AI Paper Atlas

Research map / Multi-agent systems and games / Human-AI and goal-conditioned

Self-Explaining Deviations for Coordination

Hengyuan Hu, Samuel Sokota, David J Wu, Anton Bakhtin, Andrei Lupu, Brandon Cui, Jakob Nicolaus Foerster

NeurIPS 2022 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Boundedly rational agent modeling · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Search for a joint self-explaining deviation from the no-op blueprintIMPROVISED^E and IMPROVISED^P · Trampoline-Tigeroptimal strategy discovery: 100%Improved
Search for a joint self-explaining deviation from the no-op blueprintIMPROVISED^E · Trampoline-Tigerexpected return: 0.1Improved
Use MAPPO instead of joint SED searchTrampoline-Tiger · Trampoline-Tigeroptimal strategy discovery: 2/20Hurt
Search for joint deviations by Players 1 and 2IMPROVISED^E · Hanabi, 200 finesse-able situations where Player 3 responds correctlyfinesses: 15 -> 51Improved
Choose deviations to maximize probability of improvement rather than expected returnIMPROVISED^P · Hanabi, 200 finesse-able situations where Player 3 responds correctlyfinesses: 51 -> 62Improved
Use joint self-explaining deviations during the full gameIMPROVISED^E · Hanabi, full game, 100 game seedsfinesses: 3 -> 29Improved
Use test-time deviations at finesse-able situationsIMPROVISED^E · Hanabi, finesse-able situationsaverage score: 17.52 -> 18.58Improved
Use test-time deviations at finesse-able situationsIMPROVISED^P · Hanabi, finesse-able situationsaverage score: 17.52 -> 18.36Improved
Use test-time deviations during the full gameIMPROVISED^E · Hanabi, full gameaverage score: 17.8 -> 23.54Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy