Yoshua Bengio: Deep Learning

Yoshua Bengio, along with Geoffrey Hinton and Yann Lecun, is considered one of the three people most responsible for the advancement of deep learning during the 1990s, 2000s, and now. Cited 139,000 times, he has been integral to some of the biggest breakthroughs in AI over the past 3 decades. Video version is available on YouTube https://youtu.be/azOmzumh0vQ. If you would like to get more information about this podcast go to https://lexfridman.com/ai or connect with @lexfridman on Twitter https://twitter.com/lexfridman, LinkedIn https://www.linkedin.com/in/lexfridman/, Facebook https://www.facebook.com/lexfridman, or YouTube https://www.youtube.com/lexfridman where you can watch the video versions of these conversations.

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Lex Fridman?

  Welcome to the Artificial Intelligence Podcast. My name is Lex Fridman. I'm a research scientist at MIT. If you enjoy this podcast, please rate it on iTunes or your podcast provider of choice, or simply connect with me on Twitter and other social networks at LexFriedman, spelled F R I D. Today is a conversation with Yoshua Bengio. Along with Jeff Hinton and Jan Le Cun, he's considered one of the three people most responsible for the advancement of deep learning during the 1990s and the 2000s and now. Cited 139,000 times, he has been integral to some of the biggest breakthroughs in AI over the past three decades.

• 1:00Lex Fridman?

  What difference between biological neural networks and artificial neural networks is most mysterious, captivating, and profound for you?

• 1:09Speaker 01

  First of all, there's so much we don't know about biological neural networks. And that's very mysterious and captivating because maybe it holds the key to improving artificial neural networks. One of the things I studied recently is something that we don't know how biological neural networks do, but would be really useful for artificial ones, is the ability to do credit assignment through very long time spans. There are things that we can, in principle, do with artificial neural nets, but it's not very convenient and it's not biologically plausible.

• 1:51Speaker 01

  And this mismatch, I think, this kind of mismatch, may be an interesting thing to study to, A, understand better how brains might do these things, because we don't have good corresponding theories with artificial neural nets, and B,

• 2:08Speaker 01

  Maybe provide new ideas that we could explore about things that brain do differently and that we could incorporate in artificial neural nets.

• 2:20Lex Fridman?

  So let's break credit assignment up a little bit.

• 2:23Speaker 01

  Yes.

• 2:23Lex Fridman?

  So what, it's a beautifully technical term, but it could incorporate so many things. So is it more on the RNN memory side? That thinking like that? Or is it something about knowledge, building up common sense knowledge over time? Or is it more in the reinforcement learning sense that you're picking up rewards over time for a particular, to achieve a certain kind of goal?

• 2:48Speaker 01

  So I was thinking more about the first two meanings whereby we store all kinds of memories, episodic memories in our brain, which we can access later in order to help us both infer causes of things that we are observing now and assign credit to decisions or interpretations we came up with a while ago when, you know, those memories were stored. And then we can change the way we would have reacted or interpreted things in the past. And now that's credit assignment used for learning.

• 3:34Lex Fridman?

  So in which way do you think artificial neural networks, the current LSTM, the current architectures are not able to capture the presumably you're thinking of very long term?

• 3:51Speaker 01

  Yes. So current, the current nets are doing a fairly good jobs for sequences with dozens or say hundreds of time steps. And then it gets sort of harder and harder and depending on what you have to remember and so on, as you consider longer durations. Whereas humans seem to be able to do credit assignment through essentially arbitrary times. Like I could remember something I did last year. And then now because I see some new evidence, I'm going to change my mind about the way I was thinking last year and hopefully not do the same mistake again.

• 4:31Lex Fridman?

  I think a big part of that is probably forgetting. You're only remembering the really important things. So it's very efficient forgetting.

• 4:40Speaker 01

  Yes, so there's a selection of what we remember. And I think there are really cool connections to higher level cognition here regarding consciousness, deciding and emotions, like sort of deciding what comes to consciousness and what gets stored in memory, which are not trivial either.

• 5:01Lex Fridman?

  So you've been at the forefront there all along, showing some of the amazing things that neural networks, deep neural networks can do in the field of artificial intelligence is just broadly in all kinds of applications. But we can talk about that forever. But what in your view, because we're thinking towards the future, is the weakest aspect of the way deep neural networks represent the world? What is the, what is, in your view, is missing?

• 5:30Speaker 01

  So current, current state of the art neural nets, trained on large quantities of images or texts,

• 5:42Speaker 01

  have some level of understanding of, you know, what explains those data sets. But it's very basic. It's, it's very low level. And it's not nearly as robust and abstract and general as our understanding.

• 6:01Speaker 01

  Okay, so that doesn't tell us how to fix things. But I think it encourages us to think about how we can maybe train our neural nets differently so that they would focus, for example, on causal explanations, something that we don't do currently with neural net training. Also,

• 6:28Speaker 01

  One thing I'll talk about in my talk this afternoon is instead of learning separately from images and videos on one hand and from text on the other hand, we need to do a better job of jointly learning about language and about the world to which it refers so that, you know, both sides can help each other. We need to have good world models in our neural nets for them to really understand sentences which talk about what's going on in the world. And I think we need language input to help provide clues about what high-level concepts,

• 7:13Speaker 01

  like semantic concepts, should be represented at the top levels of these neural nets. In fact, there is evidence that the purely unsupervised learning of representations doesn't give rise to high-level representations that are as powerful as the ones we're getting from supervised learning. And so the clues we're getting just with the labels, not even sentences, is already very powerful.

• 7:43Lex Fridman?

  Do you think that's an architecture challenge or is it a data set challenge?

• 7:47Speaker 01

  Neither.

• 7:52Lex Fridman?

  I'm tempted to just end it there.

• 7:55Speaker 01

  No.

• 7:56Lex Fridman?

  Can you elaborate slightly?

• 7:58Speaker 01

  Yes.

• 8:01Speaker 01

  Of course, data sets and architectures are something you want to always play with, but I think the crucial thing is more the training objectives, the training frameworks. For example, going from passive observation of data to more active agents which learn by intervening in the world, the relationships between causes and effects, the sort of objective functions which could be important to allow the highest level explanations to rise from the learning. Which I don't think we have now.

• 8:41Speaker 01

  The kinds of objective functions which could be used to reward exploration, the right kind of exploration. So these kinds of questions are neither in the data set nor in the architecture, but more in how we learn, under what objectives and so on.

• 8:59Lex Fridman?

  Yeah, that's a, I've heard you mention in several contexts, the idea of sort of the way children learn, they interact with objects in the world. And it seems fascinating because in some sense, except with some cases in reinforcement learning, that idea is not part of the learning process in artificial neural networks. It's almost like, do you envision something like an objective function saying, you know what, if you poke this object in this kind of way, it would be really helpful for me to further, further learn.

• 9:37Speaker 01

  Right, right.

• 9:38Lex Fridman?

  Sort of almost guiding some aspect of learning.

• 9:41Speaker 01

  Right, right, right. So I was talking to Rebecca Sax just an hour ago, and she was talking about lots and lots of evidence from infants seem to clearly pick what interests them in a directed way. And so they're not passive learners. They, they focus their attention on aspects of the world which are most interesting, surprising in a non-trivial way, that makes them change their theories of the world.

• 10:18Lex Fridman?

  So that's a fascinating view of the future progress, but on a, on a more maybe boring question, do you think going deeper and larger, so do you think just increasing the size of the things that have been increasing a lot in the past few years will, will also make significant progress? So some of the representational issues that you, you mentioned, they're kind of shallow in some sense.

• 10:50Speaker 01

  Oh, shallow, you mean?

• 10:51Lex Fridman?

  On a prior.

• 10:52Speaker 01

  In a sense of abstraction.

• 10:53Lex Fridman?

  In a sense of abstraction. They're not getting some.

• 10:55Speaker 01

  I don't think that having more, more depth in the network in the sense of, instead of a hundred layers, we have 10,000 is going to solve our problem.

• 11:04Lex Fridman?

  You don't think so?

• 11:07Lex Fridman?

  Is that obvious to you?

• 11:08Speaker 01

  Yes. What is clear to me is that engineers and companies and labs and grad students will continue to tune architectures and explore all kinds of tweaks to make the current state of the art slightly, ever slightly better. But I don't think that's gonna be nearly enough. I think we need some fairly drastic changes in the way that we are considering learning to achieve the goal that these learners actually understand in a deep way, the environment in which they are, you know, observing and acting.

• 11:44Lex Fridman?

  But I guess I was trying to ask a question that's more interesting than just more layers. It is basically, once you figure out a way to learn through interacting, how many parameters does it take to store that information? So I think our brain is quite bigger than most neural networks.

• 12:06Speaker 01

  Right, right. Oh, I see what you mean. Oh, I'm with you there. So I agree that in order to build neural nets with the kind of broad knowledge of the world that typical adult humans have, probably the kind of computing power we have now is gonna be insufficient. So, well, the good news is there are hardware companies building neural net chips and so it's gonna get better.

• 12:31Speaker 01

  However, the good news in a way, which is also a bad news, is that even our state-of-the-art deep learning methods fail to learn models that understand even very simple environments like some grid worlds that we have built. Even these fairly simple environments, I mean, of course, if you train them with enough examples, eventually they get it. But it's just like, instead of what, instead of what humans might need just dozens of examples, these things will need millions, right? For very, very, very simple tasks. And so I think there's an opportunity for academics who don't have the kind of computing

• 13:15Speaker 01

  power that, say, Google has, to do really important and exciting research to advance the state-of-the-art in training frameworks, learning models, agent learning in even simple environments that are synthetic, that seem trivial, but yet current machine learning fails on.

• 13:36Lex Fridman?

  We talked about priors and common sense knowledge. It seems like we humans take a lot of knowledge for granted. So what's your view of these priors of forming this broad view of the world, this accumulation of information and how we can teach neural networks or learning systems to pick that knowledge up? So knowledge, you know, for a while, the artificial intelligence was maybe in the 80s. Like, there's a time where knowledge representation, knowledge acquisition, expert systems, I mean,

• 14:15Lex Fridman?

  symbolic AI was a view, was an interesting problem set to solve and it was kind of put on hold a little bit, it seems like.

• 14:25Speaker 01

  Because it doesn't work.

• 14:26Lex Fridman?

  It doesn't work, that's right. But, that's right. But the goals of that.

• 14:33Speaker 01

  Remain important.

• 14:33Lex Fridman?

  Yes, remain important. And how do you think those goals can be addressed?

• 14:38Speaker 01

  Right. So first of all, I believe that one reason why the classical expert systems approach failed is because a lot of the knowledge we have, so you talked about common sense and intuition, there's a lot of knowledge like this which is not consciously accessible. There are lots of decisions we're taking that we can't really explain, even if sometimes we make up a story. And that knowledge is also necessary for machines to take good decisions. And that knowledge is hard to codify in expert systems, rule-based systems, and you know, classical AI formalism.

• 15:23Speaker 01

  And there are other issues, of course, with the old AI, like not really good ways of handling uncertainty. I would say something more subtle, which we understand better now, but I think still isn't enough in the minds of people. There's something really powerful that comes from distributed representations, the thing that really makes neural nets work so well.

• 15:50Speaker 01

  And it's hard to replicate that kind of power in a symbolic world. The knowledge in expert systems and so on is nicely decomposed into like a bunch of rules. Whereas if you think about a neural net, it's the opposite. You have this big blob of parameters which work intensely together to represent everything the network knows. And it's not sufficiently factorized. And so I think this is one of the weaknesses of current neural nets. That we have to take lessons from classical AI. In order to bring in another kind of compositionality, which is common in language, for example, and in these rules.

• 16:34Speaker 01

  But that isn't so native to neural nets.

• 16:38Lex Fridman?

  And on that line of thinking, disentangled representations.

• 16:43Speaker 01

  Yes.

• 16:44Lex Fridman?

  So.

• 16:46Speaker 01

  So let me connect with disentangled representations. If you might, if not, I might.

• 16:50Lex Fridman?

  Yes, that's exactly it.

• 16:51Speaker 01

  Yep. So for many years, I've thought, and I still believe that it's really important that we come up with learning algorithms, either unsupervised or supervised, but reinforcement, whatever. That build representations in which the important factors, hopefully causal factors, are nicely separated and easy to pick up from the representation. So that's the idea of disentangled representations. It says transform the data into a space where everything becomes easy. We can maybe just learn with linear models about the things we care about. And I still think this is important, but I think this is missing out on a very important ingredient, which classical AI systems can remind us of.

• 17:36Speaker 01

  So let's say we have these disentangled representations. You still need to learn about the relationships between the variables, those high level semantic variables. They're not going to be independent. I mean, this is like too much of an assumption. They're going to have some interesting relationships that allow to predict things in the future, to explain what happened in the past. The kind of knowledge about those relationships in a classical AI system is encoded in the rules. Like a rule is just like a little piece of knowledge that says, oh, I have these two, three, four variables that are linked in this interesting way, then I can say something about one or two of them, given a couple of others. Right? In addition to disentangling the.

• 18:17Speaker 01

  The elements of the representation, which are like the variables in a rule-based system, you also need to disentangle the mechanisms that relate those variables to each other. So like the rules. So the rules are neatly separated, like each rule is, you know, living on its own. And when I change a rule because I'm learning, it doesn't need to break other rules. Whereas current neural nets, for example, are very sensitive to what's called catastrophic forgetting, where after I've learned some things and then I learn new things, they can destroy the old things that I had learned, right? If the

• 18:56Speaker 01

  knowledge was better factorized and separated, disentangled, then you would avoid a lot of that. Now, you can't do this in the sensory domain, but

• 19:10Lex Fridman?

  What do you mean by sensory domain?

• 19:11Speaker 01

  Like in pixel space. But my idea is that when you project the data in the right semantic space, it becomes possible to now represent this extra knowledge beyond the transformation from input to representations, which is how representations act on each other and predict the future and so on, in a way that can be neatly disentangled. So now it's the rules that are disentangled from each other and not just the variables that are disentangled from each other.

• 19:39Lex Fridman?

  And you draw a distinction between semantic space and pixel, like, does there need to be an architectural difference?

• 19:45Speaker 01

  Well, yeah. So there's the sensory space, like pixels, which where everything is entangled. The information, like the variables, are completely interdependent in very complicated ways. And also computation, like the, it's not just the variables, it's also how they are related to each other is all intertwined. But I'm hypothesizing that in the right, high-level representation space, both the variables and how they relate to each other can be disentangled and that will provide a lot of generalization power.

• 20:19Lex Fridman?

  Generalization power?

• 20:21Speaker 01

  Yes.

• 20:22Lex Fridman?

  Distribution of the test set is assumed to be the same as the distribution of the training set?

• 20:28Speaker 01

  Right. This is where current machine learning is too weak. It doesn't tell us anything, is not able to tell us anything about how our neural nets, say, are going to generalize to a new distribution. And and you know, people may think, well, but there's nothing we can say if we don't know what the new distribution will be. The truth is, humans are able to generalize to new distributions.

• 20:51Lex Fridman?

  Yeah, how are we able to do that?

• 20:53Speaker 01

  Yeah, because there is something, these new distributions, even though they could look very different from the training distributions, they have things in common. So let me give you a concrete example. You read a science fiction novel. The science fiction novel maybe, you know, brings you in some other planet where things look very different on the surface, but it's still the same laws of physics. Right? And so you can read the book and you understand what's going on. So the distribution is very different, but because you can transport a lot of the knowledge you had from Earth about the underlying cause and effect relationships and physical mechanisms and all that, and maybe even social interactions, you can now make sense of what is going on

• 21:38Speaker 01

  on this planet where like visually, for example, things are totally different.

• 21:42Lex Fridman?

  Taking that analogy further and distorting it, let's enter a science fiction world of, say, Space Odyssey 2001 with Hal.

• 21:52Speaker 01

  Yeah.

• 21:53Lex Fridman?

  Or or maybe, which is probably one of my favorite AI movies.

• 21:59Speaker 01

  Me too.

• 22:00Lex Fridman?

  And then there's another one that a lot of people love that may be a little bit outside of the AI community is Ex Machina.

• 22:08Speaker 01

  Right.

• 22:09Lex Fridman?

  I don't know if you've seen it.

• 22:10Speaker 01

  Yes, yes.

• 22:11Lex Fridman?

  But what are your views on that movie? Does it, are you able to enjoy it?

• 22:15Speaker 01

  So there are things I liken, things I hate.

• 22:18Lex Fridman?

  So let me, you could talk about that in the context of a question I want to ask, which is, there's quite a large community of people from different backgrounds, often outside of AI, who are concerned about existential threat of artificial intelligence.

• 22:33Speaker 01

  Right.

• 22:34Lex Fridman?

  You've seen this community develop over time. You've seen you have a perspective. So what do you think is the best way to talk about AI safety, to think about it, to have discourse about it within AI community and outside and grounded in the fact that Ex Machina is one of the main sources of information for the general public about AI?

• 22:54Speaker 01

  So I think, I think you're putting it right. There's a big difference between. The sort of discussion we ought to have within the AI community and the sort of discussion that really matter in the general public. So I think the the picture of Terminator and, you know, AI loose and killing people and superintelligence that's going to destroy us whatever we try isn't really so useful for the public discussion because for the public discussion, the things I believe really matter are the short term and medium term, very likely negative impacts of AI on society,

• 23:35Speaker 01

  whether it's from security, like, you know, big brother scenarios with face recognition or killer robots, or the impact on the job market, or concentration of power and discrimination, all kinds of social issues which could actually, some of them could really threaten democracy, for example.

• 23:57Lex Fridman?

  Just to clarify, when you said killer robots, you mean autonomous weapon?

• 24:01Speaker 01

  Yes.

• 24:01Lex Fridman?

  Like the weapon systems?

• 24:02Speaker 01

  Yes.

• 24:02Lex Fridman?

  Not Terminator?

• 24:04Speaker 01

  That's right. So so I think these these short and medium term concerns should be important parts of the public debate. Now, existential risk for me is a very unlikely consideration, but still worth academic investigation. In the same way that you could say, should we study what could happen if meteorite, you know, came to Earth and destroyed it? So I think it's very unlikely that this is going to happen in or happen in a reasonable future. It's it's very the the sort of scenario of an AI getting loose goes against my understanding of at least current machine learning and current neural nets and so on. It's not plausible to me.

• 24:48Speaker 01

  But of course, I don't have a crystal ball and who knows what AI will be in 50 years from now. So I think it is worth that scientists study those problems. It's just not a pressing question as far as I'm concerned.

• 24:59Lex Fridman?

  So before I continue down that line, I have a few questions there, but what do you like and not like about Ex Machina as a movie? Because I actually watched it for the second time and enjoyed it. I hated it the first time and I enjoyed it quite a bit more the second time when I sort of learned to accept certain pieces of it, see it as a concept movie. What was your experience? What were your thoughts?

• 25:27Speaker 01

  So the negative is the picture it paints of science is totally wrong. Science in general and AI in particular. Science is not happening in some hidden place by some, you know, really smart guy.

• 25:46Lex Fridman?

  One person.

• 25:47Speaker 01

  One person. This is totally unrealistic. This is not how it happens. Even a team of people in some isolated place will not make it. Science moves by small steps, thanks to the collaboration and community of a large number of people interacting. And all the scientists who are expert in their field kind of know what is going on, even in the industrial labs. It's information flows and leaks and so on. And the spirit of it is very different from the way science is painted in this movie.

• 26:27Lex Fridman?

  Yeah, let me ask on that on that point. It's been the case to this point that kind of even if the research happens inside Google or Facebook inside companies, it still kind of comes out, ideas come out.

• 26:38Speaker 01

  Absolutely.

• 26:39Lex Fridman?

  Do you think that will always be the case with AI? Is it possible to bottle ideas to the point where there's a set of breakthroughs that go completely undiscovered by the general research community? Do you think that's even possible?

• 26:53Speaker 01

  It's possible, but it's unlikely.

• 26:55Lex Fridman?

  Unlikely.

• 26:56Speaker 01

  It's not how it is done now. It's not how I can foresee it in the foreseeable future. But of course, I don't have a crystal ball. And so who knows? This is science fiction after all. But usually.

• 27:14Lex Fridman?

  I think it's ominous that the lights went off during that discussion.

• 27:19Speaker 01

  So the problem, again, there's a, you know, one thing is the movie. And you could imagine all kinds of science fiction. The problem with, for me, maybe similar to the question about existential risk, is that this kind of movie paints such a wrong picture of what is actual, you know, the actual science and how it's going on, that it can have unfortunate effects on people's understanding of current science. And so that's kind of sad. There's an important principle in research, which is diversity. So in other words, research is exploration. Research is exploration in the space of ideas. And different people will focus on different directions.

• 28:03Speaker 01

  And this is not just good, it's essential. So I'm totally fine with people exploring directions that are contrary to mine or look orthogonal to mine. I am more than fine. I think it's important. I and my friends don't claim we have universal truth about what will, especially about what will happen in the future. Now, that being said, we have our intuitions and then we act accordingly, according to where we think we can be most useful and where society has the most to gain or to lose.

• 28:40Speaker 01

  We should have those debates and and not end up in a society where there's only one voice and one way of thinking. And research money is spread out, so.

• 28:53Lex Fridman?

  Disagreement is a sign of good research, good science, so.

• 28:57Speaker 01

  Yes.

• 28:59Lex Fridman?

  The idea of bias in the human sense of bias.

• 29:02Speaker 01

  Yep.

• 29:03Lex Fridman?

  How do you think about instilling in machine learning, something that's aligned with human values in terms of bias? We intuitively as human beings have a concept of what bias means, of what a fundamental respect for other human beings means. But how do we instill that into machine learning systems, do you think?

• 29:24Speaker 01

  So I think there are short term things that are already happening. And then there are long term things that we need to do. In the short term, there are techniques that have been proposed and I think will continue to be improved and maybe alternatives will come up to take data sets in which we know there is bias, we can measure it. Pretty much any data set where humans are being observed, taking decisions, will have some sort of bias, discrimination against particular groups and so on. And we can use machine learning techniques to try to build predictors, classifiers that are going to be less biased. We can do it, for example, using adversarial methods

• 30:09Speaker 01

  to make our systems less sensitive to these variables we should not be sensitive to. So these are clear, well-defined ways of trying to address the problem. Maybe they have weaknesses and more research is needed and so on. But I think, in fact, they're sufficiently mature that governments should start regulating companies where it matters, say like insurance companies, so that they use those techniques. Because those techniques will provably reduce the bias, but at a cost. For example, maybe their predictions will be less accurate. And so companies will not do it until you force them. All right, so this is short term.

• 30:49Speaker 01

  Long term, I'm really interested in thinking of how we can instill moral values into computers. Obviously, this is not something we'll achieve in the next five or 10 years. How can we, you know, there's already work in detecting emotions, for example, in images, in sounds, in texts, and also studying how different agents interacting in different ways may correspond to patterns of, say, injustice, which could trigger anger.

• 31:26Speaker 01

  So these are things we can do in the medium term and eventually train computers to model, for example, how humans react emotionally. I would say the simplest thing is unfair situations, which trigger anger. This is one of the most basic emotions that we share with other animals. I think it's quite feasible within the next few years, so we can build systems that can detect these kinds of things, to the extent, unfortunately, that they understand enough about the world around us, which is a long time away. But maybe we can initially do this in virtual environments.

• 32:08Speaker 01

  So you can imagine like a video game where agents interact in some ways and then some situations trigger an emotion. I think we could train machines to detect those situations and predict that the particular emotion, you know, will likely be felt if a human was playing one of the characters.

• 32:27Lex Fridman?

  You have shown excitement and done a lot of excellent work with unsupervised learning. But on the, you know, there's been a lot of success on the supervised learning.

• 32:37Speaker 01

  Yes, yes.

• 32:38Lex Fridman?

  And one of the things I'm really passionate about is how humans and robots work together. And in the context of supervised learning, that means the process of annotation. Do you think about the problem of annotation of put in a more interesting way is humans teaching machines? Is there?

• 33:00Speaker 01

  Yes, I think it's an important subject. Reducing it to annotation may be useful for somebody building a system tomorrow, but.

• 33:10Speaker 01

  Longer term, the process of teaching, I think is something that deserves a lot more attention from the machine learning community. So there are people who've coined the term "machine teaching." So what are good strategies for teaching a learning agent? And can we design, train a system that is going to be a good teacher? So in my group, we have a project called BBI or BBI game, where there is a game or scenario where there's a learning agent and a teaching agent. Presumably, the teaching agent would eventually be a human, but we're not there yet.

• 33:51Speaker 01

  And the role of the teacher is to use its knowledge of the environment, which it can acquire using whatever way, brute force, to help the learner learn as quickly as possible. So the learner is going to try to learn by itself, maybe using some exploration and whatever. But the teacher can choose, can have an influence on the interaction with the learner so as to guide the learner, maybe teach it the things that the learner has most trouble with, or just at the boundary between what it knows and doesn't know, and so on.

• 34:31Speaker 01

  So there's a tradition of these kind of ideas from other fields, like tutorial systems, for example, and AI. And of course, people in the humanities have been thinking about these questions. But I think it's time that machine learning people look at this, because in the future, we'll have more and more human-machine interaction with a human in the loop. And I think understanding how to make this work better-

• 34:58Lex Fridman?

  All the problems around that are very interesting and not sufficiently addressed. You've done a lot of work with language, too. What aspect of the traditionally formulated Turing test, a test of natural language understanding and generation, in your eyes, is the most difficult of conversation? What, in your eyes, is the hardest part of conversation to solve for machines?

• 35:22Speaker 01

  So I would say it's everything having to do with the non-linguistic knowledge, which implicitly you need in order to make sense of sentences. Things like the Wienergrad schema, so these sentences that are semantically ambiguous. In other words, you need to understand enough about the world in order to really interpret properly those sentences. I think these are interesting challenges for machine learning because they point in the direction of building systems that both understand how the world works and there's causal relationships in the world and associate that knowledge with how to express it in language,

• 36:05Speaker 01

  either for reading or writing.

• 36:10Lex Fridman?

  You speak French?

• 36:11Speaker 01

  Yes, it's my mother tongue.

• 36:12Lex Fridman?

  It's one of the Romance languages. Do you think passing the Turing test and all the underlying challenges we just mentioned depend on language? Do you think it might be easier in French than it is in English?

• 36:24Speaker 01

  No.

• 36:25Lex Fridman?

  Or is it independent of language?

• 36:27Speaker 01

  I think it's independent of language. I would like to build systems that can use the same principles, the same learning mechanisms to learn from human agents, whatever their language.

• 36:45Lex Fridman?

  Well, certainly us humans can talk more beautifully and smoothly in poetry. So I'm Russian originally. I know poetry in Russian is maybe easier to convey complex ideas than it is in English. But maybe I'm showing my bias and some people could say that about French. But of course, the goal ultimately is our human brain is able to utilize any kind of those languages to use them as tools to convey meaning.

• 37:16Speaker 01

  Yeah, of course, there are differences between languages and maybe some are slightly better at some things. But in the grand scheme of things, where we're trying to understand how the brain works and language and so on, I think these differences are minute.

• 37:31Lex Fridman?

  So you've lived perhaps through an AI winter of sorts?

• 37:36Speaker 01

  Yes.

• 37:38Lex Fridman?

  How did you stay warm and continue with your research?

• 37:42Speaker 01

  Stay warm with friends.

• 37:43Lex Fridman?

  With friends. Okay, so it's important to have friends. And what have you learned from the experience?

• 37:51Speaker 01

  Listen to your inner voice. Don't, you know, be trying to just please the crowds and the fashion. And if you have a strong intuition about something that is not contradicted by actual evidence, go for it. I mean, it could be contradicted by people.

• 38:15Lex Fridman?

  But not your own instinct of based on everything you've learned.

• 38:18Speaker 01

  So of course, of course, you have to adapt your beliefs when your experiments contradict those beliefs. But but you have to stick to your beliefs. Otherwise, it's it's it's what allowed me to go through those years. It's what allowed me to persist in directions that, you know, took time, whatever other people think, took time to mature and bring fruits.

• 38:46Lex Fridman?

  So history of AI is marked with these. Of course, it's marked with technical breakthroughs, but it's also marked with these seminal events that capture the imagination of the community. Most recent, I would say, AlphaGo beating the world champion human Go player was one of those moments. What do you think the next such moment might be?

• 39:11Speaker 01

  Okay, sir, first of all, I think that these so-called seminal events are overrated.

• 39:20Speaker 01

  As I said, science really moves by small steps. Now, what happens is you make one more small step and it's like the the drop that, you know, allows to, that fills the bucket. And and and then you have drastic consequences because now you're able to do something you were not able to. Before or now, say, the cost of building some device or solving a problem becomes cheaper than what existed and you have a new market that opens up. Right. So especially in the world of commerce and applications, the impact of a small scientific progress could be huge.

• 40:01Speaker 01

  But in the science itself, I think it's very, very gradual and.

• 40:06Lex Fridman?

  Where are these steps being taken now? So there is unsupervised.

• 40:11Speaker 01

  So if I look at, if I look at one trend that I like in, in my community. So for example, and at Milan, my institute, what are the two hottest topics? GANs and reinforcement learning, even though in Montreal in particular, like reinforcement learning was something pretty much absent just two or three years ago. So it is really a big interest from students and there's a big interest from people like me.

• 40:46Speaker 01

  So I would say this is something where we're going to see more progress, even though it hasn't yet provided much in terms of actual industrial fallout. Like even though there's AlphaGo, there's no, like Google is not making money on this right now. But I think over the long term, this is really, really important for many reasons.

• 41:07Speaker 01

  So in other words, agent, I would say reinforcement learning, maybe more generally agent learning, because it doesn't have to be with rewards. It could be in all kinds of ways that an agent is learning about its environment.

• 41:18Lex Fridman?

  Now reinforcement learning you're excited about, do you think, do you think GANs could provide something?

• 41:26Speaker 01

  Yes.

• 41:26Lex Fridman?

  Some moment in.

• 41:29Speaker 01

  Well, GANs or other.

• 41:32Speaker 01

  Generalized models, I believe, will be crucial ingredients in building agents that can understand the world. A lot of the successes in reinforcement learning in the past has been with policy gradient, where you'll just learn a policy, you don't actually learn a model of the world. But there are lots of issues with that, and we don't know how to do model-based RL right now. But I think this is where we have to go in order to build models that can generalize faster and better, like to new distributions, that capture, to some extent at least, the underlying causal mechanisms in the world.

• 42:14Lex Fridman?

  Last question. What made you fall in love with artificial intelligence? If you look back, what was the first moment in your life when you were fascinated by either the human mind or the artificial mind?

• 42:29Speaker 01

  You know, when I was an adolescent, I was reading a lot, and then I started reading science fiction.

• 42:36Lex Fridman?

  There you go.

• 42:36Speaker 01

  There I go. Yeah, that's it. That's where I got hooked. And then, you know, I had one of the first personal computers, and I got hooked in programming. And so it just, you know.

• 42:50Lex Fridman?

  Start with fiction and then make it a reality.

• 42:52Speaker 01

  That's right.

• 42:54Lex Fridman?

  Yosh, thank you so much for talking to me.

• 42:55Speaker 01

  My pleasure.