Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Collaborative Decision Making Using Action Suggestions
Dylan Asmar, Mykel Kochenderfer
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 scaled-rational suggestion update (τ=0.99) vs no suggestions | POMDP agent · Tag | mean reward: -10.7 -> -1.8 | Improved |
| add noisy-rational suggestion update (λ=5.0) vs no suggestions | POMDP agent · RockSample(7,8,20,0) | mean reward: 21.5 -> 27.5 | Improved |
| belief-update with suggestion (scaled τ=0.99) instead of naively following (ν=1.00) | POMDP agent · RockSample(7,8,20,0) | number of suggestions: 15.3 -> 6.4 | Improved |
| belief-update with suggestion (scaled τ=0.99) instead of naively following (ν=1.00) | POMDP agent · RockSample(8,4,10,-1) | number of suggestions: 8.4 -> 2.8 | Improved |
| belief-update with suggestion (scaled τ=0.99) instead of naively following (ν=1.00) | POMDP agent · Tag | mean reward: -1.6 -> -1.8 | No clear effect |
| lower naive follow rate ν from 1.00 to 0.50 | Naive agent · Tag | mean reward: -1.6 -> -6.8 | Hurt |
| lower naive follow rate ν from 1.00 to 0.50 | Naive agent · Tag | number of suggestions: 3.7 -> 15.2 | Hurt |
| lower trust τ from 0.99 to 0.50 (ablation) | Scaled agent · Tag | mean reward: -1.8 -> -3.6 | Hurt |
| lower trust τ from 0.99 to 0.50 (ablation) | Scaled agent · RockSample(8,4,10,-1) | mean reward: 16.4 -> 16.3 | No clear effect |
| lower λ from 5.0 to 1.0 (ablation) | Noisy agent · Tag | mean reward: -1.8 -> -2.4 | Hurt |
| lower λ from 5.0 to 1.0 (ablation) | Noisy agent · RockSample(7,8,20,0) | number of suggestions: 7.8 -> 10.6 | Hurt |
| increase randomness of suggester from perfect to fully random | Naive agent · Tag and RockSample(8,4,10,-1) | mean reward | Hurt |
Closest papers by meaning
- Beyond Single Stationary Policies: Meta-Task Players as Naturally Superior Collaborators (NeurIPS 2024 · similarity 0.85)
- Adaptive Collaboration with Humans: Metacognitive Policy Optimization for Multi-Agent LLMs with Continual Learning (ICLR 2026 · similarity 0.84)
- Training Generalizable Collaborative Agents via Strategic Risk Aversion (NeurIPS 2026 · similarity 0.84)
- Adaptive Agents for Mixed-Initiative Human-AI Collaborations (AAAI 2025 · similarity 0.84)
- AgentCDM: Enhancing Multi-Agent Collaborative Decision-Making via ACH-Inspired Structured Reasoning (AAAI 2026 · similarity 0.83)
- Mental-Models for Multi-Agent Systems (NeurIPS 2026 · similarity 0.83)
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.
