AI Paper Atlas

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

Off-Policy Action Anticipation in Multi-Agent Reinforcement Learning

Ariyan Bighashdel, Daan de Geus, Pavol Jancura, Gijs Dubbelman

JMLR 2024 · Multi-agent systems and games › Cooperative multi-agent RL · approach: Offline and networked MARL · 1 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
use action anticipation (LOLA-OffPA2) instead of policy parameter anticipation (LOLA-DiCE)LOLA · Exit-Room level 3normalized average episode reward: 0.56 -> 0.93Improved
use LOLA-OffPA2 instead of LOLA-DiCELOLA · Exit-Room, 4th-order reasoninglearning anticipation time complexity: 5.41 -> 0.94Improved
use LOLA-OffPA2 instead of LOLA-DiCELOLA · Exit-Room level 1normalized average episode reward: 0.91 -> 1.00Improved
use HLA-OffPA2 instead of MADDPGMADDPG · Reacher (Mujoco)normalized average episode reward: 0.02 -> 0.42Improved
use HLA-OffPA2 instead of MADDPGMADDPG · Predator-Preynormalized average episode reward: 0.21 -> 0.44Improved
use HLA-OffPA2 instead of CPG-MADDPGCPG-MADDPG · Half-Cheetahnormalized average episode reward: 0.88 -> 0.94Improved
add sorting of hierarchy levels by shaping capacity (ablation)HLA-OffPA2 · Walkernormalized average episode reward: 0.63 -> 0.67Improved
add sorting of hierarchy levels by shaping capacity (ablation)HLA-OffPA2 · Cooperative Navigationnormalized average episode reward: 0.85 -> 0.88Improved
add sorting of hierarchy levels by shaping capacity (ablation)HLA-OffPA2 · Predator-Preynormalized average episode reward: 0.38 -> 0.44Improved
raise projected prediction length from tuned value to 1.3 (ablation)LOLA-OffPA2 · Exit-Room level 3normalized average episode rewardHurt
lower projected prediction length to 0.1 (ablation)LOLA-OffPA2 · Exit-Room level 3normalized average episode rewardHurt
use LA-OffPA2 instead of MADDPG-based methodsMADDPG · Iterated Rotational Gamedistance to equilibriumImproved

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