Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Where2Act: From Pixels to Actions for Articulated 3D Objects
Kaichun Mo, Leonidas J. Guibas, Mustafa Mukadam, Abhinav Gupta, Shubham Tulsiani
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 online adaptive data sampling (ablation) | 3D-ours · SAPIEN pushing (seen categories) | F-score (%): 40.54 -> 43.76 | Improved |
| add online adaptive data sampling (ablation) | 3D-ours · SAPIEN pushing (seen categories) | sample-success rate (%): 25.18 -> 28.54 | Improved |
| add online adaptive data sampling (ablation) | 3D-ours · SAPIEN pushing (unseen categories) | sample-success rate (%): 11.76 -> 14.74 | Improved |
| add online adaptive data sampling (ablation) | 3D-ours · SAPIEN pushing-up | sample-success rate (%): 12.88 -> 12.06 (seen), 6.43 -> 6.56 (unseen) | Mixed |
| add online adaptive data sampling (ablation) | 3D-ours · SAPIEN pushing-up | F-score (%): 21.03 -> 21.64 (seen), 11.57 -> 11.20 (unseen) | Mixed |
| add online adaptive data sampling (ablation) | 3D-ours · SAPIEN pushing-left | F-score (%): 24.71 -> 26.04 | Improved |
| add online adaptive data sampling (ablation) | 3D-ours · SAPIEN pulling | sample-success rate (%): 5.62 -> 7.51 (seen), 5.87 -> 7.85 (unseen) | Improved |
| add online adaptive data sampling (ablation) | 3D-ours · SAPIEN pulling-up | F-score (%): 20.51 -> 22.24 (seen), 13.70 -> 16.28 (unseen) | Improved |
| add online adaptive data sampling (ablation) | 3D-ours · SAPIEN pulling-left | F-score (%): 23.41 -> 25.22 (seen), 15.07 -> 14.49 (unseen) | Mixed |
| use learned features instead of ground-truth normals (B-Normal) (ablation) | 3D-ours · SAPIEN pushing (seen categories) | F-score (%): 31.94 -> 43.76 | Improved |
| use learned features instead of PCPNet normals and curvatures (B-PCPNet) (ablation) | 3D-ours · SAPIEN pushing (seen categories) | F-score (%): 32.01 -> 43.76 | Improved |
| use 3D point cloud input instead of 2D RGB image (ablation) | Where2Act · SAPIEN pushing (seen categories) | F-score (%): 34.21 -> 43.76 | Improved |
Closest papers by meaning
- VAT-Mart: Learning Visual Action Trajectory Proposals for Manipulating 3D ARTiculated Objects (ICLR 2022 · similarity 0.90)
- Unsupervised Learning for Physical Interaction through Video Prediction (NeurIPS 2016 · similarity 0.89)
- IAAO: Interactive Affordance Learning for Articulated Objects in 3D Environments (CVPR 2025 · similarity 0.89)
- Act the Part: Learning Interaction Strategies for Articulated Object Part Discovery (ICCV 2021 · similarity 0.89)
- Articulate3D: Holistic Understanding of 3D Scenes as Universal Scene Description (ICCV 2025 · similarity 0.88)
- Articulate-Anything: Automatic Modeling of Articulated Objects via a Vision-Language Foundation Model (ICLR 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.
