Research map / 3D vision / 3D generation
Adaptive Articulated Object Manipulation On The Fly with Foundation Model Reasoning and Part Grounding
Xiaojie Zhang, Yuanfei Wang, Ruihai Wu, Kunqi Xu, Yu Li, Liuyu Xiang, Hao Dong, Zhaofeng He
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 GPT-4o detailed part description as GroundingDINO prompt (ablation) | AdaRPG · AdaManip sim (7 categories) | manipulation success rate: Pen 0.00 -> 0.73; Bottle 0.11 -> 0.84; Pressure Cooker 0.00 -> 1.00 | Improved |
| add part affordance model (vs uniform part-center contact point) (ablation) | AdaRPG · AdaManip sim (7 categories) | manipulation success rate: Pressure Cooker 0.83 -> 1.00; Lamp 0.40 -> 0.70; Bottle 0.78 -> 0.84 | Improved |
| use part-level affordance input instead of whole-object input | affordance model · AdaManip categories | F1 score: Pen 0.18 -> 0.66; Door 0.17 -> 0.61; Lamp 0.16 -> 0.64 | Improved |
| compare AdaRPG with SAGE, zero-shot | AdaRPG · AdaManip sim | manipulation success rate: Pressure Cooker 0.00 -> 1.00; Bottle 0.21 -> 0.84 | Improved |
| compare AdaRPG with CoPa, zero-shot | AdaRPG · AdaManip sim | manipulation success rate: Window 0.08 -> 0.84; Pressure Cooker 0.17 -> 1.00 | Improved |
| compare AdaRPG with AdaManip trained on all 7 categories | AdaRPG · AdaManip sim | manipulation success rate: Pressure Cooker 0.50 -> 1.00; Lamp 0.30 -> 0.70 | Improved |
| compare AdaRPG with AdaManip in real world | AdaRPG · Real-world pressure cooker | success rate (of 10): 5/10 -> 10/10 | Improved |
| compare AdaRPG with CoPa in real world | AdaRPG · Real-world lamp | success rate (of 10): 1/10 -> 8/10 | Improved |
| compare AdaRPG with SAGE in real world | AdaRPG · Real-world microwave | success rate (of 10): 3/10 -> 9/10 | Improved |
| apply to real world rather than simulation | AdaRPG · Pressure cooker success | success rate: 1.00 -> 10/10 | 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.
- VAT-MART — VAT-Mart: Learning Visual Action Trajectory Proposals for Manipulating 3D ARTiculated Objects (ICLR 2022) · better on 2 of 2 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
- IAAO: Interactive Affordance Learning for Articulated Objects in 3D Environments (CVPR 2025 · similarity 0.91)
- AffordDP: Generalizable Diffusion Policy with Transferable Affordance (CVPR 2025 · similarity 0.91)
- AdaManip: Adaptive Articulated Object Manipulation Environments and Policy Learning (ICLR 2025 · similarity 0.90)
- GMAP: Generalized Manipulation of Articulated Objects in Robotic Using Pre-trained Model (AAAI 2025 · similarity 0.89)
- OmniManip: Towards General Robotic Manipulation via Object-Centric Interaction Primitives as Spatial Constraints (CVPR 2025 · similarity 0.89)
- BiAssemble: Learning Collaborative Affordance for Bimanual Geometric Assembly (ICML 2025 · similarity 0.89)
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.
