Index Ai Academy
Glossary · Phase 3 · Machine learning

REINFORCE

The simplest policy gradient method: after each episode, move each action's preference by the step size times the return that followed it.

Try it now

The three probabilities always add to 1: what one action gains, the others lose.

Learn it properly

Policy gradients: learning what to do directly · A policy as probabilities you can nudge: turn up what worked, measure it against a baseline, and see how the same update trains a language model. About 40 minutes.

Other terms from the same lesson

  • advantage: In reinforcement learning, how much better an outcome was than expected: the return that followed an action minus a baseline.
  • policy gradient: Training a policy directly by making actions that were followed by better-than-expected returns more likely, and the rest less.
Keep learning

More in plain English.

1286 terms in the glossary, each with the lesson that teaches it.

1286 terms

The glossary

Every word you'll meet in AI and machine learning, explained in everyday language.

Phase 3 · lesson 40 min

Policy gradients: learning what to do directly

A policy as probabilities you can nudge: turn up what worked, measure it against a baseline, and see how the same update trains a language model.

Never trained an AI? Teach a small one to write, right in your browser. It takes about 20 seconds and needs no account. Teach it to write