Research map / Multi-agent systems and games / Cooperative multi-agent RL
Multiagent Evaluation Mechanisms
Tal Alon, Magdalen Dobson, Ariel Procaccia, Inbal Talgam-Cohen, Jamie Tucker-Foltz
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 a linear mechanism with beta=(1,1) instead of per-agent mechanisms | two-agent study-vs-cheat instance · Example 2.1, p=6 | all agents incentivized to study: marginal benefit 10 vs 9 | Improved |
| restrict to linear mechanisms when conversion parameter p drops to 4 | two-agent instance · Example 2.1, p=4 | existence of mechanism incentivizing both agents: 4β1 ≥ 5β2 and 4β2 ≥ 5β1 | Hurt |
| use nonlinear monotone mechanism min(F1,F2) instead of linear | two-agent instance · Example 2.1, p=4 | existence of mechanism incentivizing both agents | Improved |
| lower p to 1 (cheating relatively more attractive) | two-agent instance · Example 2.1, p=1 | existence of any monotone mechanism: M(4, 1) > M(1, 4) | Hurt |
| restrict feature count n to a constant when incentivizing all agents to admissible actions with monotone mechanisms | Problem 2 · complexity | computational complexity | Improved |
| switch goal from all agents to maximum number of agents (monotone) | Problems 3 and 4 · complexity | computational complexity | Hurt |
| restrict mechanism class from monotone to linear for maximizing incentivized agents | Problems 7 and 8 · complexity | computational complexity | Mixed |
Closest papers by meaning
- Agent-as-a-Judge: Evaluate Agents with Agents (ICML 2025 · similarity 0.84)
- ChatEval: Towards Better LLM-based Evaluators through Multi-Agent Debate (ICLR 2024 · similarity 0.82)
- Holistic Agent Leaderboard: The Missing Infrastructure for AI Agent Evaluation (ICLR 2026 · similarity 0.82)
- Lessons Learned: A Multi-Agent Framework for Code LLMs to Learn and Improve (NeurIPS 2025 · similarity 0.82)
- Precise but Uncoupled: Reviewer Precision Does Not Guarantee Critique Uptake in Multi-Agent Math Reasoning (NeurIPS 2026 · similarity 0.81)
- SKILL-MIX: a Flexible and Expandable Family of Evaluations for AI Models (ICLR 2024 · similarity 0.81)
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.
