Stone to Silicon

Artificial intelligence · 2019

MuZero

Mastering games without being told the rules, by learning its own model.

MuZero learned a latent dynamics model and planned inside it, matching AlphaZero in Go, chess and shogi and setting records on Atari. Learned world models for planning later found use in video compression and robotics.

See it in the living treeWikipedia

Arcs above join it to what it stood on, arcs below to what it led to. The bar places this stretch on the tree's whole timeline.

What it built on 1

Related measurements

The dashed line marks when it first worked.

Training compute of notable AI models Source: Epoch AI, Notable AI Models (CC BY 4.0)

ML hardware FLOP/s per dollar Source: Epoch AI, ML hardware (CC BY 4.0)