Research map / Vision-language and multimodal / Robot action and navigation
GraphCoT-VLA: A 3D Spatial-Aware Reasoning Vision-Language-Action Model for Robotic Manipulation with Ambiguous Instructions
Helong Huang, Min Cen, Kai Tan, Xingyue Quan, Guowei Huang, Hong Zhang
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 Pose-Object Graph (3D robot-object graph) (ablation) | GraphCoT-VLA · Outfit Selection (avg success) | success rate (%): 56.67 -> 70.00 | Improved |
| add Pose-Object Graph (3D robot-object graph) (ablation) | GraphCoT-VLA · Food Preparation (avg success) | success rate (%): 75.00 -> 76.67 | No clear effect |
| add structured CoT reasoning (ablation) | GraphCoT-VLA · Food Preparation (avg success) | success rate (%): 68.33 -> 76.67 | Improved |
| add structured CoT reasoning (ablation) | GraphCoT-VLA · Outfit Selection (avg success) | success rate (%): 58.33 -> 70.00 | Improved |
| add both CoT and Pose-Object Graph to pi0 (ablation) | pi0 · Food Preparation (avg success) | success rate (%): 56.67 -> 76.67 | Improved |
| add both CoT and Pose-Object Graph to pi0 (ablation) | pi0 · Outfit Selection (avg success) | success rate (%): 51.67 -> 70.00 | Improved |
| CoT co-training with dropout and CoT only on first frame | GraphCoT-VLA · Inference frequency | Hz: ≈10 Hz -> ≈10 Hz | No clear effect |
| use CoT and graph to resolve ambiguous instructions | GraphCoT-VLA vs baselines · Task confusion (Table 2) | task confusion | Improved |
Closest papers by meaning
- Localizing, Structuring, and Rendering: Bridging 3D and 2D Vision-Language-Action Models for Robotic Manipulation (CVPR 2026 · similarity 0.92)
- OmniManip: Towards General Robotic Manipulation via Object-Centric Interaction Primitives as Spatial Constraints (CVPR 2025 · similarity 0.91)
- Seeing Across Views: Benchmarking Spatial Reasoning of Vision-Language Models in Robotic Scenes (ICLR 2026 · similarity 0.91)
- CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models (CVPR 2025 · similarity 0.90)
- ConsisVLA-4D: Advancing Spatiotemporal Consistency in Efficient 3D-Perception and 4D-Reasoning for Robotic Manipulation (CVPR 2026 · similarity 0.90)
- RoboRefer: Towards Spatial Referring with Reasoning in Vision-Language Models for Robotics (NeurIPS 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.
