Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Transformer-based Working Memory for Multiagent Reinforcement Learning with Action Parsing
Yaodong Yang, Guangyong Chen, Weixun Wang, Xiaotian Hao, Jianye HAO, Pheng-Ann Heng
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 |
|---|---|---|---|
| add Agent Transformer Memory (ATM) in place of GRU agent network | QMIX and QPLEX · SMAC 4m_vs_5m, 5m_vs_6m, 6h_vs_8z, corridor | test win rate | Improved |
| use ATM instead of GRU agent network | QMIX and QPLEX · SMAC 6h_vs_8z and corridor | test win rate | Improved |
| add memory slots to ATM (vs w/o mem) (ablation) | ATM on QMIX · SMAC 4m_vs_5m, 5m_vs_6m | test win rate | Improved |
| add Entity-Bound Action Layer to ATM (vs w/o EBA) (ablation) | ATM on QMIX · SMAC 4m_vs_5m, 5m_vs_6m | test win rate | Improved |
| add both memory and Entity-Bound Action Layer to basic transformer (trm) (ablation) | ATM on QMIX · SMAC 4m_vs_5m, 5m_vs_6m | test win rate | Improved |
| use working memory updating instead of relational memory or AMRL (ablation) | ATM · SMAC 4m_vs_5m, 5m_vs_6m | test win rate | Improved |
| add any memory updating scheme (working, relational, AMRL, GRU) vs no memory (ablation) | ATM · SMAC 4m_vs_5m, 5m_vs_6m | test win rate | Improved |
| vary memory slot number from 0 to 5 (ablation) | ATM · SMAC 4m_vs_5m, 5m_vs_6m | episodic return | Mixed |
| use ATM instead of GRU or MLP agent network | MAPPO and MAA2C · LBF 15x15_3p_5f, 15x15_4p_5f, 15x15_4p_6f | episodic return | Improved |
| use ATM instead of GRU or MLP agent network | MAPPO · LBF 15x15_4p_6f | episodic return | Improved |
Closest papers by meaning
- Memo: Training Memory-Efficient Embodied Agents with Reinforcement Learning (NeurIPS 2025 · similarity 0.89)
- MemGen: Weaving Generative Latent Memory for Self-Evolving Agents (ICLR 2026 · similarity 0.88)
- MemEvolve: Meta-Evolution of Agent Memory Systems (ICML 2026 · similarity 0.88)
- Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions (ICLR 2026 · similarity 0.88)
- G-Memory: Tracing Hierarchical Memory for Multi-Agent Systems (NeurIPS 2025 · similarity 0.88)
- ViLoMem: Agentic Learner with Grow-and-Refine Multimodal Semantic Memory (CVPR 2026 · similarity 0.88)
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.
