Research map / Multi-agent systems and games / Cooperative multi-agent RL
Explaining Decentralized Multi-Agent Reinforcement Learning Policies
Kayla Boggess, Sarit Kraus, Lu Feng
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 |
|---|---|---|---|
| use Hasse diagram summarization instead of per-agent abstract policy graphs | decentralized MARL policy summary · SR (9,7) CTDE | number of nodes: 534 -> 8 | Improved |
| use Hasse diagram summarization instead of per-agent abstract policy graphs | decentralized MARL policy summary · RW (4,19) CTDE | number of nodes: 1,274 -> 20 | Improved |
| use Hasse diagram summarization instead of per-agent abstract policy graphs | decentralized MARL policy summary · PP (7,6) DTDE | number of nodes: 107 -> 7 | Improved |
| show HDS summaries to users instead of baseline graphs | user question answering · Summarization user study | correct answers out of 6: 3.1 -> 4.25 | Improved |
| use HDS instead of baseline | user-rated completeness · Summarization user study | Likert completeness: W=16.0, Z=-2.07, p <= 0.04 | Improved |
| use HDS instead of baseline | user-rated summary quality on other dimensions · Summarization user study | Likert ratings | No clear effect |
| use HD-When explanations (with uncertain features) instead of union baseline | explanation size · SR (9,7) CTDE | certain features: 54 -> 9 | Improved |
| use HD-When instead of baseline | explanation size · RW (4,19) CTDE | certain features: 267 -> 0 | Mixed |
| use HD explanations instead of baseline for When queries | user prediction accuracy · Explanation user study | correct answers (when): t(20)=9.65, d=2.16 | Improved |
| use HD explanations instead of baseline for Why Not queries | user prediction accuracy · Explanation user study | correct answers (why not): t(20)=13.23, d=2.96 | Improved |
| use HD explanations instead of baseline for What queries | user prediction accuracy · Explanation user study | correct answers (what): t(20)=12.05, d=2.69 | Improved |
| use HD explanations instead of baseline | user ratings of explanation quality · Explanation user study | Likert understanding: W=3.5, Z=-3.02, r=-0.47 | Improved |
Closest papers by meaning
- Decomposing Temporal Equilibrium Strategy for Coordinated Distributed Multi-Agent Reinforcement Learning (AAAI 2024 · similarity 0.88)
- Offline Multi-Agent Reinforcement Learning via In-Sample Sequential Policy Optimization (AAAI 2025 · similarity 0.88)
- Cooperative Policy Agreement: Learning Diverse Policy for Offline MARL (AAAI 2025 · similarity 0.88)
- Cautiously-Optimistic Knowledge Sharing for Cooperative Multi-Agent Reinforcement Learning (AAAI 2024 · similarity 0.88)
- Foundations of Multi-Agent Learning in Dynamic Environments: Where Reinforcement Learning Meets Strategic Decision-Making (AAAI 2025 · similarity 0.88)
- Dispelling the Mirage of Progress in Offline MARL through Standardised Baselines and Evaluation (NeurIPS 2024 · 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.
