Research map / Reinforcement learning / Visual and representation RL
Human-Object Interaction from Human-Level Instructions
Zhen Wu, Jiaman Li, Pei Xu, C. Karen Liu
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 RefineNet after CoarseNet (ablation) | CoarseNet · FullBodyManipulation | intersection volume IV: 48.56 -> 24.78 | Improved |
| add RefineNet after CoarseNet (ablation) | CoarseNet · FullBodyManipulation | contact recall: 0.81 -> 0.95 | Improved |
| add RefineNet after CoarseNet (ablation) | CoarseNet · FullBodyManipulation | contact F1: 0.84 -> 0.92 | Improved |
| add RefineNet after CoarseNet (ablation) | CoarseNet · FullBodyManipulation | waypoint error T_xy: 3.26 -> 2.81 | Improved |
| add RefineNet after CoarseNet (ablation) | CoarseNet · FullBodyManipulation | MPJPE: 14.32 -> 15.96 | Hurt |
| add RefineNet after CoarseNet (ablation) | CoarseNet · FullBodyManipulation | object position error T_obj: 10.56 -> 12.57 | Hurt |
| add FingerNet after RefineNet (ablation) | C+RNet · FullBodyManipulation | contact coverage CC: 3.84% -> 5.53% | Improved |
| add FingerNet after RefineNet (ablation) | C+RNet · FullBodyManipulation | intersection volume IV: 24.78 -> 19.06 | Improved |
| add FingerNet after RefineNet (ablation) | C+RNet · FullBodyManipulation | contact F1: 0.92 -> 0.92 | No clear effect |
| replace CNet+GRIP with the full three-stage model | CNet+GRIP · FullBodyManipulation | contact F1: 0.70 -> 0.92 | Improved |
| replace CNet+GRIP with the full three-stage model | CNet+GRIP · FullBodyManipulation | contact recall: 0.64 -> 0.95 | Improved |
| replace direct LLM position prediction with scene-graph relations | LLM planner (GPT-4o) · Scene map, 25 instructions | positional error rate PE_p: 21.9% -> 3.1% | Improved |
Closest papers by meaning
- PAMotion: Physics-Aware Motion Generation for Full-Body Interaction with Multiple Objects (CVPR 2026 · similarity 0.90)
- InterMimic: Towards Universal Whole-Body Control for Physics-Based Human-Object Interactions (CVPR 2025 · similarity 0.89)
- Instruction-Augmented Long-Horizon Planning: Embedding Grounding Mechanisms in Embodied Mobile Manipulation (AAAI 2025 · similarity 0.88)
- CG-HOI: Contact-Guided 3D Human-Object Interaction Generation (CVPR 2024 · similarity 0.87)
- OpenHOI: Open-World Hand-Object Interaction Synthesis with Multimodal Large Language Model (NeurIPS 2025 · similarity 0.87)
- TokenHSI: Unified Synthesis of Physical Human-Scene Interactions through Task Tokenization (CVPR 2025 · similarity 0.87)
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.
