Research map / Multi-agent systems and games / Human-AI and goal-conditioned
TeamHOI: Learning a Unified Policy for Cooperative Human-Object Interactions with Any Team Size
Stefan Lionar, Gim Hee Lee
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 |
|---|---|---|---|
| single unified transformer policy with teammate tokens trained on 2-8 agents instead of team-size-specific policy | CooHOI* · Cooperative carrying, 8 agents | success rate (%): 61.5 -> 97.5 | Improved |
| unified policy vs best matched team-size baseline | CooHOI*-2 · Cooperative carrying, 2 agents | success rate (%): 97.5 -> 99.1 | Improved |
| unified policy with learned formation vs pre-defined oracle contact points | CooHOI*-4 · Cooperative carrying, 4 agents | success rate (%): 94.5 -> 99.2 | Improved |
| unified policy under 5x heavy load, 8 agents | CooHOI*-4 · Cooperative carrying 5x load, 8 agents | success rate (%): 14.2 -> 81.1 | Improved |
| TeamHOI vs CooHOI*-8 on motion smoothness at 8 agents | CooHOI*-8 · Cooperative carrying, 8 agents | mean absolute jerk (m/s3): 45.2 -> 34.2 | Improved |
| TeamHOI vs CooHOI*-4 on 4-agent cooperative time ratio | CooHOI*-4 · Cooperative carrying, 4 agents | cooperative time ratio (%): 92.1 -> 96.1 | Improved |
| add masked AMP (ablation) | TeamHOI · Walking+formation+lifting | task reward | Improved |
| add principal-axes coverage reward to angular spread reward (ablation) | TeamHOI · Cooperative carrying | formation quality | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CooHOI*-2 — CooHOI: Learning Cooperative Human-Object Interaction with Manipulated Object Dynamics (NeurIPS 2024) · better on 29 of 30 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- CooHOI: Learning Cooperative Human-Object Interaction with Manipulated Object Dynamics (NeurIPS 2024 · similarity 0.90)
- Humanoid Horizon: Extending Task Horizon in Whole-Body Loco-Manipulation via Parallel Training, Dynamic Starting, and Reward Gating (NeurIPS 2026 · similarity 0.86)
- NestRL: A Nested Training Regime for Mutual Adaptation in Human–AI Teaming (NeurIPS 2026 · similarity 0.86)
- Beyond Single Stationary Policies: Meta-Task Players as Naturally Superior Collaborators (NeurIPS 2024 · similarity 0.85)
- Designs for Enabling Collaboration in Human-Machine Teaming via Interactive and Explainable Systems (NeurIPS 2024 · similarity 0.85)
- Habitat 3.0: A Co-Habitat for Humans, Avatars, and Robots (ICLR 2024 · similarity 0.85)
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.
