Research map / Multi-agent systems and games / Human-AI and goal-conditioned
HandMeThat: Human-Robot Communication in Physical and Social Environments
Yanming Wan, Jiayuan Mao, Joshua B. Tenenbaum
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 oracle subgoal input (ablation) | Seq2Seq · HandMeThat v1 fully observable Level 3 | success rate (%): 4.0% -> 14.0% | Improved |
| add oracle subgoal input (ablation) | Seq2Seq · HandMeThat v1 partially observable Level 2 | success rate (%): 10.4% -> 15.7% | Improved |
| add oracle subgoal input (ablation) | Seq2Seq · HandMeThat v1 fully observable Level 1 | success rate (%): 22.4% -> 25.9% | No clear effect |
| add oracle goal input (ablation) | Seq2Seq · HandMeThat v1 fully observable Level 1 | success rate (%): 22.4% -> 17.9% | Hurt |
| add oracle goal input (ablation) | Seq2Seq · HandMeThat v1 fully observable Level 2 | success rate (%): 6.8% -> 10.4% | Improved |
| add oracle goal input (ablation) | Seq2Seq · HandMeThat v1 partially observable Level 1 | success rate (%): 25.5% -> 21.0% | Hurt |
| add oracle subgoal input (ablation) | Seq2Seq · HandMeThat v2 fully observable Level 1 | success rate (%): 30.4% -> 69.1% | Improved |
| add oracle subgoal input (ablation) | Seq2Seq · HandMeThat v2 fully observable Level 3 | success rate (%): 12.8% -> 30.4% | Improved |
| add oracle goal input (ablation) | Seq2Seq · HandMeThat v2 fully observable Level 1 | success rate (%): 30.4% -> 26.0% | Hurt |
| train DRRN offline on expert demonstrations (ablation) | DRRN · HandMeThat v1 fully observable Level 1 | success rate (%): 0.1% -> 0.0% | No clear effect |
| use v2 dataset (40x more data, fewer ambiguities) instead of v1 | Seq2Seq · HandMeThat fully observable Level 2 | success rate (%): 6.8% -> 28.8% | Improved |
| use v2 dataset instead of v1 | Heuristic · HandMeThat fully observable Level 2 | success rate (%): 28.1% -> 64.0% | Improved |
Closest papers by meaning
- Infer Human’s Intentions Before Following Natural Language Instructions (AAAI 2025 · similarity 0.88)
- Towards Better Robot Learners: Leveraging Implicit and Explicit Human Feedback Together in Human Robot Interactions (AAAI 2025 · similarity 0.87)
- PARTNR: A Benchmark for Planning and Reasoning in Embodied Multi-agent Tasks (ICLR 2025 · similarity 0.87)
- Beyond Single Stationary Policies: Meta-Task Players as Naturally Superior Collaborators (NeurIPS 2024 · similarity 0.86)
- ThinkBot: Embodied Instruction Following with Thought Chain Reasoning (ICLR 2025 · similarity 0.86)
- Generalizing Verifiable Instruction Following (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.
