Research map / Multi-agent systems and games / Cooperative multi-agent RL
SocialJax: An Evaluation Suite for Multi-agent Reinforcement Learning in Sequential Social Dilemmas
Zihao Guo, Shuqing Shi, Richard Willis, Tristan Tomilin, Joel Z Leibo, Yali Du
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 changed | Where | Measured | Effect |
|---|---|---|---|
| raise SVO angle from 0 to 90 degrees (ablation) | SVO IPPO · Clean Up | return: 0.02 -> 1410.53 | Improved |
| raise SVO angle from 0 to 90 degrees (ablation) | SVO IPPO · Coins | return: 11.81 -> 162.46 | Improved |
| raise SVO angle from 0 to 90 degrees (ablation) | SVO IPPO · Gift Refinement | return: 104.29 -> 227.95 | Improved |
| raise SVO angle from 0 to 45 degrees (ablation) | SVO IPPO · Harvest: Open | return: 79.71 -> 68.28 | Hurt |
| use common reward instead of individual reward | IPPO · SocialJax environments | return | Improved |
| add reward exchange ratio 0.5 to IPPO (ablation) | IPPO · Clean Up | mean return: 970 -> 1780 | Improved |
| replace common reward with reward exchange (ablation) | IPPO-CR · Gift Refinement | mean return: 505 -> 342 | Hurt |
| add reward exchange ratio 0.25 to individual reward IPPO (ablation) | IPPO · Coop Mining | mean return: 446 -> 727 | Improved |
| use common reward instead of reward exchange ratio 0.75 (ablation) | IPPO-RE · Harvest: Open | mean return: 393 -> 420 | Improved |
| use centralized critic MAPPO instead of IPPO-CR | IPPO-CR · Clean Up | cooperation metric: 109.54 -> 133.99 | Improved |
| use MAPPO instead of IPPO-CR | IPPO-CR · Gift Refinement | cooperation metric: 28.37 -> 20.24 | Hurt |
| use non-shared parameter IPPO instead of shared (ablation) | P-IPPO · SocialJax environments | training time and convergence | No clear effect |
Closest papers by meaning
- FightLadder: A Benchmark for Competitive Multi-Agent Reinforcement Learning (ICML 2024 · similarity 0.86)
- ALSO: Adversarial Online Strategy Optimization for Social Agents (ICML 2026 · similarity 0.86)
- JaxMARL: Multi-Agent RL Environments and Algorithms in JAX (NeurIPS 2024 · similarity 0.85)
- Group-Graph Policy Optimization for Long-Horizon Agentic Reinforcement Learning (NeurIPS 2026 · similarity 0.85)
- Melting Pot Contest: Charting the Future of Generalized Cooperative Intelligence (NeurIPS 2024 · similarity 0.85)
- SrSv: Integrating Sequential Rollouts with Sequential Value Estimation for Multi-agent Reinforcement Learning (AAAI 2025 · similarity 0.85)
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.
