The Question of Reasoning with Neural Networks

Can neural networks be made to reason? This is the question that has sparked debate among researchers and experts in the field. The speaker believes that reasoning is not just about learning, but also about how much prior structure is needed in a neural network for human-like reasoning to emerge.

A question has been raised as to whether our current models of what reasoning is based on logic are discrete and incompatible with gradient-based learning. This raises an interesting point about the nature of learning and the type of mathematics used in deep learning. The speaker notes that the math used in deep learning is more akin to cybernetics and electrical engineering than computer science, which emphasizes precision and attention to detail.

This is in contrast to machine learning, which is often seen as the science of sloppiness. While this may seem counterintuitive, it highlights the need for new ideas and approaches to tackle the challenge of creating neural networks that can reason. The speaker notes that there are already people working on this problem, whose main goal is to develop new algorithms and techniques that can enable machines to reason like humans.

The speaker also touches upon another form of reasoning, which involves energy minimization. This type of reasoning is often used in AI systems that require planning and optimization, such as those used in robotics or control systems. The idea is to find the optimal sequence of actions that achieves a particular goal, while minimizing the amount of energy expended.

This type of reasoning has its roots in classical physics and has been applied to various fields, including computer science. The speaker notes that this approach may have contributed to the ability of humans to reason and solve problems. However, the challenge lies in how to represent knowledge and manipulate it using logic and continuous functions.

One potential solution is to replace symbols with vectors and replace logic with continuous functions. This approach has been advocated for by geoff hinton, who suggests that learning systems should be able to manipulate objects that fit into a space and then put the result back into the same space. The speaker notes that this idea may lead to new forms of reasoning that are similar to simple expert systems.

The debate surrounding how much prior structure is needed for machines to reason is an ongoing one, with some researchers arguing that more structure is required, while others believe that less structure can be sufficient. This debate highlights the need for further research and development in the field of artificial intelligence, as we strive to create machines that can think and reason like humans.

The Role of Energy Minimization in Reasoning

Energy minimization is another approach to reasoning that involves finding the optimal sequence of actions that achieves a particular goal, while minimizing the amount of energy expended. This type of reasoning has its roots in classical physics and has been applied to various fields, including computer science.

In this context, reasoning can be seen as a process of optimization, where the goal is to find the best solution among a set of possible solutions. The speaker notes that this approach may have contributed to the ability of humans to reason and solve problems, particularly in domains such as planning and control.

One example of energy minimization in AI systems is market model predictive control, which involves using an energy function to guide decision-making. This type of reasoning has been used in various applications, including robotics and autonomous vehicles.

The speaker notes that while this approach may be useful for certain types of problems, it raises questions about how to represent knowledge and manipulate it using logic and continuous functions. The challenge lies in finding a way to balance the need for precision and optimization with the need for flexibility and adaptability.

New Ideas for Reasoning with Neural Networks

One potential solution to the problem of reasoning with neural networks is to replace symbols with vectors and replace logic with continuous functions. This approach has been advocated for by geoff hinton, who suggests that learning systems should be able to manipulate objects that fit into a space and then put the result back into the same space.

The speaker notes that this idea may lead to new forms of reasoning that are similar to simple expert systems. However, the debate surrounding how much prior structure is needed for machines to reason is ongoing, with some researchers arguing that more structure is required, while others believe that less structure can be sufficient.

Gary Marcus and other researchers have argued that more structure is needed for machines to reason, while geoff hinton and others have suggested that less structure may be sufficient. The speaker notes that this debate highlights the need for further research and development in the field of artificial intelligence, as we strive to create machines that can think and reason like humans.

The Importance of Knowledge Acquisition

One of the biggest challenges facing researchers working on reasoning with neural networks is knowledge acquisition. This refers to the process of encoding knowledge into a machine learning system, which can then be used to make predictions or take actions.

The speaker notes that this challenge is particularly difficult when it comes to representing knowledge using logic and symbols. Traditional approaches often rely on human experts to encode knowledge into a graph or rules-based system, which can be brittle and inflexible.

In contrast, geoff hinton suggests replacing symbols with vectors and manipulating objects in a space. This approach may lead to new forms of reasoning that are similar to simple expert systems. However, the challenge lies in finding a way to balance the need for precision and optimization with the need for flexibility and adaptability.

The Future of Reasoning with Neural Networks

As researchers continue to explore the possibilities of reasoning with neural networks, it is clear that this field holds great promise for advancing our understanding of human cognition and intelligence. While there are many challenges to overcome, including knowledge acquisition and the role of energy minimization in reasoning, the potential rewards are significant.

By developing machines that can think and reason like humans, we may be able to create systems that are capable of complex problem-solving and decision-making. This could have far-reaching implications for fields such as robotics, autonomous vehicles, and healthcare.

Ultimately, the future of reasoning with neural networks will depend on our ability to develop new algorithms and techniques that can enable machines to learn and reason like humans. As researchers continue to push the boundaries of this field, we may uncover new insights into human cognition and intelligence that have far-reaching implications for society as a whole.

"WEBVTTKind: captionsLanguage: endo you think neural networks can be made to reason yes there's no question about that again we have a good example right the question is is how so the question is how much prior structure you have to put in the neural net so that something like human reasoning will emerge from it you know from running another question is all of our kind of model of what reasoning is that are based on logic are discrete and and are therefore incompatible with gradient based learning and I was very strong believer in this idea granion baserunning I don't believe that other types of learning that don't use kind of gradient information if you want so you don't like discrete mathematics you don't like anything discrete well that's it's not that I don't like it it's just that it's it's incompatible with learning and I'm a big fan of running right so in fact that's perhaps one reason why deep learning has been kind of looked at with suspicion by a lot of computer scientists because the math is very different the math that you use for deep running you know we kind of as more to do with you know cybernetics the kind of math you do in electrical engineering than the kind of math you doing computer science and and you know nothing in in machine learning is exact right computer science is all about sort of you know obviously compulsive attention to details of like you know every index has to be right and you can prove that an algorithm is correct right machine learning is the science of sloppiness really that's beautiful so okay maybe let's feel around in the dark of what is a neural network that reasons or a system that is works with continuous functions that's able to do build knowledge however we think about reasoning builds on previous knowledge build on extra knowledge create new knowledge generalize outside of any training set ever built what does that look like if yeah they may be do you have Inklings of thoughts of what that look like well yeah I mean yes or no if I had precise ideas about this I think you know we'd be building it right now but and there are people working on this or whose main research interest is actually exactly that right so what you need to have is a working memory so you need to have some device if you want some subsystem they can store a relatively large number of factual episodic information for you know a reasonable amount of time so you you know in the in the brain for example it kind of three main types of memory one is the sort of memory of the the state of your cortex and that sort of disappears within 20 seconds you can't remember things for more than about 20 seconds or a minute if you don't have any other form of memory the second type of memory which is longer-term is a short-term is the hippocampus so you can you know you came into this building you remember whether where the the exit is where the elevators are you have some map of that building that's stored in your hippocampus you might remember something about what I said you know a few minutes ago and forgot all right if it starts being raised but you know but that there's a marker in your hippocampus and then the the longer term memory is in the synapse the synapses right so what you need if you want for a system that's capable of reasoning is that you want the hippocampus like thing right and that's what people have tried to do with memory networks and you know in altering machines and stuff like that right and and now with transformers which have sort of a memory in they're kind of self attention system you can you can think of it this way so so that's one element you need another thing you need is some sort of network that can access this memory get an information back and then kind of crunch on it and then do this iteratively multiple times because a chain of reasoning is a process by which you you you can you update your knowledge about the state of the world about you know it's gonna happen etc and that there has to be this sort of recurrent operation basically and you think that kind of if we think about a transformer so that seems to be too small to contain the knowledge that's that's to represent the knowledge as containing Wikipedia for example well transformer doesn't have this idea of recurrence it's got a fixed number of layers and that's the number of steps that you know limits basically it's a representation but recurrence would build on the knowledge somehow I mean yeah it would evolve the knowledge and expand the amount of information perhaps or useful information within that knowledge yeah but is is this something that just can emerge with size because it seems like everything we have now is just no it's not it's not it's not clear how you access and write into an associative memory in efficient way I mean sort of the original memory network maybe had something like the right architecture but if you try to scale up a memory network so that the memory contains all Wikipedia it doesn't quite work right so so this is a need for new ideas there okay but it's not the only form of reasoning so there's another form of reasoning which is true which is very classical so in some types of AI and it's based on let's call it energy minimization okay so you have some sort of objective some energy function that represents the the the quality or the negative quality okay energy goes up when things get bad and they get low when things get good so let's say you you want to figure out you know what gestures do I need to to do to grab an object or walk out the door if you have a good model of your own body a good model of the environment using this kind of energy minimization you can make a you can make you can do planning and it's in optimum control is called it's called market model predictive control you have a model of what's gonna happen in the world as consequence of your actions and that allows you to buy energy minimization figure out a sequence of action that optimizes a particular objective function which measures you know minimize the number of times you're gonna hit something and the energy gonna spend doing the gesture and etc so so that's performer reasoning planning is a form of reasoning and perhaps what led to the ability of humans to reason is the fact that or you know species you know that appear before us had to do some sort of planning to be able to hunt and survive and survive the winter in particular and so you know it's the same capacity that you need to have so in your intuition is if you look at expert systems in encoding knowledge as logic systems as graphs in this kind of way is not a useful way to think about knowledge graphs are your brittle or logic representation so basically you know variables that have values and constraint between them that are represented by rules is real too rigid and too brittle right so one of the you know some of the early efforts in that respect were were to put probabilities on them so a rule you know you know if you have this in that symptom you know you have this disease with that probability and you should prescribe that antibiotic with that probability right this mice in system from the 70s and that that's what that branch of AI led to you know Bayesian networks in graphical models and causal inference and viral you know yeah method so so there there is I mean certainly a lot of interesting work going on in this area the main issue with this is is knowledge acquisition how do you reduce a bunch of data to a graph of this type near relies on the expert on a human being to encode at add knowledge and that's especially in practical question the second question is do you to represent knowledge symbols and you want to manipulate them with logic and again that's incomparable we're learning so one suggestion with geoff hinton has been advocating for many decades is replace symbols by vectors think of it as pattern of activities in a bunch of neurons or units or whatever you wanna call them and replace logic by continuous functions okay and that becomes now compatible there's a very good set of ideas by written in a paper about 10 years ago by Leon go-to on who is here at Facebook the title of the paper is for machine learning to machine reasoning and his idea is that learning learning systems should be able to manipulate objects that are in the sense fit in a space and then put the result back in the same space so is this idea of working memory basically and it's a very enlightening and in the sense that might learn something like the simple expert systems I mean it's with you can learn basic logic operations there yeah quite possibly yeah this is a big debate on sort of how much prior structure you have to put in for this kind of stuff to emerge that's the debate I have with Gary Marcus and people like that you\n"