AI Paper Atlas

Research map / Multi-agent systems and games / Cooperative multi-agent RL

ELIGN: Expectation Alignment as a Multi-Agent Intrinsic Reward

Zixian Ma, Rose E Wang, Li Fei-Fei, Michael S. Bernstein, Ranjay Krishna

NeurIPS 2022 · Multi-agent systems and games › Cooperative multi-agent RL · approach: MARL benchmarks and grouping · 6 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
Add ELIGN-self intrinsic reward instead of sparse reward onlySAC decentralized · Cooperative navigation 3v0, partial obsmean test episode reward: 139.07 -> 155.88Improved
ELIGN-self instead of curiosity-self intrinsic rewardSAC decentralized · Cooperative navigation 3v0, partial obsmean test episode reward: 133.93 -> 155.88Improved
Add ELIGN-adv (misalign with adversary expectations) instead of sparseSAC decentralized · Predator-prey 2v2, partial obsmean test episode reward: -4.72 -> -2.51Improved
Add ELIGN-adv instead of sparseSAC decentralized · Keep-away 2v2, partial obsmean test episode reward: 4.58 -> 19.46Improved
Add ELIGN-adv instead of sparseSAC decentralized · Google football 3v1 with keepermean test episode reward: 0.020 -> 0.025Improved
Add ELIGN-team instead of sparseSAC decentralized · Physical deception 2v1, partial obsmean test episode reward: 93.60 -> 101.72Improved
Add ELIGN-self instead of sparseSAC decentralized · Physical deception 2v1, partial obsmean test episode reward: 93.60 -> 69.91Hurt
Add ELIGN-team instead of sparseSAC decentralized · Keep-away 2v2, partial obsmean test episode reward: 4.58 -> 2.96Hurt
Misalign with adversaries (ELIGN-adv) rather than align with team (ELIGN-team) (ablation)SAC decentralized · Predator-prey 2v2, partial obsmean test episode reward: -7.69 -> -2.51Improved
Scale agents to 6 in heterogeneous navigation; ELIGN-self vs curiosity-selfSAC decentralized · Heterogeneous navigation 6v0mean test episode reward: 702.73 -> 646.70Hurt
Scale to 5 agents; ELIGN-self instead of sparseSAC decentralized · Cooperative navigation 5v0mean test episode reward: 459.92 -> 498.24Improved
Add ELIGN-self to centralized SAC under partial observabilitySAC centralized · Cooperative navigation 3v0mean test episode reward: 113.25 -> 137.14Improved

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