Learning in a world of notes

Reinforcement learning helps agents use notes more accurately and cheaply, with gains that persist as notes change. Results across three workspaces.

saved by

• Maanav Khaitan