Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Partner-Aware Algorithms in Decentralized Cooperative Bandit Teams
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 partner-aware action prediction instead of Naïve UCB | Human-robot bandit team · Burger stacking | cumulative regret: 3.6±0.31 -> 2.7±0.31 | Helped |
| Use partner-aware action prediction instead of Naïve UCB | Human-robot bandit team · Burger stacking | satisfied customers: 12.7±0.6 -> 13.5±0.6 | Helped |
| Use partner-aware action prediction instead of Naïve UCB | Human-AI casino team · Online casino study, 25 casinos | coins earned: 17,717.2±235.0 -> 18,825.8±39.3 | Helped |
| Use partner-aware action prediction rather than naïve decentralized UCB or Thompson sampling | Decentralized bandit team · Fixed and random reward-mean simulations | cumulative regret | Helped |
| Set the leader action repetition parameter to L=1 rather than larger L values (ablation) | Partner-Aware UCB · Fixed reward-mean simulation | cumulative regret | No clear effect |
| Increase the follower's action-history window from W=1 (ablation) | Partner-Aware UCB with a UCB leader · Fixed reward-mean simulation | cumulative regret | Hurt |
| Use W=25 rather than other tested window widths (ablation) | Partner-Aware UCB follower with a knowledge-gradient leader · Random reward-mean simulation | cumulative regret | Helped |
| Use a knowledge-gradient leader instead of a UCB leader (ablation) | Partner-Aware UCB follower team · Random reward-mean simulation | cumulative regret | Hurt |
| Extend the partner-aware hierarchy to more agents | Partner-Aware UCB · Simulations with 2, 3, and 4 agents | cumulative regret | Helped |
| Apply Partner-Aware UCB rather than Naïve UCB | Human-AI casino team · Online casino study, participant-level comparison | cumulative regret | Helped |
| Apply Partner-Aware UCB rather than Naïve UCB | Human-AI casino team · Online casino study, casino-level comparison | cumulative regret | Mixed |
Closest papers by meaning
- Collaborative Multi-Agent Heterogeneous Multi-Armed Bandits (ICML 2023 · similarity 0.91)
- Cooperative Stochastic Bandits with Asynchronous Agents and Constrained Feedback (NeurIPS 2021 · similarity 0.90)
- Adaptive Sample Sharing for Multi Agent Linear Bandits (ICML 2025 · similarity 0.89)
- Training Generalizable Collaborative Agents via Strategic Risk Aversion (NeurIPS 2026 · similarity 0.89)
- Distributed Multi-Agent Bandits Over Erdős-Rényi Random Networks (NeurIPS 2025 · similarity 0.88)
- Individual Regret in Cooperative Stochastic Multi-Armed Bandits (NeurIPS 2025 · similarity 0.86)
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.