AI Paper Atlas

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

Biases for Emergent Communication in Multi-agent Reinforcement Learning

Tom Eccles, Yoram Bachrach, Guy Lever, Angeliki Lazaridou, Thore Graepel

NeurIPS 2019 · Multi-agent systems and games › Cooperative multi-agent RL · approach: Learned inter-agent communication · 28 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 positive signalling lossIndependent REINFORCE agents · Summing MNIST digitsproportion of good runs: 0.00 -> 0.98Improved
Add positive listening lossIndependent REINFORCE agents · Summing MNIST digitsproportion of good runs: 0.00 -> 0.88Improved
Add both positive signalling and positive listening lossesIndependent REINFORCE agents · Summing MNIST digitsproportion of good runs: 0.00 -> 1.00Improved
Add speaker-side social influence rewardIndependent REINFORCE agents · Summing MNIST digitsproportion of good runs: 0.00 -> 0.00No clear effect
Add positive signalling lossIndependent A3C agents with V-trace · Treasure Huntproportion good: 0.28 -> 0.84Improved
Add positive listening lossIndependent A3C agents with V-trace · Treasure Huntproportion good: 0.28 -> 0.64Improved
Add both positive signalling and positive listening lossesIndependent A3C agents with V-trace · Treasure Huntfinal reward: 12.45 \pm 0.48 -> 15.14 \pm 0.33Improved
Add speaker-side social influence rewardTreasure Hunt agents · Treasure Hunttask rewardNo clear effect
Use single-timestep CIC instead of multiple-timestep CIC (ablation)Positive listening · Treasure Hunttask rewardNo clear effect
Set the cross-entropy loss weight to zero (ablation)Positive listening · Summing MNIST digitsfinal rewardHurt
Set the cross-entropy loss weight to zero (ablation)Positive listening · Treasure Hunttask rewardNo clear effect

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