AI Paper Atlas

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

DMAP: a Distributed Morphological Attention Policy for learning to locomote with a changing body

Alberto Chiappa, Alessandro Marin Vargas, Alexander Mathis

NeurIPS 2022 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Imitation and skill learning · 8 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
remove morphology encoding at test time (DMAP-ne) (ablation)DMAP · Walker sigma=0.1mean episode reward: 1229 -> 337Hurt
remove morphology encoding at test time (ablation)DMAP · Hopper sigma=0.1mean episode reward: 1842 -> 984Hurt
remove morphology encoding at test time (ablation)DMAP · Half Cheetah sigma=0.1mean episode reward: 2261 -> 1493Hurt
remove morphology encoding at test time (ablation)DMAP · Ant sigma=0.5mean episode reward: 960 -> 881Hurt
replace TCN encoder with DMAP attention encoder and distributed controllers (ablation)TCN · Ant sigma=0.3mean episode reward: 251 -> 1623Improved
replace TCN encoder with DMAP (ablation)TCN · Walker sigma=0.3mean episode reward: 518 -> 893Improved
replace TCN encoder with DMAP (ablation)TCN · Hopper sigma=0.3mean episode reward: 1368 -> 1316No clear effect
replace TCN encoder with DMAP (ablation)TCN · Half Cheetah sigma=0.1mean episode reward: 2278 -> 2261No clear effect
give Oracle morphology parameters instead of proprioception only (Simple)Simple · Ant sigma=0.5mean episode reward: 391 -> 974Improved
give Oracle morphology parameters instead of proprioception onlySimple · Ant sigma=0.1mean episode reward: 2164 -> 2148No clear effect
give Oracle morphology parametersSimple · Walker sigma=0.1mean episode reward: 1909 -> 1780Hurt
replace Oracle with two-step RMA regressed encodingOracle · Hopper sigma=0.5mean episode reward: 919 -> 730Hurt

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