Joshua ElkingtonFounder and General Partner at Axial

Demis Hassabis, AI and systems neuroscience

Name: Demis Hassabis, AI and systems neuroscience
Uploaded: 2024-05-01T15:56:05+00:00
Duration: 1 min
Channel: Joshua Elkington
Description: In this segment, Demis Hassabis explains how modern learning systems at DeepMind are rooted in ideas drawn from systems neuroscience, which historically…

1 May 2024 · 1 min

Watch on YouTube

1:00

In this segment, Demis Hassabis explains how modern learning systems at DeepMind are rooted in ideas drawn from systems neuroscience, which historically validated concepts such as reinforcement learning and neural networks by comparing them to brain function. He describes DeepMind's distinctive approach as a fusion of two algorithmic components: deep neural networks that serve as models of the environment, data, and experience, and the utilization of these models for further learning and decision-making. This perspective positions deep learning as a world‑modeling tool within a broader learning system architecture.

Chapters

Originally, these kind of learning systems took a lot of inspiration and also could be validated— some of the ideas like reinforcement learning and neural networks— by systems neuroscience and comparing what these systems do, comparing them on a systems and algorithmic level to what we know about how the brain works. Now, everything we do at DeepMind, of course, is on the learning system side, and we've been lucky enough to be in the vanguard of this almost revolution or renaissance in the last decade of these types of approaches. So, how do we think about— what's our sort of, I guess, special take on learning systems and how powerful they can be? So, there are kind of two component algorithms or approaches, one could say, that we've fused together. So, of course, there's deep learning or deep neural networks, and the way I think about this is that the deep neural network system is is there to build a model of the environment, of the data and the experience. And then what do you use that model for?