AI Paper Atlas

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

Gossip-based Actor-Learner Architectures for Deep Reinforcement Learning

Mahmoud Assran, Joshua Romoff, Nicolas Ballas, Joelle Pineau, Michael Rabbat

NeurIPS 2019 · Multi-agent systems and games › Cooperative multi-agent RL · approach: Decentralized actor-critic MARL · 22 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
replace synchronous AllReduce gradient averaging with asynchronous gossipA2C · Atari BeamRider (40M steps)game score: 9829 -> 10188Improved
replace AllReduce with gossipA2C · Atari Breakout (40M steps)game score: 495 -> 690Improved
replace AllReduce with gossipA2C · Atari Seaquest (40M steps)game score: 1894 -> 1892No clear effect
replace AllReduce with gossipA2C · Atari SpaceInvaders (40M steps)game score: 3021 -> 3074No clear effect
use gossip with many simulatorsA2C · six Atari games, 600 runspercentage of successful runsImproved
use gossip in place of exact gradient averagingA2C · Atarisample efficiencyNo clear effect
use gossip in place of AllReduceA2C · GPU hardware utilizationGPU utilization: 10-fold improvementImproved
use GALA-A2C at 25M steps instead of IMPALA at 50MGALA-A2C · Atari Qbertgame score: 18902 -> 18810Mixed

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