Matt Botvinick is the director of neuroscience research at DeepMind
Host: Lex Fridman (@lexfridman)
Our brain allows us to understand all we know of life, yet it is the most mysterious and baffling aspect of life
  • “My sense of wonder comes not from the distant, mysterious stars, but the extremely, intimately close brain” – Matt Botvinick
We understand much of the high-level functions that the brain performs, but much less so of the physical mechanisms behind how those functions operate
Matt sees cognitive science, neuroscience, and psychology all as different branches of knowledge stemming from the brain with each field only capturing different aspects of the truth
  • Understanding the foundational truths of the mind will require piecing together all the knowledge we have of the brain rather than myopically focusing on neuronal voltage channels or behavioral conditioning
Cognitive neuroscience today is in a place similar to where genetics was after Mendel established the heritability of traits but before Watson and Crick discovered the double-helix structure of DNA
  • We knew DNA was involved in the passing of genetic information but we didn’t know how. The knowledge of the double-helix allowed scientists to uncover the mechanical process of replication that drove the passing of genes. 
  • Today, we know the interactions between neurons drives our thoughts and behaviors, but we have yet to understand the exact causal processes behind how electrical signals eventually turn into philosophical musings on consciousness
Although thoughts occur at an individual level, some thoughts can only be understood within a group context
  • “It is very possible that intelligence can only arise when there are multiple intelligences” – Lex Fridman
  • A mystery of the mind, which when solved will be greatly productive for AI, is understanding the forces that drive people to rally behind ideas
Cognition stems not only from the hardware it’s operating on (brain, circuitry) but also the larger environment in which it exists 
  • Thoughts need a reality to test itself against otherwise they are nothing more than screams into the void
  • “It’s made most clear in reinforcement learning where you can only learn as much as you can simulate” – Lex Fridman
The prefrontal cortex is the cerebral cortex (outer brain layer) region around your forehead that is associated with planning, decision making, and personality expression
  • The prefrontal cortex is involved in controlled, willful thoughts and behaviors rather than our habitual, automatic thoughts and behaviors
Researchers and doctors noted that the brain-damaged veterans of WW1 who weren’t completely incapacitated and were able to live “normal” lives tended to have damaged prefrontal cortices
  • The most significant side-effect for these veterans was that they couldn’t process new information
“When you look carefully at functional differentiation in the brain what you usually end up concluding is that the difference between regions is graded rather than being discrete” – Matt Botvinick
  • You can’t simply divide the brain up between regions that perform different roles, it operates as a whole system
  • “Even regions whose function is pretty well defined, at a coarse grain, are nonetheless carrying some information about very different domains”
Neurons communicate through neurotransmitters that are emitted between each other that change the voltage of the neurons to a point where an electrical signal is eventually released
“What matters is how quickly an individual neuron is spiking” – Matt Botvinick
  • “There’s still uncertainty about whether that’s an adequate description of how information is transmitted within the brain”
  • “There are studies that suggest that the precise timing of spikes matters, there are studies that suggest that there are computations that go on within the dendritic tree that are quite rich in structure, and that really don’t equate to anything that we’re doing in our artificial neural networks”
Currently, the AI community views the activity rate of their neural networks as analogous to the spike rate of neurons within the nervous system
  • The patterns of neural networks seem “hauntingly similar” to the patterns of normal brain activity
“Meta-learning, by definition, is a situation in which you have a learning algorithm and the learning algorithm operates in such a way that it gives rise to another learning algorithm” – Matt Botvinick
  • “It relates to the old idea in psychology of learning to learn, situations where you have experiences that make you better at learning something new”
  • It’s like how learning new languages is hard at first, but you eventually anticipate what you need to learn so you can learn more languages more easily
These learning dynamics also resemble the activation patterns of the prefrontal cortex that learns based on the reinforcement mechanisms of rewards from certain behaviors 
  • The question that arises from this is how the prefrontal cortex was programmed to learn from these methods 
Based on these findings, it seems to be a natural fact that any memory system shaped by reinforcement learning will eventually result in meta-learning
  • “This will happen if the system is being trained in a setting where there is a sequence of tasks that all share some abstract structure” – Matt Botvinick
“A blind spot for at least robotics is human-robot, human-agent interaction” – Lex Fridman
  • “How do we keep these things from getting out of control; how do we keep them from doing things that harm humans?” – Matt Botvinick
Understanding even the first-order effects of robotics and AI on society will never be grasped solely by theory crafting in a research setting
  • Artificial meta-learning systems must learn not only how to perform their goals, but also how to do them in a manner that’s harmonious with people
  • For example, part of what gives meaning to life is accomplishing difficult tasks. If AI did everything for us would we even have a life worth living?
It would be beneficial to introduce non-engineers in the development of AI in order to make AI work harmoniously with society and not solely focus on technical efficiency 
  • The effects of AI will touch all aspects of society so it might be necessary to include economists, legal experts, and philosophers in its development
  • But then the question arises of whether or not these professionals even know what’s best for society
If we reach the point of artificial general intelligence (AGI) should we even make them like humans? Would it be better to have an emotionless, calculating AGI or a passionate, cultured AGI?
The best people are capable of great feats while also having great depth to their soul
  • On the other hand, AI research focuses only on the capabilities of their algorithms
“What would it mean for an AI system to display caring, compassionate behavior in a way that actually made us feel like it was for real?” – Matt Botvinick
  • “That’s the ultimate Turing test”
In Matt’s eyes, it’s impossible to handcraft an AI capable of love, but it’s possible to point AI in the direction to learn what love is