title
Biological versus Artificial Neural Networks (John Hopfield) | AI Podcast Clips

description
Full episode with John Hopfield (Feb 2020): https://www.youtube.com/watch?v=DKyzcbNr8WE Clips channel (Lex Clips): https://www.youtube.com/lexclips Main channel (Lex Fridman): https://www.youtube.com/lexfridman (more links below) Podcast full episodes playlist: https://www.youtube.com/playlist?list=PLrAXtmErZgOdP_8GztsuKi9nrraNbKKp4 Podcasts clips playlist: https://www.youtube.com/playlist?list=PLrAXtmErZgOeciFP3CBCIEElOJeitOr41 Podcast website: https://lexfridman.com/ai Podcast on Apple Podcasts (iTunes): https://apple.co/2lwqZIr Podcast on Spotify: https://spoti.fi/2nEwCF8 Podcast RSS: https://lexfridman.com/category/ai/feed/ John Hopfield is professor at Princeton, whose life's work weaved beautifully through biology, chemistry, neuroscience, and physics. Most crucially, he saw the messy world of biology through the piercing eyes of a physicist. He is perhaps best known for his work on associate neural networks, now known as Hopfield networks that were one of the early ideas that catalyzed the development of the modern field of deep learning. Subscribe to this YouTube channel or connect on: - Twitter: https://twitter.com/lexfridman - LinkedIn: https://www.linkedin.com/in/lexfridman - Facebook: https://www.facebook.com/lexfridman - Instagram: https://www.instagram.com/lexfridman - Medium: https://medium.com/@lexfridman - Support on Patreon: https://www.patreon.com/lexfridman

detail
{'title': 'Biological versus Artificial Neural Networks (John Hopfield) | AI Podcast Clips', 'heatmap': [], 'summary': 'Compares biological and artificial neural networks, detailing the influence of evolutionary biology, adaptation in the human mind, and understanding the mind through a physics lens, while also exploring ai evolution and neurobiology, presenting insightful case studies and examples.', 'chapters': [{'end': 161.557, 'segs': [{'end': 117.017, 'src': 'embed', 'start': 31.954, 'weight': 0, 'content': [{'end': 40.157, 'text': 'if you have some little quirk in how a molecule works or how a cell works and it can be made use of,', 'start': 31.954, 'duration': 8.203}, {'end': 45.659, 'text': 'evolution will sharpen it up and make it into a useful feature rather than a glitch.', 'start': 40.157, 'duration': 5.502}, {'end': 55.683, 'text': 'And so you expect, in neurobiology, for evolution to have captured all kinds of possibilities of getting neurons,', 'start': 46.94, 'duration': 8.743}, {'end': 58.104, 'text': 'of how you get neurons to do things for you.', 'start': 55.683, 'duration': 2.421}, {'end': 64.971, 'text': 'And that aspect has been completely suppressed in artificial neural networks.', 'start': 59.807, 'duration': 5.164}, {'end': 74.517, 'text': 'Do the glitches become features in the biological neural network? They can.', 'start': 66.852, 'duration': 7.665}, {'end': 78.54, 'text': 'Look, let me take one of the things that I used to do research on.', 'start': 74.537, 'duration': 4.003}, {'end': 81.843, 'text': 'If you take things which..', 'start': 78.56, 'duration': 3.283}, {'end': 88.94, 'text': 'They have rhythms which are sort of close to each other.', 'start': 84.798, 'duration': 4.142}, {'end': 96.343, 'text': 'Under some circumstances, these things will have a phase transition and suddenly the rhythm will, everybody will fall into step.', 'start': 89.34, 'duration': 7.003}, {'end': 107.734, 'text': 'There was a marvelous physical example of that in the Millennium Bridge across the Thames River built about 2001.', 'start': 97.204, 'duration': 10.53}, {'end': 112.475, 'text': "And pedestrians walking across, pedestrians don't walk synchronized.", 'start': 107.734, 'duration': 4.741}, {'end': 114.056, 'text': "They don't walk in lock step.", 'start': 112.515, 'duration': 1.541}, {'end': 117.017, 'text': 'But they all walk at about the same frequency.', 'start': 114.876, 'duration': 2.141}], 'summary': 'Evolution sharpens quirks into useful features in biological networks, unlike in artificial neural networks.', 'duration': 85.063, 'max_score': 31.954, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw31954.jpg'}], 'start': 1.71, 'title': 'Biological vs artificial neural networks', 'summary': "Delves into the contrast between biological and artificial neural networks, highlighting the influence of evolutionary biology on neurons and presenting a case study on pedestrians' walking behavior on the millennium bridge.", 'chapters': [{'end': 161.557, 'start': 1.71, 'title': 'Neural networks: biological vs artificial', 'summary': "Discusses the captivating difference between biological neural networks and artificial neural networks, emphasizing the role of evolutionary biology in shaping neurons and highlighting a real-life example of phase transition in pedestrians' walking behavior on the millennium bridge.", 'duration': 159.847, 'highlights': ["Evolutionary biology has shaped neurons in biological neural networks to capture all kinds of possibilities, turning quirks into useful features (quantifiable data: mentioned evolutionary biology's role in shaping neurons).", "The example of pedestrians' walking behavior on the Millennium Bridge demonstrates a phase transition phenomenon where pedestrians synchronized their steps due to the bridge's oscillation, highlighting the engineers' oversight in considering only the back and forth motion of walking (quantifiable data: real-life example of pedestrians synchronizing their steps on the Millennium Bridge).", 'The difference between biological neural networks and artificial neural networks lies in the suppression of the evolutionary aspect in the latter, where glitches can become features in biological neural networks (quantifiable data: comparison of evolutionary influence on neural networks).']}], 'duration': 159.847, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw1710.jpg', 'highlights': ['Evolutionary biology has shaped neurons in biological neural networks to capture all kinds of possibilities, turning quirks into useful features.', "The example of pedestrians' walking behavior on the Millennium Bridge demonstrates a phase transition phenomenon where pedestrians synchronized their steps due to the bridge's oscillation, highlighting the engineers' oversight in considering only the back and forth motion of walking.", 'The difference between biological neural networks and artificial neural networks lies in the suppression of the evolutionary aspect in the latter, where glitches can become features in biological neural networks.']}, {'end': 380.686, 'segs': [{'end': 241.394, 'src': 'embed', 'start': 185.557, 'weight': 0, 'content': [{'end': 189.459, 'text': 'Well, if they fire together, you can be sure that other cells are going to notice it.', 'start': 185.557, 'duration': 3.902}, {'end': 195.042, 'text': 'So you can make a computational feature out of this in an evolving brain.', 'start': 189.939, 'duration': 5.103}, {'end': 203.267, 'text': "Most artificial neural networks don't even have action potentials, let alone have the possibility for synchronizing them.", 'start': 196.943, 'duration': 6.324}, {'end': 208.669, 'text': 'And you mentioned the evolutionary process.', 'start': 205.848, 'duration': 2.821}, {'end': 221.471, 'text': 'The evolutionary process that builds on top of biological systems leverages that the weird mess of it somehow.', 'start': 208.69, 'duration': 12.781}, {'end': 230.22, 'text': 'So how do you make sense of that ability to leverage all the different kinds of complexities in the biological brain??', 'start': 221.932, 'duration': 8.288}, {'end': 241.394, 'text': "Well, look, in the biological molecule level, You'd have a piece of DNA which would encode for a particular protein.", 'start': 231.241, 'duration': 10.153}], 'summary': 'Synchronizing action potentials in artificial neural networks can be leveraged for computational features in an evolving brain.', 'duration': 55.837, 'max_score': 185.557, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw185557.jpg'}, {'end': 328.752, 'src': 'embed', 'start': 297.871, 'weight': 1, 'content': [{'end': 300.171, 'text': 'The evolutionary process looks a little different.', 'start': 297.871, 'duration': 2.3}, {'end': 304.392, 'text': 'Yeah Similar timescale, perhaps.', 'start': 300.691, 'duration': 3.701}, {'end': 305.792, 'text': 'Much shorter in timescale.', 'start': 304.452, 'duration': 1.34}, {'end': 308.773, 'text': 'Companies close, yeah, go bankrupt and are born.', 'start': 306.232, 'duration': 2.541}, {'end': 310.093, 'text': 'Yeah, shorter.', 'start': 309.233, 'duration': 0.86}, {'end': 311.773, 'text': 'But not much shorter.', 'start': 310.833, 'duration': 0.94}, {'end': 314.574, 'text': 'Some companies last a century.', 'start': 313.013, 'duration': 1.561}, {'end': 316.414, 'text': "But yeah, you're right.", 'start': 315.754, 'duration': 0.66}, {'end': 322.108, 'text': 'I mean if you think of companies as a single organism that builds, and you all know, yeah,', 'start': 317.725, 'duration': 4.383}, {'end': 328.752, 'text': "it's a fascinating dual correspondence there between biological.", 'start': 322.108, 'duration': 6.644}], 'summary': 'Evolution of companies mirrors biological processes, some lasting a century.', 'duration': 30.881, 'max_score': 297.871, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw297871.jpg'}], 'start': 161.637, 'title': 'Evolutionary process and biological systems', 'summary': 'Discusses the evolutionary process in biological systems, emphasizing the ability to leverage complexity and adaptability. it provides examples of action potentials and dna duplication, illustrating the differences between biological and computer-based evolution.', 'chapters': [{'end': 380.686, 'start': 161.637, 'title': 'Evolutionary process and biological systems', 'summary': 'Discusses the evolutionary process in biological systems, emphasizing the ability to leverage complexity and adaptability, with examples of action potentials and dna duplication, illustrating the differences between biological and computer-based evolution.', 'duration': 219.049, 'highlights': ['Biological systems leverage complexity and adaptability through DNA duplication and evolution, allowing for the retention of old functions and the development of new ones. The evolutionary process in biological systems leverages the ability to duplicate DNA and have molecules drift apart, retaining old functions while developing new ones, leading to evolutionary pressure for improvement.', 'Action potentials in nerve cells can synchronize, allowing for computational features in an evolving brain, a feature not present in most artificial neural networks. Nerve cells can produce action potentials and synchronize under certain circumstances, creating computational features in an evolving brain, which is not commonly observed in artificial neural networks.', 'Comparison between the evolutionary process in biological systems and the evolution of companies, noting differences in timescales and adaptability. The chapter draws parallels between the evolutionary process in biological systems and the evolution of companies, highlighting differences in timescales and adaptability, with companies facing challenges in integrating new products and adapting to new markets.']}], 'duration': 219.049, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw161637.jpg', 'highlights': ['Biological systems leverage complexity and adaptability through DNA duplication and evolution, allowing for the retention of old functions and the development of new ones.', 'Comparison between the evolutionary process in biological systems and the evolution of companies, noting differences in timescales and adaptability.', 'Action potentials in nerve cells can synchronize, allowing for computational features in an evolving brain, a feature not present in most artificial neural networks.']}, {'end': 659.361, 'segs': [{'end': 446.66, 'src': 'embed', 'start': 414.399, 'weight': 0, 'content': [{'end': 423.105, 'text': "where your learning goes on at the timescale of one individual who must learn from the environment during that individual's lifetime.", 'start': 414.399, 'duration': 8.706}, {'end': 428.565, 'text': 'And biology has both kinds of learning in it.', 'start': 425.763, 'duration': 2.802}, {'end': 442.817, 'text': "And the thing which makes neurobiology hard is that it's a mathematical system, as it were, built on this other kind of evolutionary system.", 'start': 430.167, 'duration': 12.65}, {'end': 446.66, 'text': 'What do you mean by mathematical system??', 'start': 445.159, 'duration': 1.501}], 'summary': 'Neurobiology involves learning from the environment, a mathematical system intertwined with an evolutionary one.', 'duration': 32.261, 'max_score': 414.399, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw414399.jpg'}, {'end': 512.28, 'src': 'embed', 'start': 480.719, 'weight': 1, 'content': [{'end': 484.84, 'text': 'but they really have difficulty doing three-dimensional wiring.', 'start': 480.719, 'duration': 4.121}, {'end': 496.964, 'text': 'Biology is the neocortex is actually also sheet-like and it sits on top of the white matter, which is about 10 times the volume of the gray matter,', 'start': 485.74, 'duration': 11.224}, {'end': 499.685, 'text': 'and contains all what you might call the wires.', 'start': 496.964, 'duration': 2.721}, {'end': 512.28, 'text': "But there's a huge, the effect of computer structure on what is easy and what is hard is immense.", 'start': 501.954, 'duration': 10.326}], 'summary': 'Neocortex is sheet-like, 10x white matter, struggle with 3d wiring.', 'duration': 31.561, 'max_score': 480.719, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw480719.jpg'}, {'end': 598.125, 'src': 'embed', 'start': 568.75, 'weight': 7, 'content': [{'end': 569.891, 'text': 'And which is more powerful?', 'start': 568.75, 'duration': 1.141}, {'end': 581.594, 'text': 'I find things most interesting that I begin to see how to get into the edges of them and tease them apart a little bit to see how they work.', 'start': 572.471, 'duration': 9.123}, {'end': 598.125, 'text': "And since I can't see the evolutionary process going on, I'm in awe of it, but I find it just a black hole as far as trying to understand what to do.", 'start': 583.942, 'duration': 14.183}], 'summary': 'Speaker finds evolutionary process awe-inspiring but difficult to understand and navigate.', 'duration': 29.375, 'max_score': 568.75, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw568750.jpg'}, {'end': 659.361, 'src': 'embed', 'start': 630.966, 'weight': 3, 'content': [{'end': 634.487, 'text': "the fact that you're building a system in three dimensions.", 'start': 630.966, 'duration': 3.521}, {'end': 642.511, 'text': 'In just days and months, those early days of a human life are really interesting.', 'start': 636.848, 'duration': 5.663}, {'end': 650.472, 'text': 'They are, and of course, there are times of immense cell multiplication.', 'start': 643.786, 'duration': 6.686}, {'end': 657.199, 'text': 'There are also times of the greatest cell death in the brain is during infancy.', 'start': 650.833, 'duration': 6.366}, {'end': 659.361, 'text': "It's turnover.", 'start': 658.64, 'duration': 0.721}], 'summary': 'Early human brain experiences rapid cell turnover in 3d system.', 'duration': 28.395, 'max_score': 630.966, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw630966.jpg'}], 'start': 380.686, 'title': 'Human mind adaptation', 'summary': "Discusses the importance of adaptation in the human mind, highlighting generational and individual learning differences, neurobiology's mathematical system, and contrasts evolutionary and human life scale adaptation, with emphasis on developmental neurobiology and dynamics of cell multiplication and death in early human life.", 'chapters': [{'end': 538.972, 'start': 380.686, 'title': 'The beauty of the human mind', 'summary': 'Discusses the importance of adaptation in the human mind, highlighting the differences between generational learning and individual learning, and the mathematical system underlying neurobiology and its comparison to computer structures.', 'duration': 158.286, 'highlights': ["The human mind's adaptive ability is the most beautiful aspect, encompassing both generational and individual learning, and neurobiology's mathematical system.", 'The differences between adaptation over generations and individual learning, which presents a challenge in neurobiology due to its mathematical system.', "The comparison of neurobiology's three-dimensional structure to computer chips' limitations and its impact on computational capabilities.", 'The significance of the three-dimensional structure of the neocortex and the white matter in enabling certain computational functions while limiting others.']}, {'end': 659.361, 'start': 538.972, 'title': 'Adaptation: evolutionary vs human life scale', 'summary': 'Discusses the contrast between the awe-inspiring but incomprehensible nature of evolutionary adaptation and the intricate study of human life scale adaptation, particularly focusing on developmental neurobiology and the dynamics of cell multiplication and death in the early stages of human life.', 'duration': 120.389, 'highlights': ['The human life time scale is however thing you can tease apart and study, particularly focusing on developmental neurobiology and the dynamics of cell multiplication and death in the early stages of human life.', "The fact that you're building a system in three dimensions in just days and months, those early days of a human life are really interesting.", 'There are times of immense cell multiplication and the greatest cell death in the brain is during infancy, demonstrating the dynamics of cellular changes in the early stages of human life.', 'The evolutionary process is awe-inspiring but incomprehensible, making it challenging to understand and work on.']}], 'duration': 278.675, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw380686.jpg', 'highlights': ["The human mind's adaptive ability encompasses generational and individual learning, and neurobiology's mathematical system.", "The comparison of neurobiology's three-dimensional structure to computer chips' limitations and its impact on computational capabilities.", 'The significance of the three-dimensional structure of the neocortex and the white matter in enabling certain computational functions while limiting others.', 'The human life time scale is particularly focusing on developmental neurobiology and the dynamics of cell multiplication and death in the early stages of human life.', 'There are times of immense cell multiplication and the greatest cell death in the brain is during infancy, demonstrating the dynamics of cellular changes in the early stages of human life.', 'The differences between adaptation over generations and individual learning present a challenge in neurobiology due to its mathematical system.', "The fact that you're building a system in three dimensions in just days and months, those early days of a human life are really interesting.", 'The evolutionary process is awe-inspiring but incomprehensible, making it challenging to understand and work on.']}, {'end': 977.153, 'segs': [{'end': 799.306, 'src': 'embed', 'start': 767.949, 'weight': 0, 'content': [{'end': 774.553, 'text': "When I taught freshman physics, I used to say, I wanted to get physics to understand the subject, to understand Newton's laws.", 'start': 767.949, 'duration': 6.604}, {'end': 784.358, 'text': "I didn't want them simply to memorize a set of examples to which they knew the equations to write down to generate the answers.", 'start': 775.593, 'duration': 8.765}, {'end': 787.881, 'text': 'I had this nebulous idea of understanding.', 'start': 784.859, 'duration': 3.022}, {'end': 799.306, 'text': 'so that if you looked at a situation, you could say oh, I expect the ball to make that trajectory, or I expect some intuitive notion of understanding,', 'start': 788.857, 'duration': 10.449}], 'summary': 'Teaching physics with emphasis on understanding, not memorization.', 'duration': 31.357, 'max_score': 767.949, 'thumbnail': ''}, {'end': 876.534, 'src': 'embed', 'start': 814.438, 'weight': 1, 'content': [{'end': 826.003, 'text': 'feed-forward neural nets, which do amazing things and yet you know contain nothing of the essence of what I would have felt was understanding.', 'start': 814.438, 'duration': 11.565}, {'end': 829.345, 'text': 'Understanding is more than just an enormous lookup table.', 'start': 826.783, 'duration': 2.562}, {'end': 832.287, 'text': "Let's linger on that.", 'start': 831.226, 'duration': 1.061}, {'end': 834.408, 'text': 'How sure you are of that?', 'start': 833.007, 'duration': 1.401}, {'end': 836.61, 'text': 'What if the table gets really big??', 'start': 834.749, 'duration': 1.861}, {'end': 843.695, 'text': 'So I mean asked another way these feed-forward neural networks.', 'start': 838.371, 'duration': 5.324}, {'end': 845.056, 'text': "do you think they'll ever understand??", 'start': 843.695, 'duration': 1.361}, {'end': 848.375, 'text': 'I could answer that in two ways.', 'start': 846.834, 'duration': 1.541}, {'end': 858.74, 'text': 'I think if you look at real systems, feedback is an essential aspect of how these real systems compute.', 'start': 848.475, 'duration': 10.265}, {'end': 867.704, 'text': 'On the other hand, if I have a mathematical system with feedback, I know I can unlayer this and do it.', 'start': 859.64, 'duration': 8.064}, {'end': 874.187, 'text': 'But I have an exponential expansion in the amount of stuff I have to build if I can solve the problem that way.', 'start': 867.984, 'duration': 6.203}, {'end': 876.534, 'text': 'So feedback is essential.', 'start': 875.373, 'duration': 1.161}], 'summary': 'Feed-forward neural nets lack understanding, feedback is essential for real systems.', 'duration': 62.096, 'max_score': 814.438, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw814438.jpg'}, {'end': 936.449, 'src': 'embed', 'start': 905.814, 'weight': 3, 'content': [{'end': 917.182, 'text': "They'll probably go back to saying well, I can look at the brain rhythms and tell you, this is a brain which is never going to function again.", 'start': 905.814, 'duration': 11.368}, {'end': 923.946, 'text': "This other one is one that if we treat it well, it's still recoverable.", 'start': 917.202, 'duration': 6.744}, {'end': 936.449, 'text': "And then just do that by some electrodes looking at simple electrical patterns which don't look in any detail at all at what individual neurons are doing.", 'start': 924.807, 'duration': 11.642}], 'summary': 'Electrodes can diagnose brain function with simple electrical patterns.', 'duration': 30.635, 'max_score': 905.814, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw905814.jpg'}], 'start': 661.572, 'title': 'Understanding the mind and neural networks', 'summary': 'Discusses understanding the mind through a physics lens, emphasizing genuine comprehension in education, and the limitations of feed-forward neural networks in achieving understanding, highlighting the essential role of feedback and the absence of brain rhythms in artificial systems.', 'chapters': [{'end': 787.881, 'start': 661.572, 'title': 'Understanding the mind through physics', 'summary': 'Discusses the approach to understanding the mind through a physics lens and the influence of a physics background on problem-solving, emphasizing the need for genuine comprehension rather than rote memorization in education.', 'duration': 126.309, 'highlights': ['The speaker grew up with a physics background, which influences their approach to problem-solving and understanding the world.', "Both of the speaker's parents were physicists, instilling a belief that the world is comprehensible through experimentation and structured thinking.", 'The speaker emphasizes the importance of genuine understanding over rote memorization in education, particularly in teaching physics to students.']}, {'end': 977.153, 'start': 788.857, 'title': 'Neural networks and understanding', 'summary': 'Discusses the limitations of feed-forward neural networks in achieving understanding, highlighting the essential role of feedback and the absence of brain rhythms in artificial systems when compared to biological systems.', 'duration': 188.296, 'highlights': ['The limitations of feed-forward neural networks in achieving understanding are discussed, emphasizing that understanding is more than just an enormous lookup table.', "The essential role of feedback in real systems' computation is highlighted, indicating that feedback is essential for achieving understanding through neural networks.", 'The absence of brain rhythms in artificial systems, such as those at Google, is pointed out, emphasizing the disparity between biological and artificial systems in terms of fundamental functioning.']}], 'duration': 315.581, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw661572.jpg', 'highlights': ['The speaker emphasizes genuine understanding over rote memorization in education.', 'The limitations of feed-forward neural networks in achieving understanding are discussed.', "The essential role of feedback in real systems' computation is highlighted.", 'The absence of brain rhythms in artificial systems is pointed out.']}, {'end': 1228.263, 'segs': [{'end': 1008.744, 'src': 'embed', 'start': 977.653, 'weight': 1, 'content': [{'end': 985.117, 'text': "Because you're one of the seminal figures in launching both disciplines, both sides of the river.", 'start': 977.653, 'duration': 7.464}, {'end': 988.499, 'text': "I think it's going to go on.", 'start': 985.137, 'duration': 3.362}, {'end': 1000.499, 'text': "generation after generation, the way it has, where what you might call the AI computer science community says, let's take the following.", 'start': 989.673, 'duration': 10.826}, {'end': 1003.081, 'text': 'This is our model of neurobiology at the moment.', 'start': 1000.519, 'duration': 2.562}, {'end': 1008.744, 'text': "Let's pretend it's good enough and do everything we can with it.", 'start': 1003.501, 'duration': 5.243}], 'summary': 'Seminal figure in launching ai and computer science disciplines, foreseeing continued growth and influence across generations.', 'duration': 31.091, 'max_score': 977.653, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw977653.jpg'}, {'end': 1228.263, 'src': 'embed', 'start': 1030.837, 'weight': 0, 'content': [{'end': 1042.541, 'text': "and I think there could be generations of this evolution and I don't know how many of them and each one is going to get you further into what a brain does and,", 'start': 1030.837, 'duration': 11.704}, {'end': 1048.304, 'text': 'in some sense, pass the Turing test longer and for more broad aspects.', 'start': 1042.541, 'duration': 5.763}, {'end': 1060.906, 'text': "And how many of these there are going to have to be before you say, I've made something, I've made a human, I don't know.", 'start': 1051.983, 'duration': 8.923}, {'end': 1063.667, 'text': 'But your sense is it might be a couple.', 'start': 1061.906, 'duration': 1.761}, {'end': 1065.908, 'text': 'My sense is it might be a couple more.', 'start': 1064.227, 'duration': 1.681}, {'end': 1072.25, 'text': 'Yeah And going back to my brainwaves, as it were.', 'start': 1066.208, 'duration': 6.042}, {'end': 1083.911, 'text': 'Yes From the AI point of view, they would say, ah, maybe these are an epiphenomenon and not important at all.', 'start': 1072.57, 'duration': 11.341}, {'end': 1096.398, 'text': 'The first car I had, a real wreck of a 1936 Dodge, go above about 45 miles an hour and the wheels would shimmy.', 'start': 1087.053, 'duration': 9.345}, {'end': 1101.021, 'text': 'Good speedometer, that.', 'start': 1099.18, 'duration': 1.841}, {'end': 1106.315, 'text': 'Now, nobody designed the car that way.', 'start': 1103.413, 'duration': 2.902}, {'end': 1108.257, 'text': 'The car is malfunctioning to have that.', 'start': 1106.395, 'duration': 1.862}, {'end': 1115.062, 'text': 'But in biology, if it were useful to know when are you going more than 45 miles an hour,', 'start': 1108.637, 'duration': 6.425}, {'end': 1118.465, 'text': "you just capture that and you wouldn't worry about where it came from.", 'start': 1115.062, 'duration': 3.403}, {'end': 1127.032, 'text': "Yeah It's going to be a long time before that kind of thing, which can take place..", 'start': 1122.288, 'duration': 4.744}, {'end': 1134.349, 'text': 'in large, complex networks of things is actually used in the computation.', 'start': 1128.766, 'duration': 5.583}, {'end': 1140.071, 'text': 'look the um, how many transistors are there in your laptop these days?', 'start': 1134.349, 'duration': 5.722}, {'end': 1142.192, 'text': "actually, i don't.", 'start': 1140.071, 'duration': 2.121}, {'end': 1143.053, 'text': "i don't know the number.", 'start': 1142.192, 'duration': 0.861}, {'end': 1145.774, 'text': "it's, it's on the scale of 10 to the 10.", 'start': 1143.053, 'duration': 2.721}, {'end': 1147.375, 'text': "i can't remember the number either.", 'start': 1145.774, 'duration': 1.601}, {'end': 1158.622, 'text': 'yeah, and all the transistors are somewhat similar and most physical systems with that many parts, all of which are similar,', 'start': 1147.375, 'duration': 11.247}, {'end': 1159.843, 'text': 'have collective properties.', 'start': 1158.622, 'duration': 1.221}, {'end': 1165.928, 'text': 'Yes Sound waves in air, earthquakes, what have you, have collective properties.', 'start': 1160.864, 'duration': 5.064}, {'end': 1174.476, 'text': 'Weather There are no collective properties used in artificial neural networks, in A.I.', 'start': 1166.409, 'duration': 8.067}, {'end': 1178.989, 'text': "Yeah, it's very..", 'start': 1177.648, 'duration': 1.341}, {'end': 1187.776, 'text': "If biology uses them, it's going to take us to more generations of things for people to actually dig in and see how they are used and what they mean.", 'start': 1178.989, 'duration': 8.787}, {'end': 1191.759, 'text': "See, you're very right.", 'start': 1189.758, 'duration': 2.001}, {'end': 1198.925, 'text': 'We might have to return several times to neurobiology and try to make our transistors more messy.', 'start': 1191.919, 'duration': 7.006}, {'end': 1200.366, 'text': 'Yeah, yeah.', 'start': 1199.605, 'duration': 0.761}, {'end': 1207.612, 'text': 'At the same time, the simple ones will conquer big aspects.', 'start': 1201.829, 'duration': 5.783}, {'end': 1222.98, 'text': 'And I think one of the biggest surprises to me was how well learning systems, which are manifestly non-biological,', 'start': 1210.253, 'duration': 12.727}, {'end': 1228.263, 'text': 'how important they can be actually and how important and how useful they can be in AI.', 'start': 1222.98, 'duration': 5.283}], 'summary': 'Discussion on the evolution of ai and the importance of collective properties in neural networks.', 'duration': 197.426, 'max_score': 1030.837, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw1030837.jpg'}], 'start': 977.653, 'title': 'Ai evolution and neural comparison', 'summary': 'Delves into the ongoing ai evolution and neurobiology, discussing iterative nature, potential for human-like capabilities through the turing test, comparison between biological and artificial intelligence, the concept of epiphenomenon, usefulness of traits, large-scale computational processes, and the absence of collective properties in artificial neural networks.', 'chapters': [{'end': 1063.667, 'start': 977.653, 'title': 'Ai evolution and neurobiology', 'summary': 'Discusses the ongoing evolution of ai and neurobiology, highlighting the iterative nature and potential for multiple generations to advance understanding of the brain and potentially achieve human-like capabilities through the turing test.', 'duration': 86.014, 'highlights': ['The evolution of AI and neurobiology is iterative, with potential for multiple generations to advance understanding of the brain and potentially achieve human-like capabilities through the Turing test.', 'The AI computer science community continually updates its model of neurobiology and iterates on its approach, leading to incremental advancements and potential breakthroughs in emulating human capabilities.']}, {'end': 1159.843, 'start': 1064.227, 'title': 'Understanding biological and artificial intelligence', 'summary': 'Discusses the comparison between biological and artificial intelligence, highlighting the concept of epiphenomenon, the usefulness of certain traits, and the complexity of large-scale computational processes, while also mentioning the scale and similarity of transistors in modern laptops.', 'duration': 95.616, 'highlights': ['In biology, certain traits like knowing when a car is going more than 45 miles per hour are captured without concern for their origin, similar to the concept of epiphenomenon in AI.', 'Large, complex networks of things taking place in computation will take a long time to be utilized.', 'Modern laptops contain transistors on the scale of 10 to the power of 10, with collective properties due to the similarity of the transistors.', "The concept of epiphenomenon in AI suggests that certain characteristics may be considered unimportant, similar to the case of the car's malfunction at 45 miles per hour.", "The discussion of the 1936 Dodge car's malfunction at 45 miles an hour serves as an example of an epiphenomenon."]}, {'end': 1228.263, 'start': 1160.864, 'title': 'Collective properties in artificial neural networks', 'summary': 'Discusses the absence of collective properties in artificial neural networks compared to natural phenomena, highlighting the potential impact of learning systems in ai.', 'duration': 67.399, 'highlights': ['Learning systems have proven to be crucial and useful in AI, despite being non-biological.', 'The absence of collective properties in artificial neural networks compared to natural phenomena such as sound waves and earthquakes.', 'The potential need to explore neurobiology for enhancing the performance of transistors in AI.']}], 'duration': 250.61, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/8bt6r6CIghw/pics/8bt6r6CIghw977653.jpg', 'highlights': ['The evolution of AI and neurobiology is iterative, with potential for multiple generations to advance understanding of the brain and potentially achieve human-like capabilities through the Turing test.', 'The AI computer science community continually updates its model of neurobiology and iterates on its approach, leading to incremental advancements and potential breakthroughs in emulating human capabilities.', 'Learning systems have proven to be crucial and useful in AI, despite being non-biological.', "The concept of epiphenomenon in AI suggests that certain characteristics may be considered unimportant, similar to the case of the car's malfunction at 45 miles per hour.", "The discussion of the 1936 Dodge car's malfunction at 45 miles an hour serves as an example of an epiphenomenon.", 'The potential need to explore neurobiology for enhancing the performance of transistors in AI.', 'In biology, certain traits like knowing when a car is going more than 45 miles per hour are captured without concern for their origin, similar to the concept of epiphenomenon in AI.', 'Large, complex networks of things taking place in computation will take a long time to be utilized.', 'Modern laptops contain transistors on the scale of 10 to the power of 10, with collective properties due to the similarity of the transistors.', 'The absence of collective properties in artificial neural networks compared to natural phenomena such as sound waves and earthquakes.']}], 'highlights': ['Evolutionary biology has shaped neurons in biological neural networks to capture all kinds of possibilities, turning quirks into useful features.', 'Biological systems leverage complexity and adaptability through DNA duplication and evolution, allowing for the retention of old functions and the development of new ones.', "The human mind's adaptive ability encompasses generational and individual learning, and neurobiology's mathematical system.", 'The speaker emphasizes genuine understanding over rote memorization in education.', 'The evolution of AI and neurobiology is iterative, with potential for multiple generations to advance understanding of the brain and potentially achieve human-like capabilities through the Turing test.']}