.075, 'duration': 4.323}, {'end': 4722.68, 'text': 'So the cupcake versus the muffin.', 'start': 4720.718, 'duration': 1.962}, {'end': 4724.602, 'text': "Let's have a look at our data set.", 'start': 4723.16, 'duration': 1.442}, {'end': 4726.943, 'text': 'And we have the different recipes here.', 'start': 4724.962, 'duration': 1.981}, {'end': 4729.886, 'text': 'We have a muffin recipe that has so much flour.', 'start': 4727.003, 'duration': 2.883}, {'end': 4732.328, 'text': "I'm not sure what measurement 55 is in, but it has 55.", 'start': 4729.906, 'duration': 2.422}, {'end': 4733.048, 'text': "Maybe it's ounces? Yeah.", 'start': 4732.328, 'duration': 0.72}, {'end': 4742.131, 'text': 'But it has a certain amount of flour, a certain amount of milk, sugar, butter, egg, baking powder, vanilla, and salt.', 'start': 4735.967, 'duration': 6.164}, {'end': 4747.034, 'text': "And so based on these measurements, we want to guess whether we're making a muffin or a cupcake.", 'start': 4742.612, 'duration': 4.422}, {'end': 4750.357, 'text': "And you can see in this one, we don't have just two features.", 'start': 4747.375, 'duration': 2.982}, {'end': 4754.6, 'text': "We don't just have height and weight as we did before between the male and female.", 'start': 4750.477, 'duration': 4.123}, {'end': 4756.681, 'text': 'In here, we have a number of features.', 'start': 4754.76, 'duration': 1.921}, {'end': 4762.425, 'text': "In fact, in this, we're looking at eight different features to guess whether it's a muffin or a cupcake.", 'start': 4757.341, 'duration': 5.084}, {'end': 4770.543, 'text': "What's the difference between a muffin and a cupcake? Turns out muffins have more flour while cupcakes have more butter and sugar.", 'start': 4763.339, 'duration': 7.204}, {'end': 4775.605, 'text': "So basically the cupcake's a little bit more of a dessert where the muffin's a little bit more of a fancy bread.", 'start': 4770.863, 'duration': 4.742}, {'end': 4777.486, 'text': 'But how do we do that in Python?', 'start': 4776.126, 'duration': 1.36}, {'end': 4778.847, 'text': 'How do we code that?', 'start': 4777.566, 'duration': 1.281}, {'end': 4781.629, 'text': 'to go through recipes and figure out what the recipe is?', 'start': 4778.847, 'duration': 2.782}, {'end': 4788.995, 'text': 'I really just want to say cupcakes versus muffins, like some big professional wrestling thing.', 'start': 4782.97, 'duration': 6.025}, {'end': 4794.259, 'text': 'Before we start in our cupcakes versus muffins, we are going to be working in Python.', 'start': 4789.375, 'duration': 4.884}, {'end': 4797.182, 'text': "There's many versions of Python, many different editors.", 'start': 4794.519, 'duration': 2.663}, {'end': 4802.606, 'text': 'That is one of the strengths and weaknesses of Python is it just has so much stuff attached to it.', 'start': 4797.402, 'duration': 5.204}, {'end': 4807.17, 'text': "It's one of the more popular data science programming packages you can use.", 'start': 4802.666, 'duration': 4.504}, {'end': 4812.414, 'text': "In this case, we're going to go ahead and use Anaconda in Jupyter Notebook.", 'start': 4807.81, 'duration': 4.604}, {'end': 4816.482, 'text': 'The Anaconda Navigator has all kinds of fun tools.', 'start': 4812.799, 'duration': 3.683}, {'end': 4820.565, 'text': "Once you're into the Anaconda Navigator, you can change environments.", 'start': 4816.902, 'duration': 3.663}, {'end': 4823.087, 'text': 'I actually have a number of environments on here.', 'start': 4820.585, 'duration': 2.502}, {'end': 4825.909, 'text': "We'll be using Python 3.6 environment.", 'start': 4823.307, 'duration': 2.602}, {'end': 4831.013, 'text': "So this is in Python version 3.6, although it doesn't matter too much which version you use.", 'start': 4826.089, 'duration': 4.924}], 'summary': 'Using support vector machines to classify muffin and cupcake recipes based on multiple features and dimensions.', 'duration': 215.27, 'max_score': 4615.743, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM4615743.jpg'}, {'end': 4692.077, 'src': 'embed', 'start': 4662.827, 'weight': 5, 'content': [{'end': 4669.108, 'text': "So we actually have a value, and it should be equally distant between the two points that we're comparing it to.", 'start': 4662.827, 'duration': 6.281}, {'end': 4673.849, 'text': 'When we draw the hyperplanes we observe that line one has a maximum distance.', 'start': 4669.428, 'duration': 4.421}, {'end': 4677.53, 'text': 'so we observe that line one has a maximum distance margin.', 'start': 4673.849, 'duration': 3.681}, {'end': 4679.67, 'text': "so we'll classify the new data point correctly.", 'start': 4677.53, 'duration': 2.14}, {'end': 4683.431, 'text': 'And our result on this one is going to be that the new data point is mel.', 'start': 4680.11, 'duration': 3.321}, {'end': 4692.077, 'text': "One of the reasons we call it a hyperplane versus a line is that a lot of times we're not looking at just weight and height.", 'start': 4684.071, 'duration': 8.006}], 'summary': "Using a hyperplane to classify data points, achieving a maximum distance margin, resulting in correct classification of a new data point as 'mel'.", 'duration': 29.25, 'max_score': 4662.827, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM4662827.jpg'}, {'end': 4947.553, 'src': 'embed', 'start': 4919.83, 'weight': 0, 'content': [{'end': 4922.773, 'text': "And then finally, we're working with the support vector machine.", 'start': 4919.83, 'duration': 2.943}, {'end': 4931.42, 'text': "So from SK learn, we're going to use the SK learn model, import SVM support vector machine.", 'start': 4923.073, 'duration': 8.347}, {'end': 4933.722, 'text': 'And then.', 'start': 4933.522, 'duration': 0.2}, {'end': 4938.945, 'text': 'As a data scientist, you should always try to visualize your data.', 'start': 4934.721, 'duration': 4.224}, {'end': 4943.009, 'text': "Some data obviously is too complicated or doesn't make any sense to the human.", 'start': 4939.325, 'duration': 3.684}, {'end': 4947.553, 'text': "But if it's possible, it's good to take a second look at it so that you can actually see what you're doing.", 'start': 4943.329, 'duration': 4.224}], 'summary': 'Using support vector machine model from sk learn for data visualization and analysis.', 'duration': 27.723, 'max_score': 4919.83, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM4919830.jpg'}, {'end': 5586.267, 'src': 'embed', 'start': 5558.055, 'weight': 1, 'content': [{'end': 5561.098, 'text': "We're using a package called SVC.", 'start': 5558.055, 'duration': 3.043}, {'end': 5566.842, 'text': "In this case, we're going to go ahead and set the kernel equals linear.", 'start': 5563.079, 'duration': 3.763}, {'end': 5568.944, 'text': "So it's using a specific setup on there.", 'start': 5566.902, 'duration': 2.042}, {'end': 5576.428, 'text': "And if we go to the reference on their website for the SVM, You'll see that there's eight of them here.", 'start': 5569.344, 'duration': 7.084}, {'end': 5578.552, 'text': 'Three of them are for regression.', 'start': 5577.049, 'duration': 1.503}, {'end': 5581.237, 'text': 'Three are for classification.', 'start': 5579.594, 'duration': 1.643}, {'end': 5586.267, 'text': 'The SVC, support vector classification, is probably one of the most commonly used.', 'start': 5581.638, 'duration': 4.629}], 'summary': 'Using svc package with linear kernel for support vector classification, one of the most commonly used methods with 8 types, 3 for regression and 3 for classification.', 'duration': 28.212, 'max_score': 5558.055, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM5558055.jpg'}], 'start': 3770.114, 'title': 'Machine learning fundamentals', 'summary': 'Covers fundamental concepts of linear regression, basics of decision trees, entropy, and information gain, using entropy in decision making, using numpy, pandas for data analysis, and training an svm model for classifying cupcake vs muffin recipes.', 'chapters': [{'end': 4065.909, 'start': 3770.114, 'title': 'Linear regression basics', 'summary': 'Discusses the fundamental concepts of linear regression, including calculating the mean, finding the regression equation, computing predicted values, and minimizing error in the model.', 'duration': 295.795, 'highlights': ['The linear regression model should go through the mean values of x and y when plotted on a graph, with the regression equation being y = mx + c.', 'Calculation of the regression equation involves finding the slope (m) and the coefficient (c), where the slope is determined using the formula m = Σ(x - x̄)(y - ȳ) / Σ(x - x̄)².', 'The process of minimizing error in linear regression involves various methods such as sum of squared errors, sum of absolute errors, and root mean square error.', 'The chapter emphasizes the importance of understanding and minimizing error in linear regression models, aiming to reduce the distance between the line and the data points.']}, {'end': 4384.718, 'start': 4066.289, 'title': 'Understanding decision trees and entropy', 'summary': 'Discusses the basics of decision trees and how entropy and information gain are used to make decisions, emphasizing the importance of low entropy and high information gain. it also provides a practical example of computing entropy for a set of data.', 'duration': 318.429, 'highlights': ['Decision trees and how entropy and information gain are used to make decisions', 'Importance of low entropy and high information gain', 'Practical example of computing entropy for a set of data']}, {'end': 4847.365, 'start': 4385.219, 'title': 'Decision making with entropy and support vector machines', 'summary': 'Delves into using entropy to make decisions in a decision tree, with examples of calculating entropy for different predictors and information gains. it also explains the concept of support vector machine and its application in classifying data with the example of muffin and cupcake recipes, highlighting the process of creating a separation line to divide classes.', 'duration': 462.146, 'highlights': ['Decision tree: Entropy calculation for different predictors and information gain', 'Decision tree: Building the decision tree based on information gain', 'Support Vector Machine: Creating a separation line for classifying data']}, {'end': 5062.734, 'start': 4847.585, 'title': 'Data analysis with numpy, pandas, and support vector machine (svm)', 'summary': 'Discusses the use of numpy and pandas for data manipulation, visualization using matplotlib and seaborn, and the importance of importing necessary packages for data analysis for supporting the support vector machine model from sk learn.', 'duration': 215.149, 'highlights': ['The importance of using NumPy and Pandas for data manipulation and creating data frames for easy referencing and additional features.', 'The significance of visualizing data using Matplotlib and Seaborn for understanding and analyzing the data effectively.', 'The necessity of importing necessary packages such as NumPy, Pandas, Matplotlib, Seaborn, and SK learn for supporting the support vector machine model.']}, {'end': 5368.108, 'start': 5063.015, 'title': 'Analyzing cupcakes vs muffins data', 'summary': "Discusses opening and analyzing a csv file containing data on cupcakes vs muffins, using python's pandas and seaborn libraries to plot and process the data, and preparing the data for a machine learning model, showcasing the power and efficiency of these tools in data analysis.", 'duration': 305.093, 'highlights': ['The chapter discusses opening and analyzing a CSV file containing data on cupcakes vs muffins.', "Using Python's pandas and seaborn libraries to plot and process the data.", 'Preparing the data for a machine learning model.']}, {'end': 5783.107, 'start': 5368.289, 'title': 'Creating recipe features and training svm model', 'summary': 'Explains how to create recipe features and train an svm model using flour and sugar ingredients, with a focus on support vector classification and the kernel setting, while also delving into the coefficients for the separating plane.', 'duration': 414.818, 'highlights': ['Explaining Support Vector Classification and Kernel Setting', 'Creating Recipe Features and Ingredients', 'Training the SVM Model', 'Understanding Coefficients and Separating Plane']}], 'duration': 2012.993, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM3770114.jpg', 'highlights': ['The linear regression model involves finding the slope and coefficient for minimizing error.', 'Decision trees use entropy and information gain for decision making.', 'Importance of using NumPy and Pandas for data manipulation and visualization.', "Using Python's pandas and seaborn libraries to plot and process data.", 'Preparing the data for a machine learning model.', 'Training the SVM model for classifying cupcake vs muffin recipes.']}, {'end': 6561.933, 'segs': [{'end': 5891.77, 'src': 'embed', 'start': 5860.447, 'weight': 1, 'content': [{'end': 5863.291, 'text': "Here's our YY, which we now know is a set of data.", 'start': 5860.447, 'duration': 2.844}, {'end': 5870.379, 'text': "We're going to create YY down equals A times XX plus B1 minus A times B0.", 'start': 5863.571, 'duration': 6.808}, {'end': 5877.158, 'text': 'And then model support vector b is going to be set that to a new value, the minus 1 set up.', 'start': 5871.934, 'duration': 5.224}, {'end': 5883.783, 'text': 'And yy up equals a times xx plus b1 minus a times b0.', 'start': 5877.659, 'duration': 6.124}, {'end': 5887.226, 'text': 'And we can go ahead and just run this to load these variables up.', 'start': 5884.784, 'duration': 2.442}, {'end': 5891.77, 'text': "If you wanted to understand a little bit more of what's going on, you can see if we print.", 'start': 5887.746, 'duration': 4.024}], 'summary': 'Creating a model to calculate yy using specific formulas and setting support vector b to a new value.', 'duration': 31.323, 'max_score': 5860.447, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM5860447.jpg'}, {'end': 6054.122, 'src': 'embed', 'start': 6021.363, 'weight': 4, 'content': [{'end': 6022.464, 'text': 'And this is what you expect.', 'start': 6021.363, 'duration': 1.101}, {'end': 6025.726, 'text': 'You expect these two lines to go through the nearest data point.', 'start': 6022.684, 'duration': 3.042}, {'end': 6030.731, 'text': "So the dashed lines go through the nearest muffin and the nearest cupcake when it's plotting it.", 'start': 6025.847, 'duration': 4.884}, {'end': 6033.133, 'text': 'And then your SVM goes right down the middle.', 'start': 6030.991, 'duration': 2.142}, {'end': 6034.954, 'text': 'So it gives it a nice split in our data.', 'start': 6033.213, 'duration': 1.741}, {'end': 6041.399, 'text': "And you can see how easy it is to see, based just on sugar and flour, which one's a muffin or a cupcake.", 'start': 6035.274, 'duration': 6.125}, {'end': 6054.122, 'text': "Let's go ahead and create a function to predict muffin or cupcake.", 'start': 6043.261, 'duration': 10.861}], 'summary': 'Svm creates a clear split between muffins and cupcakes based on sugar and flour, making it easy to predict using a function.', 'duration': 32.759, 'max_score': 6021.363, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM6021363.jpg'}, {'end': 6095.593, 'src': 'embed', 'start': 6070.845, 'weight': 2, 'content': [{'end': 6076.406, 'text': "and remember, we're just doing flour and sugar today, not doing all the ingredients, and that actually is a pretty good split.", 'start': 6070.845, 'duration': 5.561}, {'end': 6077.606, 'text': "you really don't need all the ingredients.", 'start': 6076.406, 'duration': 1.2}, {'end': 6082.329, 'text': "you know it's flour and sugar, And let's go ahead and do an if-else statement.", 'start': 6077.606, 'duration': 4.723}, {'end': 6093.012, 'text': 'So if model predict is of flour and sugar equals zero, so we take our model and we run a predict.', 'start': 6082.629, 'duration': 10.383}, {'end': 6095.593, 'text': "It's very common in sklearn where you have a dot predict.", 'start': 6093.092, 'duration': 2.501}], 'summary': 'Focusing on flour and sugar today, if-else statement used in sklearn with a good split.', 'duration': 24.748, 'max_score': 6070.845, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM6070845.jpg'}, {'end': 6411.844, 'src': 'embed', 'start': 6386.782, 'weight': 5, 'content': [{'end': 6392.105, 'text': 'Now with books, we can easily see fiction and horror and history books.', 'start': 6386.782, 'duration': 5.323}, {'end': 6397.527, 'text': "But a lot of times with data, some of that information isn't so easy to see right when we first look at it.", 'start': 6392.505, 'duration': 5.022}, {'end': 6402.45, 'text': 'And so k-means is one of those tools where we can start finding things that connect, that match with each other.', 'start': 6397.688, 'duration': 4.762}, {'end': 6406.692, 'text': 'Suppose we have these data points and want to assign them into a cluster.', 'start': 6402.93, 'duration': 3.762}, {'end': 6411.844, 'text': 'Now when I look at these data points, I would probably group them into two clusters just by looking at them.', 'start': 6407.199, 'duration': 4.645}], 'summary': 'K-means helps find connections in data for clustering, like grouping data points into two clusters.', 'duration': 25.062, 'max_score': 6386.782, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM6386782.jpg'}, {'end': 6506.215, 'src': 'embed', 'start': 6481.363, 'weight': 0, 'content': [{'end': 6488.812, 'text': 'We assign random centroids to clusters and sometimes you pick the centroids because you might look at the data in a graph and say oh,', 'start': 6481.363, 'duration': 7.449}, {'end': 6490.313, 'text': 'these are probably the central points.', 'start': 6488.812, 'duration': 1.501}, {'end': 6494.598, 'text': 'Then we compute the distance from the objects to the centroids.', 'start': 6490.834, 'duration': 3.764}, {'end': 6500.045, 'text': 'We take that and we form new clusters based on minimum distance and calculate their centroids.', 'start': 6495.339, 'duration': 4.706}, {'end': 6503.814, 'text': 'Then we compute the distance from objects to the new centroids.', 'start': 6500.592, 'duration': 3.222}, {'end': 6506.215, 'text': 'And then we go back and repeat those last two steps.', 'start': 6504.254, 'duration': 1.961}], 'summary': 'Process involves assigning centroids, computing distances, and forming clusters iteratively.', 'duration': 24.852, 'max_score': 6481.363, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM6481363.jpg'}], 'start': 5783.347, 'title': 'Classifying recipes and k-means clustering', 'summary': 'Covers the implementation of a support vector machine to create a classifier for predicting muffin or cupcake recipes using flour and sugar, and introduces upcoming topics on k-means clustering and logistic regression. it also explains k-means clustering as a commonly used tool in unsupervised learning for organizing unlabeled data into clusters based on feature similarities.', 'chapters': [{'end': 6330.087, 'start': 5783.347, 'title': 'Support vector machine and classifying recipes', 'summary': 'Explains the implementation of a support vector machine to create a classifier for predicting muffin or cupcake recipes using flour and sugar, and also introduces the upcoming k-means clustering and logistic regression topics, providing a comprehensive understanding of the machine learning concepts.', 'duration': 546.74, 'highlights': ['The chapter extensively covers the implementation of a support vector machine to classify muffin or cupcake recipes using flour and sugar, demonstrating the process of creating a classifier and predicting the recipe type, with a clear explanation of the code structure and functionality.', 'The upcoming topics of k-means clustering and logistic regression are introduced, offering a glimpse into the upcoming machine learning concepts and practical demonstrations, providing a comprehensive learning experience for the audience.', 'The chapter includes practical Python code demonstrations for clustering cars based on brands and classifying tumors as malignant or benign, showcasing real-world applications of machine learning techniques and their impact on data analysis and decision making.']}, {'end': 6561.933, 'start': 6330.107, 'title': 'Understanding k-means clustering', 'summary': 'Introduces k-means clustering as a commonly used tool in unsupervised learning for organizing unlabeled data into clusters based on feature similarities, and explores the iterative process of assigning centroids, calculating distances, and converging to static clusters.', 'duration': 231.826, 'highlights': ['K-means clustering is a commonly used tool in unsupervised learning for organizing unlabeled data into clusters based on feature similarities.', 'The iterative process of assigning centroids, calculating distances, and converging to static clusters involves repeatedly picking k clusters, assigning random centroids, computing distances from objects to centroids, forming new clusters based on minimum distances, and recalculating centroids until convergence.', 'The process of selecting initial cluster centroids involves choosing either random points or the farthest apart points, depending on the data and knowledge about it.']}], 'duration': 778.586, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM5783347.jpg', 'highlights': ['The chapter extensively covers the implementation of a support vector machine to classify muffin or cupcake recipes using flour and sugar, demonstrating the process of creating a classifier and predicting the recipe type, with a clear explanation of the code structure and functionality.', 'The upcoming topics of k-means clustering and logistic regression are introduced, offering a glimpse into the upcoming machine learning concepts and practical demonstrations, providing a comprehensive learning experience for the audience.', 'The chapter includes practical Python code demonstrations for clustering cars based on brands and classifying tumors as malignant or benign, showcasing real-world applications of machine learning techniques and their impact on data analysis and decision making.', 'K-means clustering is a commonly used tool in unsupervised learning for organizing unlabeled data into clusters based on feature similarities.', 'The iterative process of assigning centroids, calculating distances, and converging to static clusters involves repeatedly picking k clusters, assigning random centroids, computing distances from objects to centroids, forming new clusters based on minimum distances, and recalculating centroids until convergence.', 'The process of selecting initial cluster centroids involves choosing either random points or the farthest apart points, depending on the data and knowledge about it.']}, {'end': 10188.413, 'segs': [{'end': 6604.006, 'src': 'embed', 'start': 6579.394, 'weight': 5, 'content': [{'end': 6587.436, 'text': 'we measure that distance and you can see that if we measure each of those distances and you use the Pythagorean theorem for a triangle in this case,', 'start': 6579.394, 'duration': 8.042}, {'end': 6593.998, 'text': 'because you know the x and the y and you can figure out the diagonal line from that or you can just take a ruler and put it on your monitor.', 'start': 6587.436, 'duration': 6.562}, {'end': 6597.099, 'text': "That'd be kind of silly, but it would work if you're just eyeballing it.", 'start': 6594.198, 'duration': 2.901}, {'end': 6600.42, 'text': 'You can see how they naturally come together in certain areas.', 'start': 6597.299, 'duration': 3.121}, {'end': 6604.006, 'text': 'Now we again calculate the centroids of each cluster.', 'start': 6601.024, 'duration': 2.982}], 'summary': 'Measuring distances using pythagorean theorem and calculating centroids of clusters.', 'duration': 24.612, 'max_score': 6579.394, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM6579394.jpg'}, {'end': 6808.499, 'src': 'embed', 'start': 6785.862, 'weight': 2, 'content': [{'end': 6794.489, 'text': 'using k-means, clustering to cluster cars into brands, using parameters such as horsepower, cubic inches, make, year, etc.', 'start': 6785.862, 'duration': 8.627}, {'end': 6801.914, 'text': "so we're going to use the dataset cars data having information about three brands of cars toyota, honda and nissan.", 'start': 6794.489, 'duration': 7.425}, {'end': 6808.499, 'text': "we'll go back to my favorite tool, the anaconda navigator with the jupiter notebook,", 'start': 6801.914, 'duration': 6.585}], 'summary': 'Using k-means to cluster cars into brands based on parameters like horsepower and cubic inches.', 'duration': 22.637, 'max_score': 6785.862, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM6785862.jpg'}, {'end': 7584.176, 'src': 'embed', 'start': 7562.622, 'weight': 0, 'content': [{'end': 7572.048, 'text': "And you can already see the problem that if I'm going to iterate through a terabyte of data 11 times and then the k-means itself is iterating through the data multiple times,", 'start': 7562.622, 'duration': 9.426}, {'end': 7573.349, 'text': "that's a heck of a process.", 'start': 7572.048, 'duration': 1.301}, {'end': 7575.69, 'text': "So you've got to be a little careful with this.", 'start': 7573.369, 'duration': 2.321}, {'end': 7579.753, 'text': 'A lot of times, though, you can find your ELBO using the ELBO method.', 'start': 7575.71, 'duration': 4.043}, {'end': 7584.176, 'text': "find your optimal number on a sample of data, especially if you're working with larger data sources.", 'start': 7579.753, 'duration': 4.423}], 'summary': 'Iterating through a terabyte of data 11 times and then k-means iterating through the data multiple times can be a time-consuming process. finding the optimal number on a sample of data can be a more efficient approach, particularly for larger data sources.', 'duration': 21.554, 'max_score': 7562.622, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM7562622.jpg'}, {'end': 7870.926, 'src': 'embed', 'start': 7838.686, 'weight': 6, 'content': [{'end': 7842.649, 'text': "And then let's go ahead and print x and see what we have for x.", 'start': 7838.686, 'duration': 3.963}, {'end': 7848.252, 'text': "And we'll see that x is an array, it's a matrix, so we have our different values in the array.", 'start': 7842.649, 'duration': 5.603}, {'end': 7853.136, 'text': "And what we're going to do, it's very hard to plot all the different values in the array.", 'start': 7848.272, 'duration': 4.864}, {'end': 7859.52, 'text': "So we're only going to be looking at the first two, or positions 0 and 1.", 'start': 7853.376, 'duration': 6.144}, {'end': 7864.463, 'text': 'And if you were doing a full presentation in front of the board meeting,', 'start': 7859.52, 'duration': 4.943}, {'end': 7870.926, 'text': 'you might actually do a little different and dig a little deeper into the different aspects, because this is all the different columns we looked at.', 'start': 7864.463, 'duration': 6.463}], 'summary': 'Analyzing x, an array with matrix values, focusing on positions 0 and 1.', 'duration': 32.24, 'max_score': 7838.686, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM7838686.jpg'}, {'end': 8058.668, 'src': 'embed', 'start': 8029.408, 'weight': 8, 'content': [{'end': 8036.773, 'text': "it'll automatically bring in it, since we've already labeled the different aspects of the legend with toyota, nissan and honda.", 'start': 8029.408, 'duration': 7.365}, {'end': 8041.977, 'text': "and finally we want to go ahead and show so we can actually see it and remember it's in line.", 'start': 8036.773, 'duration': 5.204}, {'end': 8049.222, 'text': "so if you're using a different editor that's not the jupiter notebook you'll get a pop-up of this and you should have a nice set of clusters here.", 'start': 8041.977, 'duration': 7.245}, {'end': 8050.042, 'text': 'so we can look at this.', 'start': 8049.222, 'duration': 0.82}, {'end': 8058.668, 'text': 'we have a clusters of honda and green, toyota and red, nissan and purple, and you can see where they put the centroids to separate them.', 'start': 8050.042, 'duration': 8.626}], 'summary': 'Visualizing clusters of car brands with colors and centroids.', 'duration': 29.26, 'max_score': 8029.408, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM8029408.jpg'}, {'end': 8351.752, 'src': 'embed', 'start': 8327.068, 'weight': 4, 'content': [{'end': 8332.571, 'text': "so why would you want to do that if you know you're just going to go get a biopsy because you know it's that serious.", 'start': 8327.068, 'duration': 5.503}, {'end': 8335.554, 'text': 'this is like an all or nothing, Just referencing the domain.', 'start': 8332.571, 'duration': 2.983}, {'end': 8336.193, 'text': "it's important.", 'start': 8335.554, 'duration': 0.639}, {'end': 8343.222, 'text': 'It might help the doctor know where to look just by understanding what kind of tumor it is.', 'start': 8336.495, 'duration': 6.727}, {'end': 8346.846, 'text': 'So it might help them or aid them in something they missed from before.', 'start': 8343.583, 'duration': 3.263}, {'end': 8351.752, 'text': "So let's go ahead and dive into the code and I'll come back to the domain part of it in just a minute.", 'start': 8347.387, 'duration': 4.365}], 'summary': 'Understanding tumor type may aid in targeted treatment.', 'duration': 24.684, 'max_score': 8327.068, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM8327067.jpg'}, {'end': 8725.993, 'src': 'embed', 'start': 8697.903, 'weight': 9, 'content': [{'end': 8703.984, 'text': "So when you get an answer or something like that, or you start looking at some of these individual pieces, you might go, hey, that doesn't match.", 'start': 8697.903, 'duration': 6.081}, {'end': 8708.925, 'text': 'According to showing our heat map, this should not correlate with each other.', 'start': 8704.044, 'duration': 4.881}, {'end': 8711.165, 'text': "And if it is, you're going to have to start asking well, why?", 'start': 8708.965, 'duration': 2.2}, {'end': 8711.905, 'text': "What's going on?", 'start': 8711.225, 'duration': 0.68}, {'end': 8713.365, 'text': 'What else is coming in there?', 'start': 8711.945, 'duration': 1.42}, {'end': 8716.526, 'text': 'But it does show some really cool information on here.', 'start': 8714.166, 'duration': 2.36}, {'end': 8725.993, 'text': "And we can see from the ID, there's no real one feature that just says, if you go across the top line, that lights up.", 'start': 8717.026, 'duration': 8.967}], 'summary': 'Analyzing data for correlations, identifying anomalies, and finding insights.', 'duration': 28.09, 'max_score': 8697.903, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM8697903.jpg'}, {'end': 9203.6, 'src': 'embed', 'start': 9175.852, 'weight': 3, 'content': [{'end': 9180.754, 'text': "So once we've run this, we'll have a model that fits this data, that 70% of our training data.", 'start': 9175.852, 'duration': 4.902}, {'end': 9184.196, 'text': 'And, of course, it prints this out.', 'start': 9183.135, 'duration': 1.061}, {'end': 9187.297, 'text': 'It tells us all the different variables that you can set on there.', 'start': 9184.296, 'duration': 3.001}, {'end': 9189.178, 'text': "There's a lot of different choices you can make.", 'start': 9187.317, 'duration': 1.861}, {'end': 9191.717, 'text': "But for what we're doing, we're just going to let all the defaults sit.", 'start': 9189.537, 'duration': 2.18}, {'end': 9194.578, 'text': "We don't really need to mess with those on this particular example.", 'start': 9192.018, 'duration': 2.56}, {'end': 9200.679, 'text': "And there's nothing in here that really stands out as super important until you start fine-tuning it.", 'start': 9194.818, 'duration': 5.861}, {'end': 9203.6, 'text': "But for what we're doing, the basics will work just fine.", 'start': 9200.899, 'duration': 2.701}], 'summary': 'Model fits 70% of training data with default settings, fine-tuning possible later.', 'duration': 27.748, 'max_score': 9175.852, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM9175852.jpg'}, {'end': 9802.121, 'src': 'embed', 'start': 9770.874, 'weight': 1, 'content': [{'end': 9775.96, 'text': 'Fake accounts, where these accounts only last for how long the transaction takes place and stop existing after that.', 'start': 9770.874, 'duration': 5.086}, {'end': 9779.885, 'text': 'And man in the middle attacks, where they steal your money while the transaction is taking place.', 'start': 9776.241, 'duration': 3.644}, {'end': 9784.59, 'text': 'The feed-forward neural network helps determine whether a transaction is genuine or fraudulent.', 'start': 9780.105, 'duration': 4.485}, {'end': 9792.555, 'text': 'So what happens with feed forward neural networks are that the outputs are converted into hash values and these values become the inputs for the next round.', 'start': 9784.75, 'duration': 7.805}, {'end': 9795.997, 'text': "So for every real transaction that takes place, there's a specific pattern.", 'start': 9792.735, 'duration': 3.262}, {'end': 9800.94, 'text': 'A fraudulent transaction would stand out because of the significant changes that it would cause with the hash values.', 'start': 9796.157, 'duration': 4.783}, {'end': 9802.121, 'text': 'Stock market trading.', 'start': 9801.2, 'duration': 0.921}], 'summary': 'Feed-forward neural network detects fraudulent transactions by analyzing hash value changes.', 'duration': 31.247, 'max_score': 9770.874, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM9770874.jpg'}], 'start': 6562.653, 'title': 'Machine learning applications', 'summary': 'Covers k-means clustering, data cleaning, logistic regression, heat maps, and model testing, focusing on applications like tumor prediction achieving 92% precision and various machine learning domains such as traffic prediction, social media personalization, and stock market trading.', 'chapters': [{'end': 7247.397, 'start': 6562.653, 'title': 'K-means clustering: understanding the process', 'summary': 'Explains the process of k-means clustering and its application in determining the optimal number of clusters using the elbo method, with a use case of clustering cars into brands based on parameters such as horsepower, cubic inches, make, year, etc.', 'duration': 684.744, 'highlights': ['The chapter provides a detailed explanation of the K-means clustering process, including the initial selection of cluster centroids, assignment of points to clusters based on proximity to centroids, calculation of new centroids, and comparison of distances to determine cluster reassignment.', 'The application of the ELBO method to find the optimal number of clusters is illustrated, with a focus on using the slope of the ELBO curve to determine the number of clusters, leading to the identification of the appropriate value for k.', 'A specific use case of using K-means clustering to group cars into brands based on parameters such as horsepower, cubic inches, make, year, etc., is presented, showcasing the practical application of the clustering technique.']}, {'end': 7909.889, 'start': 7248.917, 'title': 'Data cleaning and k-means clustering', 'summary': 'Discusses data cleaning using pandas to fill missing values and ensure all data is numeric, followed by a detailed explanation of applying the k-means clustering method from the sklearn library to find the optimal number of clusters for the dataset, visually represented using the elbow method, and then applying the k-means clustering to the dataset to group the data into clusters based on the chosen number of clusters.', 'duration': 660.972, 'highlights': ['Explaining how to fill missing data and ensure all data is numeric using Pandas', 'Applying the k-means clustering method to find the optimal number of clusters using the sklearn library', 'Demonstrating the application of k-means clustering to the dataset to group the data into clusters']}, {'end': 8612.532, 'start': 7909.949, 'title': 'Clustering cars and logistic regression', 'summary': 'Discusses clustering cars based on attributes, using k-means to create distinct clusters and logistic regression for categorizing data, and its application in classifying tumors, along with data exploration using seaborn and matplotlib.', 'duration': 702.583, 'highlights': ['The chapter discusses clustering cars based on attributes, using k-means to create distinct clusters and logistic regression for categorizing data', 'The application of logistic regression in classifying tumors is explained', 'Data exploration using Seaborn and Matplotlib is demonstrated']}, {'end': 9342.428, 'start': 8612.913, 'title': 'Exploring data with heat maps and model testing', 'summary': "Explores the use of heat maps in data analysis, showing how the data correlates and identifies the need for a comprehensive approach in solving for malignant or benign diagnosis. it then delves into the process of model testing using logistic regression, demonstrating the model's functionality and evaluating its performance through a classification report.", 'duration': 729.515, 'highlights': ['The chapter explores the use of heat maps in data analysis, showing how the data correlates and identifies the need for a comprehensive approach in solving for malignant or benign diagnosis.', "The chapter delves into the process of model testing using logistic regression, demonstrating the model's functionality and evaluating its performance through a classification report."]}, {'end': 9611.218, 'start': 9345.091, 'title': 'Machine learning applications & performance metrics', 'summary': 'Discusses the performance metrics of a tumor prediction model achieving 92% precision and the applications of machine learning, including virtual personal assistants and classification examples.', 'duration': 266.127, 'highlights': ['The tumor prediction model achieves 92% precision in predicting the type of tumor, emphasizing its significance in the medical domain with catastrophic outcomes.', 'Machine learning has various applications, including virtual personal assistants like Google Assistant, Alexa, Cortana, and Siri, which enhance daily tasks such as making calls, playing music, and scheduling appointments.', 'Examples of classification are provided, such as identifying handwritten digits in images and grouping documents into different categories, along with the mention of relevant models like k-means and SVM.', 'The concept of anomaly detection is explained, showcasing its relevance in understanding abnormal behaviors in contexts like website functionality and stock market fluctuations.', "Regression modeling is exemplified in predicting an individual's salary based on their years of experience, highlighting the mathematical relationship between independent and dependent variables."]}, {'end': 10188.413, 'start': 9611.218, 'title': 'Applications of machine learning', 'summary': 'Explores the application of machine learning in various domains including traffic prediction, social media personalization, email spam filtering, online fraud detection, stock market trading, assistive medical technology, and automatic translation, emphasizing the role of neural networks and their impact on different sectors.', 'duration': 577.195, 'highlights': ["Machine learning's role in traffic prediction", 'Social media personalization using machine learning', 'Email spam filtering with machine learning', 'Online fraud detection through feed-forward neural networks', "Machine learning's impact on stock market trading", 'Implications of machine learning in assistive medical technology', 'Automatic translation using deep learning and neural networks']}], 'duration': 3625.76, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM6562653.jpg', 'highlights': ['Tumor prediction achieves 92% precision', 'K-means clustering for car grouping', 'Machine learning applications in traffic prediction, social media personalization, and stock market trading', 'Logistic regression for tumor classification', 'Heat maps for data correlation analysis', 'Anomaly detection in website functionality and stock market fluctuations', 'Regression modeling for salary prediction', 'Application of ELBO method for determining optimal number of clusters', 'Data cleaning and numeric conversion using Pandas', 'Model testing using logistic regression']}, {'end': 12397.39, 'segs': [{'end': 10266.682, 'src': 'embed', 'start': 10228.488, 'weight': 7, 'content': [{'end': 10235.251, 'text': 'So just like the bias is your y-intercept in Euclidean geometry, you could look at the one weight.', 'start': 10228.488, 'duration': 6.763}, {'end': 10237.752, 'text': "Remember, this is very complicated, so we're not looking at just one weight.", 'start': 10235.271, 'duration': 2.481}, {'end': 10240.093, 'text': 'You can look at the weight as your slope of the line.', 'start': 10237.772, 'duration': 2.321}, {'end': 10245.055, 'text': "Or if you're doing x equals my plus c, it would be the m value.", 'start': 10240.373, 'duration': 4.682}, {'end': 10248.676, 'text': 'Neurons of each layer transmit information to neurons of the next layer.', 'start': 10245.235, 'duration': 3.441}, {'end': 10252.218, 'text': 'And you can see here as they light up going across into the final layer.', 'start': 10248.797, 'duration': 3.421}, {'end': 10253.718, 'text': 'and then to the output.', 'start': 10252.618, 'duration': 1.1}, {'end': 10260.32, 'text': 'And in this case, the output is going to be either a square in this one, or it might light up the other one, which is a circle.', 'start': 10254.058, 'duration': 6.262}, {'end': 10262.841, 'text': 'The output layer emits a predicted output.', 'start': 10260.44, 'duration': 2.401}, {'end': 10265.742, 'text': "So in this case, we're looking at a classification.", 'start': 10263.141, 'duration': 2.601}, {'end': 10266.682, 'text': 'True, false.', 'start': 10266.122, 'duration': 0.56}], 'summary': 'Neural network weights represent slope and bias, transmitting information to predict outputs in a classification scenario.', 'duration': 38.194, 'max_score': 10228.488, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM10228488.jpg'}, {'end': 10526.711, 'src': 'embed', 'start': 10500.572, 'weight': 4, 'content': [{'end': 10505.456, 'text': 'Rectifier is a pretty common one, although I see the sigmoid function used to be the basic one.', 'start': 10500.572, 'duration': 4.884}, {'end': 10506.117, 'text': "But it's up there.", 'start': 10505.556, 'duration': 0.561}, {'end': 10508.199, 'text': 'The rectifier function is very commonly used.', 'start': 10506.177, 'duration': 2.022}, {'end': 10511.761, 'text': 'because the output of x if x is positive and 0 otherwise.', 'start': 10508.339, 'duration': 3.422}, {'end': 10517.545, 'text': "and you can see here again, just like it's either kind of get a value going up there.", 'start': 10511.761, 'duration': 5.784}, {'end': 10519.726, 'text': "so max of x of 0, so it's.", 'start': 10517.545, 'duration': 2.181}, {'end': 10520.146, 'text': "it's again.", 'start': 10519.726, 'duration': 0.42}, {'end': 10521.447, 'text': "it's like the threshold function.", 'start': 10520.146, 'duration': 1.301}, {'end': 10526.711, 'text': "yes, no, true, false, it's either 0 or it's some kind of progressive value.", 'start': 10521.447, 'duration': 5.264}], 'summary': 'Rectifier and sigmoid functions are commonly used in machine learning for their ability to handle positive and zero values effectively.', 'duration': 26.139, 'max_score': 10500.572, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM10500572.jpg'}, {'end': 10750.857, 'src': 'embed', 'start': 10722.372, 'weight': 3, 'content': [{'end': 10726.916, 'text': "But what happens within these neurons? So let's look at a little example of this.", 'start': 10722.372, 'duration': 4.544}, {'end': 10728.778, 'text': 'Kind of helps if you have some kind of visual.', 'start': 10727.157, 'duration': 1.621}, {'end': 10733.222, 'text': "Let's build a neural network that predict bike prices based on a few of its features.", 'start': 10728.818, 'duration': 4.404}, {'end': 10736.685, 'text': "And we'll see here we have our CC, our mileage, and our ABS.", 'start': 10733.543, 'duration': 3.142}, {'end': 10738.687, 'text': 'And these are our three input layers.', 'start': 10737.006, 'duration': 1.681}, {'end': 10741.39, 'text': 'And then we have the bike price and the output layer.', 'start': 10739.008, 'duration': 2.382}, {'end': 10745.453, 'text': "Now, it doesn't do us very good to just pump it in from the beginning and pump it out.", 'start': 10741.59, 'duration': 3.863}, {'end': 10750.857, 'text': 'And to be honest, I would use a machine learning linear regression model on this since these are just straight numbers.', 'start': 10745.573, 'duration': 5.284}], 'summary': 'Exploring neural network for bike price prediction using input layers and output layer with straight numbers.', 'duration': 28.485, 'max_score': 10722.372, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM10722372.jpg'}, {'end': 10845.789, 'src': 'embed', 'start': 10818.301, 'weight': 1, 'content': [{'end': 10824.742, 'text': 'they now have neurons that fire into the same layer or back a layer, so that you now have a time series,', 'start': 10818.301, 'duration': 6.441}, {'end': 10828.223, 'text': "and there's all kinds of wild things that they're experimenting with on these layers.", 'start': 10824.742, 'duration': 3.481}, {'end': 10831.303, 'text': 'This basic setup has been around since the mid-90s.', 'start': 10828.403, 'duration': 2.9}, {'end': 10836.724, 'text': "It's only now, because of our technology, that it's open to almost everybody to play with it,", 'start': 10831.663, 'duration': 5.061}, {'end': 10839.425, 'text': "and that's why I say this is in an infant stage in development.", 'start': 10836.724, 'duration': 2.701}, {'end': 10845.789, 'text': "is this basic math is here but what we can do with it is amazing and what they're actually doing with all these different things is amazing.", 'start': 10839.685, 'duration': 6.104}], 'summary': 'Neural network technology evolving since mid-90s, now accessible to almost everyone, still in early development stages.', 'duration': 27.488, 'max_score': 10818.301, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM10818301.jpg'}, {'end': 10901.509, 'src': 'embed', 'start': 10872.908, 'weight': 9, 'content': [{'end': 10875.13, 'text': 'um, i always look at this as like a group of people.', 'start': 10872.908, 'duration': 2.222}, {'end': 10879.633, 'text': "they're all looking at the bulletin board and the first person says this is what i project sells for the company,", 'start': 10875.13, 'duration': 4.503}, {'end': 10885.538, 'text': 'and the second person and the third and so on, and then their perspectives are weighted based on their expertise.', 'start': 10879.633, 'duration': 5.905}, {'end': 10893.304, 'text': 'so your accountant might have a very high weight where the um, maybe your janitor has a very low weight because their expertise is not in accounting.', 'start': 10885.538, 'duration': 7.766}, {'end': 10897.126, 'text': 'and then that goes into the output layer and once in the output layer it goes.', 'start': 10893.304, 'duration': 3.822}, {'end': 10901.509, 'text': 'the output, which is the predicted value, is compared against the original value.', 'start': 10897.126, 'duration': 4.383}], 'summary': 'Group perspectives weighted based on expertise for predicting values.', 'duration': 28.601, 'max_score': 10872.908, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM10872908.jpg'}, {'end': 10970.571, 'src': 'embed', 'start': 10943.979, 'weight': 6, 'content': [{'end': 10947.501, 'text': 'Remember, you might have a data pool with a terabyte of data.', 'start': 10943.979, 'duration': 3.522}, {'end': 10953.044, 'text': "You don't want to solve for the first set of data that comes in and that be the main solution because everything else will be off.", 'start': 10947.641, 'duration': 5.403}, {'end': 10954.285, 'text': 'This is going to confuse you.', 'start': 10953.244, 'duration': 1.041}, {'end': 10955.866, 'text': "That's also called a bias.", 'start': 10954.525, 'duration': 1.341}, {'end': 10964.03, 'text': "So we have the bias in the cell where we're adding a value, kind of like the y-intercept, and we have a bias of the whole neural network.", 'start': 10956.106, 'duration': 7.924}, {'end': 10967.151, 'text': 'which means it is weighted towards one set of answers.', 'start': 10964.35, 'duration': 2.801}, {'end': 10970.571, 'text': "So we want to make small changes in these weights so we don't create a bias.", 'start': 10967.451, 'duration': 3.12}], 'summary': 'Avoid bias by making small changes in neural network weights.', 'duration': 26.592, 'max_score': 10943.979, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM10943979.jpg'}, {'end': 11126.419, 'src': 'embed', 'start': 11098.243, 'weight': 2, 'content': [{'end': 11101.604, 'text': 'This error is based on the error of each cell generated.', 'start': 11098.243, 'duration': 3.361}, {'end': 11105.506, 'text': "How far off is that cell as far as its weights? We're not going to show you.", 'start': 11101.764, 'duration': 3.742}, {'end': 11109.807, 'text': "It's actually a very complicated differential equation, and you can probably write it out if you wanted to.", 'start': 11105.526, 'duration': 4.281}, {'end': 11111.308, 'text': 'You just write out each formula.', 'start': 11109.827, 'duration': 1.481}, {'end': 11115.349, 'text': 'That goes into the next level, and you add them all together, and you can write it out all the way through.', 'start': 11111.428, 'duration': 3.921}, {'end': 11117.11, 'text': "Computers make it so you don't have to.", 'start': 11115.529, 'duration': 1.581}, {'end': 11121.894, 'text': 'And our neural network is considered trained when the value for the cost function is minimum.', 'start': 11117.27, 'duration': 4.624}, {'end': 11126.419, 'text': "So when we get our error way down as low as we can, that's when our neural network is trained.", 'start': 11121.954, 'duration': 4.465}], 'summary': 'Neural network training aims to minimize error through cost function optimization.', 'duration': 28.176, 'max_score': 11098.243, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM11098243.jpg'}, {'end': 11180.182, 'src': 'embed', 'start': 11149.958, 'weight': 0, 'content': [{'end': 11157.563, 'text': "we're at the beginning stages in neural networks and it's just really cool what they can do now and how much of it's automated and how much of it is experimental right now.", 'start': 11149.958, 'duration': 7.605}, {'end': 11159.325, 'text': "Let's take a look at gradient descent.", 'start': 11157.744, 'duration': 1.581}, {'end': 11166.27, 'text': 'But what approach do we take to minimize the cost function? So here we have a nice error thing coming in.', 'start': 11159.605, 'duration': 6.665}, {'end': 11167.771, 'text': 'This is our cost or our error.', 'start': 11166.29, 'duration': 1.481}, {'end': 11171.534, 'text': "Let's start with plotting the cost function against the predicted value.", 'start': 11167.791, 'duration': 3.743}, {'end': 11180.182, 'text': "And so you can see they've fed in multiple y's and these are the errors coming in and the cost of each of these inputs and changes going on.", 'start': 11171.795, 'duration': 8.387}], 'summary': 'Intro to neural networks, discussing gradient descent and cost function analysis.', 'duration': 30.224, 'max_score': 11149.958, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM11149958.jpg'}, {'end': 11468.813, 'src': 'embed', 'start': 11438.254, 'weight': 5, 'content': [{'end': 11444.559, 'text': 'You know, it does increase the processing, but I did run into some version problems with my Python and stuff like that.', 'start': 11438.254, 'duration': 6.305}, {'end': 11449.564, 'text': "And when I did finally work it out, I went back to the CPU because it didn't increase my speed enough for what I was working on.", 'start': 11444.579, 'duration': 4.985}, {'end': 11451.885, 'text': 'But in a larger group, you might be able to put that on.', 'start': 11449.744, 'duration': 2.141}, {'end': 11455.909, 'text': "If you're working with a larger stack of computers, you might want to run it in the GPU.", 'start': 11451.926, 'duration': 3.983}, {'end': 11459.39, 'text': 'You can create a data flow graphs that have nodes and edges.', 'start': 11456.109, 'duration': 3.281}, {'end': 11461.19, 'text': "So there's our edges coming in.", 'start': 11459.71, 'duration': 1.48}, {'end': 11468.813, 'text': "We didn't talk about edges, but that's a very up and coming way of looking at your analytical data is how do different nodes connect?", 'start': 11461.37, 'duration': 7.443}], 'summary': 'Using gpu did not increase speed enough, data flow graphs include nodes and edges.', 'duration': 30.559, 'max_score': 11438.254, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM11438254.jpg'}], 'start': 10188.553, 'title': 'Neural network fundamentals', 'summary': 'Covers neural network layers, biases, weights, activation functions, back propagation, and tensorflow, emphasizing their roles in information transmission, model performance, and practical applications in deep learning, with specific focus on tensorflow 1.12 setup and data exploration.', 'chapters': [{'end': 10285.531, 'start': 10188.553, 'title': 'Neural network layers and information transmission', 'summary': 'Explains the concept of neural network layers, biases, weights, and output layers, and their role in information transmission, leading to the predicted output, such as classification as true or false.', 'duration': 96.978, 'highlights': ['Neurons of each layer transmit information to neurons of the next layer over channels.', 'The output layer emits a predicted output, such as a classification as true or false.', 'The concept of biases and weights, and their role in determining the information passed between neurons.']}, {'end': 10560.435, 'start': 10285.711, 'title': 'Neural network activation functions', 'summary': 'Discusses the operations performed within each neuron, including the computation of weighted sum and bias unique to the neuron, the significance of activation functions on model performance, and popular types of activation functions such as sigmoid, threshold, and rectifier functions, each with specific use cases and formulas.', 'duration': 274.724, 'highlights': ["The sigmoid function is used for models where we have to predict the probability as an output, existing between 0 and 1, with the formula 1 / (1 + e^-x), and it's probably the default on most models.", 'The threshold function is a threshold-based activation function that provides a yes/no, true/false output based on a certain value of x.', 'The rectifier function is widely used, providing an output of x if x is positive and 0 otherwise, making it suitable for handling non-linearities in data.']}, {'end': 11330.295, 'start': 10560.435, 'title': 'Neural network basics', 'summary': 'Discusses the basics of neural networks, including activation functions, cost functions, back propagation, training process, gradient descent, and popular deep learning platforms, emphasizing the practical applications and significance of these concepts in deep learning.', 'duration': 769.86, 'highlights': ["The cost function measures the difference between the neural net's predicted output and the actual output, and the network is trained by making adjustments to the weights and biases iteratively throughout the training process in order to minimize the error.", 'Back propagation involves sending the error or cost back and adjusting the weights in small increments across large amounts of data to minimize the error and prevent bias in the neural network.', 'The chapter explains the process of gradient descent to minimize the cost function, emphasizing the significance of finding the least value of the cost function for training the neural network.', 'The transcript introduces popular deep learning platforms, such as TensorFlow, Deep Learning 4J, Keras, and Torch, highlighting their significance and interconnections in the field of deep learning.']}, {'end': 11709.211, 'start': 11330.495, 'title': 'Understanding tensors and tensorflow', 'summary': 'Explains the concept of tensors and their use in tensorflow, highlighting its development by google, its application in processing high-level data, and the implementation of a neural network to identify handwritten digits using the mnist database.', 'duration': 378.716, 'highlights': ['TensorFlow is the most popular library in deep learning, developed by Google.', 'The MNIST database contains 70,000 handwritten digits, and a neural network is implemented to identify these digits.', 'TensorFlow implementation involves creating a neural network to identify handwritten digits using the MNIST database.', 'Tensors are arrays and are used to process high-level data, such as images and features, in TensorFlow.', 'TensorFlow can be run on either a CPU or a GPU, offering faster processing on the latter.']}, {'end': 12397.39, 'start': 11709.391, 'title': 'Tensorflow 1.12 setup and data exploration', 'summary': 'Demonstrates setting up a python 3.6 environment with tensorflow 1.12 and numpy, focusing on the importance of consistent package management in anaconda environments. it then delves into data exploration, plotting, and examining the mnist dataset, followed by the setup of tensorflow variables, placeholders, and the gradient descent optimizer for training.', 'duration': 687.999, 'highlights': ['Setting up Python 3.6 environment with TensorFlow 1.12 and NumPy', 'Exploring the MNIST dataset and data plotting', 'Setting up TensorFlow variables, placeholders, and gradient descent optimizer']}], 'duration': 2208.837, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM10188553.jpg', 'highlights': ['TensorFlow is the most popular library in deep learning, developed by Google.', "The cost function measures the difference between the neural net's predicted output and the actual output.", 'Back propagation involves sending the error or cost back and adjusting the weights in small increments.', 'The sigmoid function is used for models where we have to predict the probability as an output.', 'The concept of biases and weights, and their role in determining the information passed between neurons.', 'The output layer emits a predicted output, such as a classification as true or false.', 'The chapter explains the process of gradient descent to minimize the cost function.', 'Tensors are arrays and are used to process high-level data, such as images and features, in TensorFlow.', 'The MNIST database contains 70,000 handwritten digits, and a neural network is implemented to identify these digits.', 'Setting up Python 3.6 environment with TensorFlow 1.12 and NumPy']}, {'end': 14400.759, 'segs': [{'end': 12824.473, 'src': 'embed', 'start': 12795.961, 'weight': 5, 'content': [{'end': 12801.384, 'text': "And you can simply, we'll do this, I have it installed, but you could do all, we'll do all.", 'start': 12795.961, 'duration': 5.423}, {'end': 12804.925, 'text': 'You can do a search under all for tensor, and you can see all the different tensors.', 'start': 12801.564, 'duration': 3.361}, {'end': 12808.627, 'text': "It's actually installed in here, version 1.1.", 'start': 12805.085, 'duration': 3.542}, {'end': 12813.049, 'text': "3.1 If it wasn't, you could check the box and then run the install on there, and it'd bring it right in.", 'start': 12808.627, 'duration': 4.422}, {'end': 12815.73, 'text': "But we'll go ahead and start up our JupyterLab, which is going to open up.", 'start': 12813.109, 'duration': 2.621}, {'end': 12817.69, 'text': 'In this case, I use Google Chrome.', 'start': 12815.99, 'duration': 1.7}, {'end': 12824.473, 'text': "And in our JupyterLab or Jupyter Notebook, if you're in the notebook, you only have one tab and you won't have the added options,", 'start': 12817.891, 'duration': 6.582}], 'summary': 'Installing version 1.1.3 of tensor, open jupyterlab in google chrome.', 'duration': 28.512, 'max_score': 12795.961, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM12795961.jpg'}, {'end': 12857.75, 'src': 'embed', 'start': 12829.335, 'weight': 4, 'content': [{'end': 12831.456, 'text': "But everything we're going to do, you can easily do in the notebook.", 'start': 12829.335, 'duration': 2.121}, {'end': 12833.113, 'text': "You'll start up a new project.", 'start': 12831.812, 'duration': 1.301}, {'end': 12835.054, 'text': "Deep learning is what I'm going to call this.", 'start': 12833.373, 'duration': 1.681}, {'end': 12837.696, 'text': "And if you're not familiar, you can definitely.", 'start': 12835.675, 'duration': 2.021}, {'end': 12843.16, 'text': 'we have some tutorials out on the use of Jupyter Notebook and how to run it and set it up and things like that.', 'start': 12837.696, 'duration': 5.464}, {'end': 12845.382, 'text': 'The most basic is we put our code in here.', 'start': 12843.28, 'duration': 2.102}, {'end': 12849.544, 'text': 'It has a nice display, a nice interface, especially for data science.', 'start': 12845.402, 'duration': 4.142}, {'end': 12851.686, 'text': 'I can display all kinds of things on this page.', 'start': 12849.685, 'duration': 2.001}, {'end': 12853.667, 'text': 'And then you can just run this page right here.', 'start': 12851.826, 'duration': 1.841}, {'end': 12857.75, 'text': "There's no code in it, so it's not going to show anything until I put some code in there.", 'start': 12853.867, 'duration': 3.883}], 'summary': 'Introduction to using jupyter notebook for deep learning projects.', 'duration': 28.415, 'max_score': 12829.335, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM12829335.jpg'}, {'end': 13097.36, 'src': 'embed', 'start': 13067.804, 'weight': 6, 'content': [{'end': 13071.066, 'text': "So we'll use the pandas D types.", 'start': 13067.804, 'duration': 3.262}, {'end': 13072.026, 'text': "We'll run that.", 'start': 13071.326, 'duration': 0.7}, {'end': 13076.588, 'text': 'And you can see here where age came in as an integer, integer 64.', 'start': 13072.326, 'duration': 4.262}, {'end': 13079.39, 'text': 'Working class as an object, which makes sense.', 'start': 13076.588, 'duration': 2.802}, {'end': 13084.292, 'text': 'And then we have our FNLWGT, the education as objects.', 'start': 13080.07, 'duration': 4.222}, {'end': 13088.114, 'text': 'Well, this is an integer, integer object, so on, all the way down.', 'start': 13084.732, 'duration': 3.382}, {'end': 13093.758, 'text': "And if we flash back to the data, we look at the last column, it's less than or equal to 50k.", 'start': 13088.314, 'duration': 5.444}, {'end': 13097.36, 'text': "And if we scroll down enough, we'll see it's also greater than or equal to 50k.", 'start': 13093.778, 'duration': 3.582}], 'summary': 'Analyzing data types using pandas, including integer and object types, with income categorized as <=50k and >=50k.', 'duration': 29.556, 'max_score': 13067.804, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM13067804.jpg'}, {'end': 13290.277, 'src': 'embed', 'start': 13268.099, 'weight': 1, 'content': [{'end': 13276.044, 'text': "fnlwgt is an integer, capital gain is an integer, educational number, there's our educational number as an integer, and so on.", 'start': 13268.099, 'duration': 7.945}, {'end': 13280.607, 'text': "So if we're going to have these as continuous features where they're an integer,", 'start': 13276.505, 'duration': 4.102}, {'end': 13289.096, 'text': 'we also need to make a list of the categorical features we want to work with, such as working class education, marital occupation, relationship, race,', 'start': 13280.607, 'duration': 8.489}, {'end': 13290.277, 'text': 'sex and native country.', 'start': 13289.096, 'duration': 1.181}], 'summary': 'Identify integer and categorical features for analysis.', 'duration': 22.178, 'max_score': 13268.099, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM13268099.jpg'}, {'end': 13759.074, 'src': 'embed', 'start': 13731.338, 'weight': 0, 'content': [{'end': 13735.639, 'text': "We've told it what columns it's going to pull in so it knows what columns it's looking at.", 'start': 13731.338, 'duration': 4.301}, {'end': 13738.641, 'text': "know what the definition where it's going to get the information from.", 'start': 13735.979, 'duration': 2.662}, {'end': 13740.062, 'text': 'So now we want to go ahead and train it.', 'start': 13738.801, 'duration': 1.261}, {'end': 13744.424, 'text': "And here's our input function equals, in this case, get input function.", 'start': 13740.202, 'duration': 4.222}, {'end': 13747.827, 'text': 'And here we tell it that our data set is a DF train.', 'start': 13744.745, 'duration': 3.082}, {'end': 13751.629, 'text': 'Number of epochs equals none, which is already automatically set up there.', 'start': 13748.007, 'duration': 3.622}, {'end': 13755.031, 'text': 'In batch equals 128, which is what we have up here.', 'start': 13751.849, 'duration': 3.182}, {'end': 13756.492, 'text': 'Shuffle equals false.', 'start': 13755.331, 'duration': 1.161}, {'end': 13759.074, 'text': "And we're going to do 1, 000 steps.", 'start': 13756.972, 'duration': 2.102}], 'summary': 'Training the model with df train dataset, using 128 batch size for 1000 steps.', 'duration': 27.736, 'max_score': 13731.338, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM13731338.jpg'}, {'end': 14031.055, 'src': 'embed', 'start': 14005.381, 'weight': 3, 'content': [{'end': 14012.146, 'text': 'And our training set has 32, 561 rows in it, and this one has just over 16, 000 in there.', 'start': 14005.381, 'duration': 6.765}, {'end': 14019.35, 'text': "And everything's pretty much the same, except that in our continuous features new, we now have the new column we put in there.", 'start': 14012.326, 'duration': 7.024}, {'end': 14020.491, 'text': 'So we need to change that.', 'start': 14019.45, 'duration': 1.041}, {'end': 14025.803, 'text': "And we have our continuous features new down here, which we're going to load up with our TF feature column numerical.", 'start': 14020.771, 'duration': 5.032}, {'end': 14027.707, 'text': 'This would all look familiar because we just did this.', 'start': 14025.963, 'duration': 1.744}, {'end': 14031.055, 'text': "And so we're going to load that up with the new being the new value in there.", 'start': 14027.867, 'duration': 3.188}], 'summary': 'The training set has 32,561 rows, and the new set has over 16,000 rows. changes are made to the continuous features with the addition of a new column.', 'duration': 25.674, 'max_score': 14005.381, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM14005381.jpg'}, {'end': 14094.096, 'src': 'embed', 'start': 14057.783, 'weight': 2, 'content': [{'end': 14064.306, 'text': "So where's the input coming from? So we built the model with the features in it, and now we need to go ahead and create our get input function.", 'start': 14057.783, 'duration': 6.523}, {'end': 14071.21, 'text': "This is the same as before, but you'll see we now have it with the data set coming in is going to be the same data set up here.", 'start': 14064.466, 'duration': 6.744}, {'end': 14074.692, 'text': 'And so if we scroll a little bit to the right, we should see the new.', 'start': 14071.45, 'duration': 3.242}, {'end': 14075.392, 'text': 'And there it is.', 'start': 14074.872, 'duration': 0.52}, {'end': 14076.933, 'text': "Sure enough, there's our new on the right.", 'start': 14075.472, 'duration': 1.461}, {'end': 14078.513, 'text': "And so let's go ahead and run that.", 'start': 14077.313, 'duration': 1.2}, {'end': 14080.214, 'text': "Now we've defined our model.", 'start': 14078.813, 'duration': 1.401}, {'end': 14087.995, 'text': "we've defined where the information is coming from, and again you can go back and review the first part, because this is identical that we did before,", 'start': 14080.214, 'duration': 7.781}, {'end': 14091.036, 'text': 'except now we have new as the column for the age.', 'start': 14087.995, 'duration': 3.041}, {'end': 14094.096, 'text': "And we'll go ahead and do our model one, and let's go ahead and train it.", 'start': 14091.416, 'duration': 2.68}], 'summary': 'Building and training a model with new data set and feature columns.', 'duration': 36.313, 'max_score': 14057.783, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM14057783.jpg'}], 'start': 12397.41, 'title': 'Tensorflow implementation and model analysis', 'summary': 'Covers tensorflow training, achieving 91.65% accuracy, use case implementation, data preprocessing, model training, and evaluation with 79% accuracy, with a focus on feature extraction and age analysis.', 'chapters': [{'end': 12756.462, 'start': 12397.41, 'title': 'Tensorflow training and network setup', 'summary': 'Covers the process of initializing and running a tensorflow object, setting up a neural network, training it, and evaluating its accuracy, achieving an accuracy of 0.9165 on the test data by training a neural network for recognizing symbols.', 'duration': 359.052, 'highlights': ['The process of initializing and running a TensorFlow object involves bundling all the steps into one, initializing variables, and creating a session.', 'Training the neural network involves running it for 1000 iterations with batch sizes of 100, allowing for efficient training and improving accuracy.', 'The accuracy achieved on the test data after training the neural network is 0.9165, indicating a good performance in recognizing symbols.']}, {'end': 13007.438, 'start': 12756.462, 'title': 'Tensorflow use case implementation', 'summary': 'Covers the implementation of use cases with tensorflow through tools like jupyterlab and anaconda, including the installation of tensorflow and pandas, setting up data paths, and loading data using pandas, with a focus on a machine learning database.', 'duration': 250.976, 'highlights': ['The chapter covers the implementation of use cases with TensorFlow through tools like JupyterLab and Anaconda.', 'The installation of TensorFlow and Pandas is demonstrated, with emphasis on verifying the installation and accessing the required libraries.', 'Setting up data paths and obtaining a machine learning database are crucial aspects of the implementation.', 'Loading data using pandas is illustrated, showcasing the practical application of the library in data manipulation and analysis.']}, {'end': 13350.925, 'start': 13007.978, 'title': 'Data preprocessing and feature extraction', 'summary': 'Covers preprocessing of data using pandas, including skipping and handling rows, checking data shape, data types, and labeling, as well as creating categorical and continuous feature lists for manipulation using tensorflow, with a focus on ensuring the data is in a format that the computer can work with, and includes looking at the shape of the data, data types, and labels, as well as creating categorical and continuous feature lists for manipulation using tensorflow.', 'duration': 342.947, 'highlights': ['The data sets have 32,561 and 16,281 rows respectively, and 15 columns each, checked using the shape attribute.', 'The data types of various columns are inspected with pandas D types, revealing information such as age being an integer and education being an object.', 'The labeling of data is performed by setting specific labels based on conditions, with the label being set to 0 or 1 depending on whether the value is less than or equal to 50k or greater than or equal to 50k.', 'Creation of categorical and continuous feature lists is done to prepare the data for manipulation using TensorFlow, with a focus on ensuring the data is in a format that the computer can work with.']}, {'end': 13839.858, 'start': 13350.965, 'title': 'Tensorflow data loading and model training', 'summary': 'Covers the process of creating feature columns, setting up relationships, loading continuous and categorical features, creating a linear classifier model with two classes, and training it with 1000 steps using tensorflow.', 'duration': 488.893, 'highlights': ['Creating feature columns and setting up relationships', 'Loading continuous and categorical features with vocabulary lists and hash buckets', 'Creating a linear classifier model with two classes and training it with 1000 steps']}, {'end': 14400.759, 'start': 13840.199, 'title': 'Tensorflow model evaluation and age analysis', 'summary': 'Discusses evaluating a tensorflow model with an accuracy of 0.79, tweaking the model by analyzing the square value of age, and utilizing the model for predictions, highlighting its limitations and data improvement strategies.', 'duration': 560.56, 'highlights': ['The TensorFlow model evaluation yielded an accuracy of 0.79, surpassing the baseline accuracy of 0.76, indicating the effectiveness of the model.', 'The process of tweaking the model involved analyzing the square value of age to account for the non-linear relationship between age and income, showcasing the data-specific approach to model optimization.', 'Utilizing the model for predictions, it was observed that the accuracy did not significantly improve due to the data partitioning and the limited impact of age on income, highlighting the importance of understanding data limitations and improvement strategies.']}], 'duration': 2003.349, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM12397410.jpg', 'highlights': ['The accuracy achieved on the test data after training the neural network is 0.9165, indicating a good performance in recognizing symbols.', 'The TensorFlow model evaluation yielded an accuracy of 0.79, surpassing the baseline accuracy of 0.76, indicating the effectiveness of the model.', 'The process of initializing and running a TensorFlow object involves bundling all the steps into one, initializing variables, and creating a session.', 'Training the neural network involves running it for 1000 iterations with batch sizes of 100, allowing for efficient training and improving accuracy.', 'The chapter covers the implementation of use cases with TensorFlow through tools like JupyterLab and Anaconda.', 'The data sets have 32,561 and 16,281 rows respectively, and 15 columns each, checked using the shape attribute.', 'The labeling of data is performed by setting specific labels based on conditions, with the label being set to 0 or 1 depending on whether the value is less than or equal to 50k or greater than or equal to 50k.', 'The process of tweaking the model involved analyzing the square value of age to account for the non-linear relationship between age and income, showcasing the data-specific approach to model optimization.']}, {'end': 15055.765, 'segs': [{'end': 14439.127, 'src': 'embed', 'start': 14400.979, 'weight': 1, 'content': [{'end': 14406.046, 'text': 'And this is probably one of the more complicated parts of this is our input function.', 'start': 14400.979, 'duration': 5.067}, {'end': 14408.048, 'text': 'And this input function is so important.', 'start': 14406.226, 'duration': 1.822}, {'end': 14411.533, 'text': 'So I want to just rehash the input function for that reason.', 'start': 14408.088, 'duration': 3.445}, {'end': 14413.695, 'text': "This lets us know how we're pulling the data.", 'start': 14411.853, 'duration': 1.842}, {'end': 14417.5, 'text': "It lets us know if we're going to go through all the data 20 times.", 'start': 14413.896, 'duration': 3.604}, {'end': 14424.624, 'text': "or we're just going to let TensorFlow itself keep going until the training model reaches a point.", 'start': 14419.022, 'duration': 5.602}, {'end': 14429.125, 'text': 'That point is based on what they call bias.', 'start': 14424.944, 'duration': 4.181}, {'end': 14431.245, 'text': 'You can become biased on your training data.', 'start': 14429.345, 'duration': 1.9}, {'end': 14438.367, 'text': "And so when it hits a certain point where it becomes overly biased to just that data, then it doesn't really work really good in the outside world.", 'start': 14431.405, 'duration': 6.962}, {'end': 14439.127, 'text': 'You start losing.', 'start': 14438.407, 'duration': 0.72}], 'summary': 'The input function is crucial for data handling and preventing bias in training data, impacting model performance.', 'duration': 38.148, 'max_score': 14400.979, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM14400979.jpg'}, {'end': 14503.838, 'src': 'embed', 'start': 14475.828, 'weight': 0, 'content': [{'end': 14479.911, 'text': "So it gives you a whole roundabout setup on here and how this is set up and how it's working.", 'start': 14475.828, 'duration': 4.083}, {'end': 14485.318, 'text': 'Certainly there are a lot more things you can do with TensorFlow.', 'start': 14480.071, 'duration': 5.247}, {'end': 14488.982, 'text': "This is the basic TensorFlow, and it's always developing, so it's exciting.", 'start': 14485.478, 'duration': 3.504}, {'end': 14496.071, 'text': 'This is going to be one of the most exciting fields right now because it is really in an infant stage and just exploding in the market.', 'start': 14489.042, 'duration': 7.029}, {'end': 14503.838, 'text': "In today's exciting world of AI, artificial intelligence, we're going to cover some of the many uses it's being applied to.", 'start': 14496.371, 'duration': 7.467}], 'summary': 'Basic tensorflow is constantly developing, with ai in an infant stage and booming in the market.', 'duration': 28.01, 'max_score': 14475.828, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM14475828.jpg'}, {'end': 14629.534, 'src': 'embed', 'start': 14599.468, 'weight': 7, 'content': [{'end': 14601.649, 'text': 'And a number of cities are exploring that for medical use.', 'start': 14599.468, 'duration': 2.181}, {'end': 14607.074, 'text': 'to how do you reprogram T cells in the human body to combat disease.', 'start': 14601.949, 'duration': 5.125}, {'end': 14609.256, 'text': "There's whole industries in data analysis.", 'start': 14607.354, 'duration': 1.902}, {'end': 14611.057, 'text': 'Another one is chemistry.', 'start': 14609.516, 'duration': 1.541}, {'end': 14618.244, 'text': 'How do you use a chemical cell or a chemistry molecule not cell, a chemical molecule to imitate the T cell?', 'start': 14611.217, 'duration': 7.027}, {'end': 14624.089, 'text': 'So you can then put those molecules in there and then they grab the disease cells out or the cancer cells or whatever out,', 'start': 14618.364, 'duration': 5.725}, {'end': 14625.25, 'text': 'just like the human body would.', 'start': 14624.089, 'duration': 1.161}, {'end': 14629.534, 'text': 'So discovering new drugs and how to apply that is just a huge industry.', 'start': 14625.45, 'duration': 4.084}], 'summary': 'Cities are exploring medical use of reprogrammed t cells and new drugs, with a focus on data analysis and chemistry.', 'duration': 30.066, 'max_score': 14599.468, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM14599468.jpg'}, {'end': 14899.097, 'src': 'embed', 'start': 14870.153, 'weight': 6, 'content': [{'end': 14873.474, 'text': 'And then you take the new one and it recreates the image by adding the colors.', 'start': 14870.153, 'duration': 3.321}, {'end': 14878.179, 'text': 'And this is also used in the movie industry to add in special effects.', 'start': 14873.894, 'duration': 4.285}, {'end': 14880.922, 'text': "There's all kinds of cool things it can do with this.", 'start': 14878.78, 'duration': 2.142}, {'end': 14883.265, 'text': 'It goes way beyond just coloring pictures.', 'start': 14881.022, 'duration': 2.243}, {'end': 14888.09, 'text': 'Of course, the coloring pictures is something that you can probably do in your own lab, on your own computer,', 'start': 14883.525, 'duration': 4.565}, {'end': 14890.352, 'text': 'and build your own neural network to do that.', 'start': 14888.09, 'duration': 2.262}, {'end': 14892.453, 'text': 'Robotics, one of my favorites.', 'start': 14890.552, 'duration': 1.901}, {'end': 14899.097, 'text': 'Robotics, all these tools that make them so exciting, is that you can go out and even now used to be.', 'start': 14893.013, 'duration': 6.084}], 'summary': 'Neural networks recreate images, used in movie industry for special effects, and extend beyond coloring pictures.', 'duration': 28.944, 'max_score': 14870.153, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM14870153.jpg'}], 'start': 14400.979, 'title': 'Neural network input function and deep learning in various industries', 'summary': 'Covers the importance of the input function in a neural network, including considerations for data pulling, bias, generic answers for large data, batch size, shuffling, and estimator inputs. it also discusses the use of deep learning in reshaping healthcare, entertainment, music generation, robotics, and image caption generation, highlighting its impact and application in each sector.', 'chapters': [{'end': 14458.677, 'start': 14400.979, 'title': 'Neural network input function', 'summary': 'Covers the importance of the input function in a neural network, highlighting the considerations for data pulling, bias, generic answers for large data, batch size, shuffling, and estimator inputs.', 'duration': 57.698, 'highlights': ['The input function determines how data is pulled and whether it will go through all the data or continue until the training model reaches a certain point of bias, impacting its effectiveness in the outside world.', 'Maintaining generality while being as close as possible to the right answer is crucial for a large amount of data in the input function.', 'Considerations such as batch size, shuffling, and estimator inputs are essential components of the input function for a neural network.']}, {'end': 15055.765, 'start': 14458.717, 'title': 'Deep learning in various industries', 'summary': 'Discusses the use of deep learning in various industries, including reshaping healthcare industry, entertainment (netflix, amazon, filmmaking), music and audio generation, robotics, and image caption generation, highlighting its impact and application in each sector.', 'duration': 597.048, 'highlights': ["Deep learning reshapes healthcare industry by delivering new possibilities to improve people's lives, such as early detection of cancer cells and tumors, synthesizing new drugs, and enhancing medical instruments.", 'Deep learning is used in the entertainment industry, including Netflix, Amazon, and filmmaking, for personalized recommendations and content curation.', "IBM Watson was used in Wimbledon 2018 to analyze player emotions and auto-generate highlights, showcasing deep learning's application in sports.", 'Deep learning is utilized in music and audio generation for training networks to produce music compositions and reading lip movements, demonstrating its impact on creative and technological advancements.', 'Deep learning is applied in robotics to build robots capable of performing human-like tasks, making better and safer decisions without supervision, and replacing humans in hazardous environments, such as nuclear reactors and outer space.', 'Deep learning is used for image caption generation, training systems to detect objects in photographs and turn labels into coherent sentences, improving image organization and search capabilities.']}], 'duration': 654.786, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM14400979.jpg', 'highlights': ['Deep learning reshapes healthcare industry with early cancer detection and drug synthesis.', 'Deep learning applied in robotics for human-like tasks and hazardous environment work.', 'Deep learning used in entertainment industry for personalized recommendations and content curation.', 'Deep learning utilized in music generation and reading lip movements for creative and technological advancements.', 'Deep learning applied in image caption generation for improved image organization and search capabilities.', 'IBM Watson used in Wimbledon 2018 to analyze player emotions and auto-generate highlights.', 'Considerations such as batch size, shuffling, and estimator inputs are essential components of the input function for a neural network.', 'Maintaining generality while being as close as possible to the right answer is crucial for a large amount of data in the input function.', 'The input function determines how data is pulled and whether it will go through all the data or continue until the training model reaches a certain point of bias, impacting its effectiveness in the outside world.']}, {'end': 16215.588, 'segs': [{'end': 15117.868, 'src': 'embed', 'start': 15077.961, 'weight': 0, 'content': [{'end': 15083.887, 'text': 'Allows a network to reduce cost by dropping the cost per acquisition of a campaign from $60 to $30.', 'start': 15077.961, 'duration': 5.926}, {'end': 15089.654, 'text': 'Create data-driven predictive advertising, real-time bidding of their ads, and target display advertising.', 'start': 15083.887, 'duration': 5.767}, {'end': 15094.359, 'text': "And again, this goes from experimental where I'm just going to guess what people want.", 'start': 15089.794, 'duration': 4.565}, {'end': 15100.341, 'text': "and we'll put out a pink sign that says, you want to add on widget to your order.", 'start': 15094.799, 'duration': 5.542}, {'end': 15103.603, 'text': 'Well, nobody wants widgets and nobody wants to see the pink sign.', 'start': 15100.561, 'duration': 3.042}, {'end': 15105.263, 'text': 'Maybe you got three people want to see the pink sign.', 'start': 15103.743, 'duration': 1.52}, {'end': 15111.846, 'text': 'And you find out that the blue sign, a lot of people are reading that and they really want a widget number two, whatever that is.', 'start': 15105.363, 'duration': 6.483}, {'end': 15117.868, 'text': 'So when you talk about advertising, It also deals with marketing and also preference of merchandise.', 'start': 15112.526, 'duration': 5.342}], 'summary': 'Network reduced cost per acquisition from $60 to $30, using data-driven advertising and real-time bidding.', 'duration': 39.907, 'max_score': 15077.961, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM15077961.jpg'}], 'start': 15056.105, 'title': 'Ai impact and skills', 'summary': 'Covers the impact of ai in advertising with a 50,000% improvement in earthquake prediction, essential skills for ai engineers, including knowledge of ai models, spark and big data technologies, algorithms and frameworks, and highlights career roles and salary insights for ai engineers.', 'chapters': [{'end': 15344.223, 'start': 15056.105, 'title': 'Ai in advertising and earthquake prediction', 'summary': 'Discusses the impact of ai in advertising, highlighting a 50,000% improvement in earthquake prediction using deep learning, and it also provides an overview of the responsibilities and skills required to become an ai engineer.', 'duration': 288.118, 'highlights': ['Deep learning in advertising reduces cost per acquisition of a campaign from $60 to $30, improving relevancy of ads and boosting advertising campaigns.', 'AI in advertising and marketing generates more sales, impacting various industries including real estate, property management, rentals, and Amazon merchandise.', 'Deep learning model improved earthquake prediction by 50,000%, enabling more reliable forecasts and human safety.', 'AI engineer responsibilities include developing, testing, and deploying AI models, converting machine learning models into APIs, and building data ingestion and transformation infrastructure.', 'Skills required for an AI engineer include programming languages such as Python, R, Java, and C++, as well as knowledge in linear algebra, probability, and statistics.']}, {'end': 15692.442, 'start': 15344.443, 'title': 'Skills for ai engineers & career opportunities', 'summary': "Discusses the essential skills for ai engineers including knowledge of ai models, spark and big data technologies, algorithms and frameworks, as well as communication and problem-solving skills. it also highlights the average salary of an ai engineer, career roles in artificial intelligence, and the ai engineers master's program offered by simply learn in collaboration with ibm.", 'duration': 347.999, 'highlights': ['The average salary of an AI engineer is $111,000 per annum in the United States and Rs. 13 lakhs per annum in India.', "Simply Learn has tied up with IBM and is offering the AI Engineers Master's Program which consists of various courses, such as Introduction to Artificial Intelligence, Data Science with Python, Machine Learning, Deep Learning Fundamentals by IBM, Deep Learning with TensorFlow, and the AI Capstone Project.", 'The third skill required is knowledge about Spark and Big Data technologies, such as Apache Spark, Hadoop, Cassandra, and MongoDB.', 'The final skill required to become an AI engineer is communication and problem-solving skills.', 'The chapter discusses the essential skills for AI engineers including knowledge of AI models, Spark and big data technologies, algorithms and frameworks, as well as communication and problem-solving skills.']}, {'end': 16215.588, 'start': 15692.915, 'title': 'Types of machine learning and overfitting', 'summary': 'Discusses the two main types of machine learning: supervised and unsupervised learning, and reinforcement learning, emphasizing the importance of labeled data in supervised learning and providing an analogy for overfitting, along with techniques to avoid it.', 'duration': 522.673, 'highlights': ['Supervised learning requires labeled data and is used for regression and classification, while unsupervised learning involves unlabeled data and is used for clustering.', 'Reinforcement learning involves an agent working in an environment to achieve a target, where rewards and punishments are used to guide the agent, as demonstrated in the example of AlphaGo.', 'Overfitting is when the model memorizes the training data, leading to high accuracy during training but low accuracy during testing, and can be avoided using techniques like regularization.', 'Training set and test set are used in machine learning to train and evaluate models, with the training set used for model training and the test set used to assess model performance.']}], 'duration': 1159.483, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM15056105.jpg', 'highlights': ['Deep learning model improved earthquake prediction by 50,000%, enabling more reliable forecasts and human safety.', 'Deep learning in advertising reduces cost per acquisition of a campaign from $60 to $30, improving relevancy of ads and boosting advertising campaigns.', 'AI in advertising and marketing generates more sales, impacting various industries including real estate, property management, rentals, and Amazon merchandise.', 'The average salary of an AI engineer is $111,000 per annum in the United States and Rs. 13 lakhs per annum in India.', 'AI engineer responsibilities include developing, testing, and deploying AI models, converting machine learning models into APIs, and building data ingestion and transformation infrastructure.', 'Supervised learning requires labeled data and is used for regression and classification, while unsupervised learning involves unlabeled data and is used for clustering.']}, {'end': 17727.657, 'segs': [{'end': 16281.368, 'src': 'embed', 'start': 16215.588, 'weight': 2, 'content': [{'end': 16217.631, 'text': 'so we have a three step process.', 'start': 16215.588, 'duration': 2.043}, {'end': 16221.615, 'text': 'We train the model and then we test the model.', 'start': 16218.051, 'duration': 3.564}, {'end': 16226.28, 'text': 'And then once we are satisfied with the test, only then we deploy the model.', 'start': 16221.755, 'duration': 4.525}, {'end': 16232.827, 'text': 'So what happens in the train and test is that you remember the labeled data.', 'start': 16226.5, 'duration': 6.327}, {'end': 16237.112, 'text': "So let's say you have 1000 records with labeling information.", 'start': 16233.047, 'duration': 4.065}, {'end': 16244.24, 'text': 'Now, one way of doing it is you use all the 1000 records for training and then maybe right,', 'start': 16237.292, 'duration': 6.948}, {'end': 16248.845, 'text': 'which means that you have exposed all this 1000 records during the training process.', 'start': 16244.24, 'duration': 4.605}, {'end': 16254.389, 'text': 'And then you take a small set of the same data and then you say, OK, I will test it with this.', 'start': 16249.085, 'duration': 5.304}, {'end': 16258.632, 'text': 'OK, and then you probably what will happen? You may get some good results.', 'start': 16254.789, 'duration': 3.843}, {'end': 16260.854, 'text': 'But there is a flaw there.', 'start': 16259.393, 'duration': 1.461}, {'end': 16263.576, 'text': 'What is the flaw? This is very similar to human beings.', 'start': 16260.914, 'duration': 2.662}, {'end': 16269.02, 'text': 'It is like you are showing this model the entire data as a part of training.', 'start': 16263.656, 'duration': 5.364}, {'end': 16272.502, 'text': 'So obviously it has become familiar with the entire data.', 'start': 16269.66, 'duration': 2.842}, {'end': 16278.867, 'text': "So when you're taking a part of that again and you're saying that I want to test it, obviously you will get good results.", 'start': 16272.642, 'duration': 6.225}, {'end': 16281.368, 'text': 'So that is not a very accurate way of testing.', 'start': 16278.907, 'duration': 2.461}], 'summary': 'Model training and testing should avoid exposing entire labeled data to ensure accurate results.', 'duration': 65.78, 'max_score': 16215.588, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM16215588.jpg'}, {'end': 16589.973, 'src': 'embed', 'start': 16560.033, 'weight': 4, 'content': [{'end': 16561.433, 'text': 'So that is the easiest way to do.', 'start': 16560.033, 'duration': 1.4}, {'end': 16569.895, 'text': "But then the downside is, as I said in the first case, if let's say 50% of your data is like that because some column or the other is missing.", 'start': 16561.613, 'duration': 8.282}, {'end': 16573.716, 'text': 'So it is not like in every place, in every row, the same column is missing.', 'start': 16569.935, 'duration': 3.781}, {'end': 16582.305, 'text': 'but you have in maybe 10 of the records, column one is missing and another 10 column two is missing, another 10 column three is missing,', 'start': 16574.036, 'duration': 8.269}, {'end': 16583.045, 'text': 'and so on and so forth.', 'start': 16582.305, 'duration': 0.74}, {'end': 16585.268, 'text': 'so it adds up to maybe half of your data set.', 'start': 16583.045, 'duration': 2.223}, {'end': 16588.111, 'text': 'so you cannot completely remove half of your data set.', 'start': 16585.268, 'duration': 2.843}, {'end': 16589.973, 'text': 'then the whole purpose is lost.', 'start': 16588.111, 'duration': 1.862}], 'summary': 'Half of the dataset has missing data, impacting analysis.', 'duration': 29.94, 'max_score': 16560.033, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM16560033.jpg'}, {'end': 16968.11, 'src': 'embed', 'start': 16937.562, 'weight': 1, 'content': [{'end': 16941.844, 'text': 'and maybe it may be a good idea to be prepared with an example so that it becomes easy for you.', 'start': 16937.562, 'duration': 4.282}, {'end': 16944.686, 'text': "you don't have to calculate these numbers on the fly, right.", 'start': 16941.844, 'duration': 2.842}, {'end': 16950.682, 'text': 'so a couple of hints are that you take some numbers which are with which add up to 100.', 'start': 16945.154, 'duration': 5.528}, {'end': 16951.904, 'text': 'that is always a good idea.', 'start': 16950.682, 'duration': 1.222}, {'end': 16954.628, 'text': "so you don't have to really do this complex calculations.", 'start': 16951.904, 'duration': 2.724}, {'end': 16956.111, 'text': 'so the total value will be 100.', 'start': 16954.628, 'duration': 1.483}, {'end': 16957.192, 'text': 'and then diagonal values.', 'start': 16956.111, 'duration': 1.081}, {'end': 16959.235, 'text': 'you divide once you find the diagonal values.', 'start': 16957.192, 'duration': 2.043}, {'end': 16960.758, 'text': 'that is equal to your percentage.', 'start': 16959.235, 'duration': 1.523}, {'end': 16961.945, 'text': 'okay, All right.', 'start': 16960.758, 'duration': 1.187}, {'end': 16968.11, 'text': 'so the next question can be a related question about false positive and false negative.', 'start': 16961.945, 'duration': 6.165}], 'summary': 'Suggest using numbers adding up to 100 for easy calculations. diagonal values equal percentage. next topic: false positive and false negative.', 'duration': 30.548, 'max_score': 16937.562, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM16937562.jpg'}, {'end': 17281.873, 'src': 'embed', 'start': 17255.518, 'weight': 0, 'content': [{'end': 17260.242, 'text': 'And then you use the test data set to check the accuracy, whether it is working fine or not.', 'start': 17255.518, 'duration': 4.724}, {'end': 17261.263, 'text': 'So you test the model.', 'start': 17260.262, 'duration': 1.001}, {'end': 17267.869, 'text': "before you actually put it into production, right? So once you test the model, you're satisfied, it's working fine.", 'start': 17261.403, 'duration': 6.466}, {'end': 17271.533, 'text': 'Then you go to the next level, which is putting it for production.', 'start': 17268.009, 'duration': 3.524}, {'end': 17273.875, 'text': 'And then in production, obviously new data will come.', 'start': 17271.713, 'duration': 2.162}, {'end': 17281.873, 'text': 'inference happens, so the model is readily available and only thing that happens is new data comes and the model predicts the values,', 'start': 17275.307, 'duration': 6.566}], 'summary': 'Test model with data set for accuracy before production.', 'duration': 26.355, 'max_score': 17255.518, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM17255518.jpg'}], 'start': 16215.588, 'title': 'Machine learning concepts and process', 'summary': 'Covers the explanation of false positive and false negative in a confusion matrix, the steps involved in the machine learning process, and the methodology for developing a machine learning model, emphasizing the importance of model training and testing, with a focus on classifier selection and confusion matrix accuracy calculation.', 'chapters': [{'end': 16713.205, 'start': 16215.588, 'title': 'Model training, testing, and deployment', 'summary': 'Explains the process of model training, testing, and deployment, emphasizing the importance of splitting labeled data into training and test sets, multiple passes through the training data, ratio preferences for splitting data, handling missing data, and choosing a classifier based on training set size.', 'duration': 497.617, 'highlights': ['The process of splitting labeled data into training and test sets is crucial for accurate model testing, with the training process involving multiple passes through the data to achieve high accuracy and minimize errors.', 'Ratio preferences for splitting training and test data vary from individual to individual, with options ranging from 50-50 to 70-30 or even more specific ratios like 63.33 and 33.', 'Handling missing data involves considering the percentage of missing values, the criticality of the situation, and the effort required to fix the data, with options including removing records or filling missing values with mean, minimum, or maximum values.', "Choosing a classifier based on training set data size may not have a one-size-fits-all answer, and it's recommended to try out multiple classifiers and decide based on the specific situation."]}, {'end': 16937.562, 'start': 16713.205, 'title': 'Understanding confusion matrix in classification', 'summary': 'Explains the importance of testing multiple classifiers to determine the most accurate one, provides a detailed explanation of confusion matrix including its components and how to calculate accuracy, with an example demonstrating 85% accuracy.', 'duration': 224.357, 'highlights': ['The chapter emphasizes the necessity of testing multiple classifiers to determine the most accurate one before making a decision.', 'A detailed explanation of confusion matrix, including the illustration of its components and the importance of the values across the diagonal for accuracy assessment.', 'A step-by-step demonstration of calculating accuracy using the confusion matrix, resulting in an 85% accuracy for the given example.']}, {'end': 17273.875, 'start': 16937.562, 'title': 'Machine learning concepts and process', 'summary': 'Covers the explanation of false positive and false negative in a confusion matrix, the steps involved in the machine learning process, and the methodology for developing a machine learning model, emphasizing the importance of model training and testing.', 'duration': 336.313, 'highlights': ['Explaining false positive and false negative in a confusion matrix, illustrating with examples and providing a clear distinction between the terms, emphasizing the importance of understanding these concepts in machine learning applications.', 'Detailing the three key steps in the process of developing a machine learning model, including understanding the problem, selecting algorithms, training, testing, and putting the model into production, highlighting the significance of model evaluation and testing before deployment.', 'Providing guidance on selecting suitable algorithms, emphasizing the trial and error process involved in algorithm selection and the importance of training the model using supervised learning techniques for accuracy assessment and validation.']}, {'end': 17727.657, 'start': 17275.307, 'title': 'Machine learning iterative process & deep learning explanation', 'summary': 'Discusses the iterative process of machine learning, the differences between machine learning and deep learning, and real-life applications of supervised machine learning in modern business, including examples such as email spam detection and healthcare diagnostics.', 'duration': 452.35, 'highlights': ['Machine learning involves an iterative process of training, testing, and deployment, which may require retraining due to failures in testing or deployment, or deterioration in accuracy over time.', 'Deep learning is a subset of machine learning that primarily utilizes neural networks for training and classification, and differs from machine learning in its automatic feature engineering.', 'The differences between machine learning and deep learning include the manual feature engineering in machine learning, as opposed to automatic feature engineering in deep learning, and the use of neural networks in deep learning.', 'Real-life applications of supervised machine learning include email spam detection, which involves training the model with historical email data labeled as spam or not spam.', 'Supervised machine learning is also utilized in healthcare diagnostics, where models are trained to detect specific features from medical images.']}], 'duration': 1512.069, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM16215588.jpg', 'highlights': ['The process of splitting labeled data into training and test sets is crucial for accurate model testing, with the training process involving multiple passes through the data to achieve high accuracy and minimize errors.', 'A step-by-step demonstration of calculating accuracy using the confusion matrix, resulting in an 85% accuracy for the given example.', 'Explaining false positive and false negative in a confusion matrix, illustrating with examples and providing a clear distinction between the terms, emphasizing the importance of understanding these concepts in machine learning applications.', "Choosing a classifier based on training set data size may not have a one-size-fits-all answer, and it's recommended to try out multiple classifiers and decide based on the specific situation.", 'Detailing the three key steps in the process of developing a machine learning model, including understanding the problem, selecting algorithms, training, testing, and putting the model into production, highlighting the significance of model evaluation and testing before deployment.', 'Machine learning involves an iterative process of training, testing, and deployment, which may require retraining due to failures in testing or deployment, or deterioration in accuracy over time.', 'The chapter emphasizes the necessity of testing multiple classifiers to determine the most accurate one before making a decision.', 'Handling missing data involves considering the percentage of missing values, the criticality of the situation, and the effort required to fix the data, with options including removing records or filling missing values with mean, minimum, or maximum values.', 'Real-life applications of supervised machine learning include email spam detection, which involves training the model with historical email data labeled as spam or not spam.', 'Supervised machine learning is also utilized in healthcare diagnostics, where models are trained to detect specific features from medical images.']}, {'end': 18510.03, 'segs': [{'end': 17875.112, 'src': 'embed', 'start': 17839.379, 'weight': 0, 'content': [{'end': 17843.34, 'text': 'So all your data for training your model has to be labeled.', 'start': 17839.379, 'duration': 3.961}, {'end': 17847.921, 'text': 'Now, this is a big problem in many industries or under many situations.', 'start': 17843.4, 'duration': 4.521}, {'end': 17853.482, 'text': "Getting the label data is not that easy because there's a lot of effort in labeling this data.", 'start': 17848.161, 'duration': 5.321}, {'end': 17857.063, 'text': "Let's take an example of diagnostic images.", 'start': 17853.922, 'duration': 3.141}, {'end': 17860.139, 'text': "we can just let's say take x-ray images.", 'start': 17857.677, 'duration': 2.462}, {'end': 17868.106, 'text': 'Now there are actually millions of x-ray images available all over the world, but the problem is they are not labeled.', 'start': 17860.319, 'duration': 7.787}, {'end': 17875.112, 'text': 'So their images are there, but whether it is effective or whether it is good that information is not available, along with that, right?', 'start': 17868.166, 'duration': 6.946}], 'summary': 'Labeling training data is a big challenge, e.g., x-ray images lack labels.', 'duration': 35.733, 'max_score': 17839.379, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM17839379.jpg'}, {'end': 17954.011, 'src': 'embed', 'start': 17900.469, 'weight': 1, 'content': [{'end': 17902.649, 'text': 'maybe a part of it is labeled.', 'start': 17900.469, 'duration': 2.18}, {'end': 17912.351, 'text': 'then we try some techniques to label the remaining part of the data so that we get completely labeled data and then we train model.', 'start': 17902.649, 'duration': 9.702}, {'end': 17921.195, 'text': 'So I know this is a little long winding explanation, but unfortunately there is no quick and easy definition for semi supervised machine learning.', 'start': 17912.411, 'duration': 8.784}, {'end': 17924.476, 'text': 'This is the only way probably to explain this concept.', 'start': 17921.215, 'duration': 3.261}, {'end': 17927.718, 'text': 'We may have another question,', 'start': 17925.377, 'duration': 2.341}, {'end': 17937.541, 'text': 'as what are unsupervised machine learning techniques or what are some of the techniques used for performing unsupervised machine learning?', 'start': 17927.718, 'duration': 9.823}, {'end': 17942.206, 'text': 'so it can be worded in So how do we answer this question?', 'start': 17937.541, 'duration': 4.665}, {'end': 17944.287, 'text': 'So, unsupervised learning?', 'start': 17942.626, 'duration': 1.661}, {'end': 17948.669, 'text': 'you can say that there are two types clustering and association.', 'start': 17944.287, 'duration': 4.382}, {'end': 17954.011, 'text': 'And clustering is a technique where similar objects are put together.', 'start': 17949.089, 'duration': 4.922}], 'summary': 'Semi-supervised machine learning uses techniques to label data and train models without quick definition.', 'duration': 53.542, 'max_score': 17900.469, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM17900469.jpg'}, {'end': 18129.109, 'src': 'embed', 'start': 18101.534, 'weight': 3, 'content': [{'end': 18109.999, 'text': "So, whenever you're doing some explanation, try as much as possible, as I said, to give examples from your work experience or give some analogies,", 'start': 18101.534, 'duration': 8.465}, {'end': 18115.783, 'text': 'and that will also help a lot in explaining as well and for the interviewer also to understand.', 'start': 18109.999, 'duration': 5.784}, {'end': 18119.644, 'text': "so here we'll take an example, or rather we will use an analogy.", 'start': 18115.783, 'duration': 3.861}, {'end': 18129.109, 'text': 'so inductive training is when we induce some knowledge or the learning process into a person without the person actually experiencing it.', 'start': 18119.644, 'duration': 9.465}], 'summary': 'Use work examples and analogies for effective explanations, like in inductive training.', 'duration': 27.575, 'max_score': 18101.534, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM18101534.jpg'}], 'start': 17727.657, 'title': 'Machine learning concepts and algorithms', 'summary': 'Delves into supervised, unsupervised, and semi-supervised machine learning, covering labeled data, healthcare diagnostics, and challenges in obtaining labeled data. it also explores inductive and deductive machine learning, knn vs k-means clustering, and reinforcement learning with relevant examples and applications.', 'chapters': [{'end': 18064.086, 'start': 17727.657, 'title': 'Supervised, unsupervised, and semi-supervised machine learning', 'summary': 'Explores supervised, unsupervised, and semi-supervised machine learning, discussing labeled data, healthcare diagnostics, and the challenges of obtaining labeled data, with examples of classification and clustering techniques.', 'duration': 336.429, 'highlights': ['Supervised machine learning uses labeled data to categorize images for healthcare diagnostics, with the goal of predicting illnesses such as cancer.', 'Semi-supervised learning bridges the gap between supervised and unsupervised learning by addressing the challenge of obtaining labeled data, with techniques to label a large portion of the available data.', 'Unsupervised learning encompasses clustering and association techniques, with clustering involving grouping similar objects and association identifying links between items, as demonstrated in e-commerce scenarios.']}, {'end': 18510.03, 'start': 18064.086, 'title': 'Machine learning concepts and algorithms', 'summary': 'Discusses inductive and deductive machine learning, knn vs k-means clustering, and reinforcement learning with examples and explanations, highlighting the differences and applications of each concept.', 'duration': 445.944, 'highlights': ['Inductive learning involves inducing knowledge into a person without the person experiencing it, while deductive learning involves drawing conclusions from experience.', 'KNN is a classification process under supervised learning, while K-means clustering is an unsupervised clustering process.', 'Naive base classifier is a probability-based classifier making assumptions about the independence of features, despite potential interrelatedness.', 'Reinforcement learning involves an environment and an agent performing actions to achieve a goal, such as playing a game to score high or maintain a high number of lives.']}], 'duration': 782.373, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM17727657.jpg', 'highlights': ['Supervised machine learning uses labeled data for healthcare diagnostics, predicting illnesses like cancer.', 'Semi-supervised learning addresses challenges in obtaining labeled data, labeling a large portion of available data.', 'Unsupervised learning involves clustering and association techniques, demonstrated in e-commerce scenarios.', 'Reinforcement learning involves an agent performing actions to achieve a goal, such as playing a game.', 'Inductive learning involves inducing knowledge into a person without the person experiencing it.', 'Deductive learning involves drawing conclusions from experience.', 'K-means clustering is an unsupervised clustering process.', 'KNN is a classification process under supervised learning.', 'Naive base classifier is a probability-based classifier making assumptions about the independence of features.']}, {'end': 19802.301, 'segs': [{'end': 19164.758, 'src': 'embed', 'start': 19142.084, 'weight': 2, 'content': [{'end': 19149.427, 'text': 'And then they load up that neural network into, in this case I have a Pixel 2, which actually has a built-in neural network for processing pictures.', 'start': 19142.084, 'duration': 7.343}, {'end': 19153.81, 'text': "And so it's just the forward propagation I use when it processes my photos.", 'start': 19149.727, 'duration': 4.083}, {'end': 19157.993, 'text': 'but when they were training it it used the back propagation to train it with the errors they had.', 'start': 19153.81, 'duration': 4.183}, {'end': 19160.775, 'text': "We'll be coming back to different models that are used.", 'start': 19158.353, 'duration': 2.422}, {'end': 19164.758, 'text': 'For right now, though, multi-layer perceptron, MLP.', 'start': 19161.015, 'duration': 3.743}], 'summary': 'Pixel 2 uses built-in neural network for processing photos, trained with back propagation and used for forward propagation.', 'duration': 22.674, 'max_score': 19142.084, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM19142084.jpg'}, {'end': 19556.577, 'src': 'embed', 'start': 19529.106, 'weight': 0, 'content': [{'end': 19534.791, 'text': "So what is the cost function? Cost function is a measure to evaluate how good your model's performance is.", 'start': 19529.106, 'duration': 5.685}, {'end': 19542.013, 'text': 'It is also referred to as loss or error, used to compute the error of the output layer during back propagation.', 'start': 19535.251, 'duration': 6.762}, {'end': 19544.374, 'text': "There's our back propagation where we're training our model.", 'start': 19542.133, 'duration': 2.241}, {'end': 19545.754, 'text': "That's one of our key words.", 'start': 19544.594, 'duration': 1.16}, {'end': 19549.875, 'text': 'Mean squared error is an example of a popular cost function.', 'start': 19546.114, 'duration': 3.761}, {'end': 19556.577, 'text': 'And so here we have the cost function C equals half of Y minus Y predicted, and then you square that.', 'start': 19550.335, 'duration': 6.242}], 'summary': 'The cost function measures model performance, using mean squared error as an example, expressed as c = 0.5 * (y - y_predicted)^2.', 'duration': 27.471, 'max_score': 19529.106, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM19529106.jpg'}, {'end': 19631.386, 'src': 'embed', 'start': 19603.428, 'weight': 1, 'content': [{'end': 19607.335, 'text': 'It might work great to test your data on what you have on your computer.', 'start': 19603.428, 'duration': 3.907}, {'end': 19608.917, 'text': "But that's different than in the field.", 'start': 19607.637, 'duration': 1.28}, {'end': 19614.239, 'text': "So, when we're talking about all these different tests and the error test as far as your loss,", 'start': 19609.338, 'duration': 4.901}, {'end': 19618.041, 'text': "you want to make sure that you're in a closed environment when you do initial testing,", 'start': 19614.239, 'duration': 3.802}, {'end': 19623.143, 'text': 'but you also want to open that up and make sure you follow up with the testing on the larger scale of data, because it will change.', 'start': 19618.041, 'duration': 5.102}, {'end': 19624.623, 'text': 'It might not fit the larger scale.', 'start': 19623.223, 'duration': 1.4}, {'end': 19631.386, 'text': 'There might be something in there in the way you brought the data in specifically or the data group you used or any of those could cause an error.', 'start': 19624.683, 'duration': 6.703}], 'summary': 'Initial testing in a closed environment may differ from larger scale testing due to potential data discrepancies and errors.', 'duration': 27.958, 'max_score': 19603.428, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM19603428.jpg'}, {'end': 19776.345, 'src': 'embed', 'start': 19751.069, 'weight': 4, 'content': [{'end': 19757.053, 'text': "We have our predicted Y coming out and then we have, since it's a training set, we already know the answer.", 'start': 19751.069, 'duration': 5.984}, {'end': 19758.655, 'text': 'And the answer comes back.', 'start': 19757.214, 'duration': 1.441}, {'end': 19765.56, 'text': 'and based on case of the square means, was one of the functions we looked at, one of the activation functions based on cost function.', 'start': 19758.655, 'duration': 6.905}, {'end': 19772.763, 'text': "That cost function then, depending on what you choose for your back propagation method, and there's a number of them, will change the weights.", 'start': 19765.7, 'duration': 7.063}, {'end': 19776.345, 'text': 'It will change the weight going to each one of those nodes in the hidden layer.', 'start': 19773.043, 'duration': 3.302}], 'summary': 'Training set predicts y, cost function alters weights in hidden layer.', 'duration': 25.276, 'max_score': 19751.069, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM19751069.jpg'}], 'start': 18510.03, 'title': 'Reinforcement learning applications', 'summary': 'Introduces reinforcement learning in industries like automotive, and game-playing systems like alphago. it discusses teaching machine learning through visual reinforcement and choosing algorithms, understanding multilayer perceptron in neural networks, and covers neural network basics, emphasizing data integrity. it also discusses testing data in closed and open environments, and the role of gradient descent in model optimization.', 'chapters': [{'end': 18579.066, 'start': 18510.03, 'title': 'Reinforcement learning in various industries', 'summary': 'Introduces reinforcement learning and its applications in industries, such as self-driving cars in the automotive sector, with examples of how the system works, and its use in creating game-playing systems like alphago.', 'duration': 69.036, 'highlights': ['Reinforcement learning is used in the automotive industry for self-driving cars, teaching the car how to navigate through roads using reinforcement learning.', 'The system in reinforcement learning consists of an agent and environment, where the agent is rewarded for taking steps towards the goal and penalized for steps going against it.', 'Reinforcement learning was used in creating game-playing systems like AlphaGo, where the system is allowed to learn and play games like chess.']}, {'end': 19015.051, 'start': 18579.246, 'title': 'Teaching machine learning and deep learning', 'summary': 'Discusses teaching a new learning system to play chess through visual reinforcement, choosing machine learning algorithms based on trial and error, and how recommendation engines work based on data collection and user profiling.', 'duration': 435.805, 'highlights': ['The system is taught to play chess through visual reinforcement by observing games and making random moves, then playing hundreds of thousands of games to learn on its own.', 'Choosing the right machine learning algorithm involves trying out a variety of algorithms, assessing their performance and accuracy, and then selecting the most suitable one.', 'Recommendation engines work by collecting data on customer behavior, associating and linking items through unsupervised learning, and profiling users based on demographic information to provide personalized recommendations.']}, {'end': 19245.105, 'start': 19015.071, 'title': 'Understanding multilayer perceptron and data normalization', 'summary': 'Introduces the concept of multilayer perceptron (mlp) in neural networks, highlighting its structure, backpropagation training, and capability to classify nonlinear classes. it also emphasizes the importance of data normalization for standardizing and rescaling values to improve convergence in neural networks.', 'duration': 230.034, 'highlights': ['Multilayer perceptron (MLP) has the same structure as a single-layer perceptron with one or more hidden layers, except the input layer, and uses a nonlinear activation function for each node in the hidden layers.', 'MLP employs supervised learning method called backpropagation for training the model, allowing the network to adjust its weights based on errors and improve accuracy.', 'Data normalization is crucial for standardizing and rescaling values to reduce redundancy and achieve better convergence in neural networks.']}, {'end': 19603.027, 'start': 19245.105, 'title': 'Neural network basics and data integrity', 'summary': 'Covers the basics of boltzmann machines, activation functions, and cost functions in neural networks, emphasizing the importance of data integrity and model evaluation, with examples and explanations.', 'duration': 357.922, 'highlights': ['The Boltzmann machine is a basic model with two layers making stochastic decisions, known as restricted Boltzmann machine', 'Activation functions in a neural network decide whether a neuron should be fired and play a crucial role in determining the output', "Cost functions measure the model's performance and are used to compute the error of the output layer during back propagation", 'Data integrity is crucial, ensuring clean data input and reliable output']}, {'end': 19802.301, 'start': 19603.428, 'title': 'Testing data and gradient descent', 'summary': "Discusses the importance of testing data in a closed and open environment, and emphasizes the role of gradient descent in minimizing error and optimizing the model's performance, with a brief overview of back propagation in neural network training.", 'duration': 198.873, 'highlights': ['Gradient descent is an optimization algorithm to minimize the cost function or to minimize the error, aiming to find the local or global minima of a function and determine the direction the model should take to reduce the error.', 'The chapter emphasizes the importance of testing data in both closed and open environments, highlighting the need to consider the local and global context of errors, and the potential impact of data scale on testing results.', 'Back propagation is described as a neural network technique to minimize the cost function, improve network performance, back propagate the error, and update weights to reduce the error.']}], 'duration': 1292.271, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM18510030.jpg', 'highlights': ['Reinforcement learning used in automotive for self-driving cars', 'Teaching machine learning through visual reinforcement', 'Choosing algorithms involves trying out a variety', 'Understanding multilayer perceptron in neural networks', 'Testing data in closed and open environments', 'Role of gradient descent in model optimization', 'Neural network basics emphasize data integrity']}, {'end': 20995.562, 'segs': [{'end': 20280.859, 'src': 'embed', 'start': 20254.536, 'weight': 7, 'content': [{'end': 20258.36, 'text': 'So it says my x value is going to be somewhere between 0 or 1.', 'start': 20254.536, 'duration': 3.824}, {'end': 20263.605, 'text': "And then usually, unless it's really uncertain, the output is usually a 1 or 0.", 'start': 20258.36, 'duration': 5.245}, {'end': 20270.031, 'text': "And then you have that little piece of uncertainty there that you can send forward to another network or you can look at to know that there's uncertainty involved.", 'start': 20263.605, 'duration': 6.426}, {'end': 20272.653, 'text': "And it's often used in the hidden layers.", 'start': 20270.191, 'duration': 2.462}, {'end': 20276.035, 'text': "This is what's coming out of the hidden layers into the output layer usually.", 'start': 20272.933, 'duration': 3.102}, {'end': 20280.859, 'text': 'Or as we referenced the convolutional neural network, the CNN.', 'start': 20276.376, 'duration': 4.483}], 'summary': 'X value between 0-1, output usually 1 or 0, uncertainty used in hidden layers.', 'duration': 26.323, 'max_score': 20254.536, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM20254536.jpg'}, {'end': 20514.524, 'src': 'embed', 'start': 20484.663, 'weight': 6, 'content': [{'end': 20486.205, 'text': 'They flip back and forth too easy.', 'start': 20484.663, 'duration': 1.542}, {'end': 20491.609, 'text': "And you see down here we've introduced two new terms, converge and diverge.", 'start': 20486.445, 'duration': 5.164}, {'end': 20500.114, 'text': "A converge means that our model has reached a point where it's able to give a fairly good answer for all the data we put in.", 'start': 20492.149, 'duration': 7.965}, {'end': 20502.936, 'text': "All those weights have adjusted and it's minimized the error.", 'start': 20500.234, 'duration': 2.702}, {'end': 20508.68, 'text': 'Diverge means that the data is so chaotic that it can never manage to train to that data.', 'start': 20503.196, 'duration': 5.484}, {'end': 20510.561, 'text': 'The data is just too chaotic for it to train.', 'start': 20508.72, 'duration': 1.841}, {'end': 20514.524, 'text': 'So we have two new words there, converge and diverge are important to know also.', 'start': 20510.661, 'duration': 3.863}], 'summary': 'Understanding the concepts of converge and diverge is essential for model training.', 'duration': 29.861, 'max_score': 20484.663, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM20484663.jpg'}, {'end': 20750.915, 'src': 'embed', 'start': 20725.728, 'weight': 2, 'content': [{'end': 20731.934, 'text': 'It usually happens when there is less and improper data to train a model, has a poor performance and accuracy.', 'start': 20725.728, 'duration': 6.206}, {'end': 20736.818, 'text': "So if you're using underfitted data and you generate a model and you distribute that in a commercial zone,", 'start': 20732.114, 'duration': 4.704}, {'end': 20740.422, 'text': "you'll have a lot of people unhappy with you, because it's not going to give them very good answers.", 'start': 20736.818, 'duration': 3.604}, {'end': 20745.828, 'text': "So we've explained overfitting and underfitting, so now we want to ask how to combat them.", 'start': 20740.862, 'duration': 4.966}, {'end': 20748.091, 'text': 'Combating overfitting and underfitting.', 'start': 20745.989, 'duration': 2.102}, {'end': 20750.915, 'text': 'Resampling the data to estimate the model accuracy.', 'start': 20748.251, 'duration': 2.664}], 'summary': 'Insufficient data leads to poor model performance. combat overfitting and underfitting by resampling data for accurate model estimation.', 'duration': 25.187, 'max_score': 20725.728, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM20725728.jpg'}, {'end': 20992.621, 'src': 'heatmap', 'start': 20787.894, 'weight': 0.844, 'content': [{'end': 20794.038, 'text': 'All the neurons in every layer perform the same operation given the same output and making the deep net useless.', 'start': 20787.894, 'duration': 6.144}, {'end': 20795.279, 'text': 'Right there is a key word.', 'start': 20794.339, 'duration': 0.94}, {'end': 20797.641, 'text': "It's going to be useless if you initialize everything to zero.", 'start': 20795.319, 'duration': 2.322}, {'end': 20801.203, 'text': 'At that point, be looking into some other machine learning tools.', 'start': 20797.821, 'duration': 3.382}, {'end': 20803.165, 'text': 'Initializing all weights randomly.', 'start': 20801.504, 'duration': 1.661}, {'end': 20807.169, 'text': 'Here the weights are assigned randomly by initializing them very close to zero.', 'start': 20803.365, 'duration': 3.804}, {'end': 20811.893, 'text': 'It gives better accuracy to the model since every neuron performs different computations.', 'start': 20807.309, 'duration': 4.584}, {'end': 20814.496, 'text': 'And here we have the weights are set randomly.', 'start': 20812.314, 'duration': 2.182}, {'end': 20817.058, 'text': 'We have our input layer, the hidden layers, and the output layer.', 'start': 20814.616, 'duration': 2.442}, {'end': 20819.381, 'text': 'And W equals NP random, random in.', 'start': 20817.259, 'duration': 2.122}, {'end': 20822.143, 'text': 'Layer size L, layer size L minus 1.', 'start': 20819.501, 'duration': 2.642}, {'end': 20825.687, 'text': 'This is the most commonly used, is to randomly generate your weights.', 'start': 20822.143, 'duration': 3.544}, {'end': 20835.156, 'text': 'What are the different layers in CNN, convolutional neural network? First is the convolutional layer that performs a convolutional operation.', 'start': 20826.107, 'duration': 9.049}, {'end': 20838.199, 'text': 'We have our other video out if you want to explore that more,', 'start': 20835.496, 'duration': 2.703}, {'end': 20847.028, 'text': 'so you can go into detail exactly how the convolutional layer works in the CNN as far as creating a number of smaller picture windows to go over the data.', 'start': 20838.199, 'duration': 8.829}, {'end': 20849.871, 'text': 'The second step is as a ReLU layer.', 'start': 20847.589, 'duration': 2.282}, {'end': 20854.577, 'text': 'ReLU brings non-linearity to the network and converts all the negative pixels to zero.', 'start': 20850.031, 'duration': 4.546}, {'end': 20856.839, 'text': 'Output is rectified feature map.', 'start': 20854.817, 'duration': 2.022}, {'end': 20858.821, 'text': 'So it goes into a mapping feature there.', 'start': 20857.099, 'duration': 1.722}, {'end': 20859.722, 'text': 'Pooling layer.', 'start': 20858.982, 'duration': 0.74}, {'end': 20864.468, 'text': 'Pooling is a down sampling operation that reduces the dimensionality of the feature map.', 'start': 20859.803, 'duration': 4.665}, {'end': 20869.651, 'text': 'So we have all our ReLU layer, which is pulling all these little maps out of our convolutional layer.', 'start': 20864.708, 'duration': 4.943}, {'end': 20874.333, 'text': "It's taking that picture and creating little tiny neural networks to look at different parts of the picture.", 'start': 20869.791, 'duration': 4.542}, {'end': 20878.695, 'text': 'Then we need to pool it together, and then finally the fully connected layer.', 'start': 20874.873, 'duration': 3.822}, {'end': 20884.958, 'text': 'So we flatten our pooling layer out, and we have a fully connected layer recognizes and classifies the objects in the image.', 'start': 20878.775, 'duration': 6.183}, {'end': 20889.921, 'text': "And that's actually your forward propagation, reverse propagation training model, usually.", 'start': 20885.218, 'duration': 4.703}, {'end': 20891.862, 'text': "I mean, there's a number of different models out there, of course.", 'start': 20889.941, 'duration': 1.921}, {'end': 20895.563, 'text': 'What is pooling in CNN and how does it work?', 'start': 20892.262, 'duration': 3.301}, {'end': 20903.286, 'text': 'Pooling used to reduce the spatial dimensions of a CNN performs downsampling operation to reduce the dimensionality,', 'start': 20896.023, 'duration': 7.263}, {'end': 20907.868, 'text': 'creates a pooled feature map by sliding a filter matrix over the input matrix.', 'start': 20903.286, 'duration': 4.582}, {'end': 20910.363, 'text': 'I mentioned that briefly on the previous slide.', 'start': 20908.462, 'duration': 1.901}, {'end': 20914.746, 'text': "It's important to know that you have, if you can see here, they have a rectified feature map.", 'start': 20910.903, 'duration': 3.843}, {'end': 20921.75, 'text': 'And so each one of those colors, like the yellow color, that might be one of the smaller little neural network using the ReLU.', 'start': 20915.046, 'duration': 6.704}, {'end': 20927.153, 'text': "It'll look at, it'll just kind of go over the main picture and look at all the different areas on the main picture.", 'start': 20921.93, 'duration': 5.223}, {'end': 20929.674, 'text': 'So you might step one, two, three, four spaces.', 'start': 20927.393, 'duration': 2.281}, {'end': 20932.556, 'text': "And then you have another one that's also looking at features.", 'start': 20930.094, 'duration': 2.462}, {'end': 20933.476, 'text': 'And it has a 2785.', 'start': 20932.976, 'duration': 0.5}, {'end': 20934.197, 'text': 'Each one of those is a map.', 'start': 20933.476, 'duration': 0.721}, {'end': 20941.883, 'text': 'So the first one might be a map looking for cat ears, and the second one looking for human eyes.', 'start': 20936.978, 'duration': 4.905}, {'end': 20946.528, 'text': 'When it does this, you then have this rectified feature map looking at these different features.', 'start': 20942.343, 'duration': 4.185}, {'end': 20950.011, 'text': 'And the max pooling with a 2x2 filter is in a stride of 2.', 'start': 20946.848, 'duration': 3.163}, {'end': 20953.174, 'text': "Stride means instead of skipping every pixel, you're going to go every 2 pixels.", 'start': 20950.011, 'duration': 3.163}, {'end': 20954.936, 'text': 'You take the maximum values.', 'start': 20953.394, 'duration': 1.542}, {'end': 20960.982, 'text': 'And you can see over here when we look at a pooled feature map, one of the features says, hey, I had a max value of 8.', 'start': 20955.256, 'duration': 5.726}, {'end': 20965.527, 'text': 'So somewhere in here we saw a human eye labeled as 8, pretty high label.', 'start': 20960.982, 'duration': 4.545}, {'end': 20971.173, 'text': 'And maybe 7 was a human hand and maybe 4 was cat whiskers or something that we thought might be cat whiskers.', 'start': 20965.667, 'duration': 5.506}, {'end': 20974.257, 'text': '4 is kind of a low number in this particular case compared to the other ones.', 'start': 20971.534, 'duration': 2.723}, {'end': 20976.317, 'text': 'So you have your full pooled feature map.', 'start': 20974.477, 'duration': 1.84}, {'end': 20979.238, 'text': 'You can see the process here as we have our stepping.', 'start': 20976.337, 'duration': 2.901}, {'end': 20983.699, 'text': 'We look for the max value and then we create a pooled feature map of the maxed values.', 'start': 20979.258, 'duration': 4.441}, {'end': 20985.72, 'text': 'I hope this helps you in your interview.', 'start': 20984.119, 'duration': 1.601}, {'end': 20989.521, 'text': "With that, we've reached the end of the complete artificial intelligence course.", 'start': 20986.04, 'duration': 3.481}, {'end': 20991.261, 'text': 'I hope you enjoyed this video.', 'start': 20989.941, 'duration': 1.32}, {'end': 20992.621, 'text': 'Do like and share it.', 'start': 20991.621, 'duration': 1}], 'summary': 'Random weight initialization improves accuracy in deep nets, while cnn involves convolutional, relu, pooling, and fully connected layers for image recognition.', 'duration': 204.727, 'max_score': 20787.894, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM20787894.jpg'}, {'end': 20825.687, 'src': 'embed', 'start': 20797.821, 'weight': 0, 'content': [{'end': 20801.203, 'text': 'At that point, be looking into some other machine learning tools.', 'start': 20797.821, 'duration': 3.382}, {'end': 20803.165, 'text': 'Initializing all weights randomly.', 'start': 20801.504, 'duration': 1.661}, {'end': 20807.169, 'text': 'Here the weights are assigned randomly by initializing them very close to zero.', 'start': 20803.365, 'duration': 3.804}, {'end': 20811.893, 'text': 'It gives better accuracy to the model since every neuron performs different computations.', 'start': 20807.309, 'duration': 4.584}, {'end': 20814.496, 'text': 'And here we have the weights are set randomly.', 'start': 20812.314, 'duration': 2.182}, {'end': 20817.058, 'text': 'We have our input layer, the hidden layers, and the output layer.', 'start': 20814.616, 'duration': 2.442}, {'end': 20819.381, 'text': 'And W equals NP random, random in.', 'start': 20817.259, 'duration': 2.122}, {'end': 20822.143, 'text': 'Layer size L, layer size L minus 1.', 'start': 20819.501, 'duration': 2.642}, {'end': 20825.687, 'text': 'This is the most commonly used, is to randomly generate your weights.', 'start': 20822.143, 'duration': 3.544}], 'summary': 'Using random weight initialization for better accuracy in machine learning models.', 'duration': 27.866, 'max_score': 20797.821, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM20797821.jpg'}, {'end': 20921.75, 'src': 'embed', 'start': 20896.023, 'weight': 1, 'content': [{'end': 20903.286, 'text': 'Pooling used to reduce the spatial dimensions of a CNN performs downsampling operation to reduce the dimensionality,', 'start': 20896.023, 'duration': 7.263}, {'end': 20907.868, 'text': 'creates a pooled feature map by sliding a filter matrix over the input matrix.', 'start': 20903.286, 'duration': 4.582}, {'end': 20910.363, 'text': 'I mentioned that briefly on the previous slide.', 'start': 20908.462, 'duration': 1.901}, {'end': 20914.746, 'text': "It's important to know that you have, if you can see here, they have a rectified feature map.", 'start': 20910.903, 'duration': 3.843}, {'end': 20921.75, 'text': 'And so each one of those colors, like the yellow color, that might be one of the smaller little neural network using the ReLU.', 'start': 20915.046, 'duration': 6.704}], 'summary': 'Pooling in cnn reduces dimensions, creates feature maps using filter matrix, and utilizes rectified feature maps for neural networks.', 'duration': 25.727, 'max_score': 20896.023, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM20896023.jpg'}], 'start': 19802.421, 'title': 'Neural network architectures and techniques', 'summary': 'Delves into the differences between feed-forward and recurrent neural networks, their applications in sentiment analysis, text mining, image captioning, and time series problems, and the impact of hyperparameters and techniques such as dropout, batch normalization, gradient descent, overfitting, underfitting, weight initialization, cnn layers, and pooling.', 'chapters': [{'end': 19929.409, 'start': 19802.421, 'title': 'Neural networks: feed-forward vs recurrent', 'summary': 'Explores the differences between feed-forward and recurrent neural networks, emphasizing their directional signal flow, feedback loops, and ability to memorize past inputs, with a focus on the cnn as an example of a feed-forward neural network.', 'duration': 126.988, 'highlights': ['Recurrent neural network signals travel in both directions, making it a looped network, and has the ability to memorize past data due to its internal memory.', 'Feed-forward neural network signals travel in one direction from input to output, with no feedback loops, and considers only the current input.', 'CNN is an example of a feed-forward neural network, where information goes forward and it splits the picture apart to process individual pixels before categorizing the output.']}, {'end': 20342.166, 'start': 19929.629, 'title': 'Applications of recurrent neural network', 'summary': 'Discusses the architecture and applications of recurrent neural network (rnn), touching on its usage in sentiment analysis, text mining, image captioning, and time series problems like stock price prediction, with a focus on the activation functions relu and softmax.', 'duration': 412.537, 'highlights': ['RNN applications in sentiment analysis and text mining', 'RNN applications in image captioning', 'RNN applications in time series problems like stock price prediction', 'Explanation of Softmax activation function', 'Explanation of ReLU activation function']}, {'end': 20995.562, 'start': 20342.326, 'title': 'Neural network hyperparameters & techniques', 'summary': 'Explains neural network hyperparameters such as hidden units, layers, learning rate, and epochs, and their impacts, including the effects of setting learning rate too low or too high. it also covers techniques like dropout, batch normalization, gradient descent, overfitting, underfitting, weight initialization, cnn layers, and pooling in cnn.', 'duration': 653.236, 'highlights': ['Explaining the impact of setting learning rate too low or too high', 'Describing the techniques of dropout and batch normalization to prevent overfitting and improve network performance', 'Explaining the differences between batch gradient descent and stochastic gradient descent', 'Defining overfitting and underfitting and how to combat them', 'Detailing weight initialization in a network']}], 'duration': 1193.141, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8Pyy2d3SZuM/pics/8Pyy2d3SZuM19802421.jpg', 'highlights': ['RNN applications in time series problems like stock price prediction', 'RNN applications in image captioning', 'RNN applications in sentiment analysis and text mining', 'CNN is an example of a feed-forward neural network, processing individual pixels before categorizing the output', 'Recurrent neural network signals travel in both directions, making it a looped network and has the ability to memorize past data', 'Feed-forward neural network signals travel in one direction from input to output, with no feedback loops, and considers only the current input', 'Explaining the impact of setting learning rate too low or too high', 'Describing the techniques of dropout and batch normalization to prevent overfitting and improve network performance', 'Detailing weight initialization in a network', 'Explaining the differences between batch gradient descent and stochastic gradient descent']}], 'highlights': ['AI applications in smartphones, cars, social media, banking, and surveillance are widespread.', 'Deep learning reshapes healthcare industry with early cancer detection and drug synthesis.', 'The linear regression model involves finding the slope and coefficient for minimizing error.', 'The importance of defining objectives before collecting data and selecting algorithms, as it is crucial to know what is being predicted.', 'The use of AI in banking for fraud detection, online customer support, cybersecurity, and virtual assistants is detailed.', 'The process of initializing and running a TensorFlow object involves bundling all the steps into one, initializing variables, and creating a session.', 'The potential applications of artificial intelligence and machine learning include the detection of crimes before they happen, increased efficiency in healthcare, and advancements in marketing techniques.', 'The accuracy achieved on the test data after training the neural network is 0.9165, indicating a good performance in recognizing symbols.', 'The process of splitting labeled data into training and test sets is crucial for accurate model testing, with the training process involving multiple passes through the data to achieve high accuracy and minimize errors.', 'Reinforcement learning used in automotive for self-driving cars']}