Chapter

Simulation of Large Materials using Reinforcement Learning
listen on SpotifyListen on Youtube
1:21:47 - 1:26:26 (04:39)

The problem of simulating large materials can be tackled using reinforcement learning prediction problem which can simulate larger and more complex molecules to understand how they interact with each other, the electron clouds and their resulting shapes. This was done in a Nature study last year where plasma was held in specific shapes using these approximations.

Clips
The challenge of simulating large materials has been tackled with reinforcement learning prediction models, as traditional simulation methods are only able to simulate small and simple molecules.
1:21:47 - 1:26:26 (04:39)
listen on SpotifyListen on Youtube
Reinforcement Learning
Summary

The challenge of simulating large materials has been tackled with reinforcement learning prediction models, as traditional simulation methods are only able to simulate small and simple molecules. This approach allows for better approximations of electron cloud interactions in specific shapes, as described in a Nature paper.

Chapter
Simulation of Large Materials using Reinforcement Learning
Episode
#299 – Demis Hassabis: DeepMind
Podcast
Lex Fridman Podcast