Research map / Vision-language and multimodal / Robot action and navigation
VULCAN: Tool-Augmented Multi Agents for Iterative 3D Object Arrangement
Zhengfei Kuang, Rui Lin, Long Zhao, Gordon Wetzstein, Saining Xie, Sanghyun Woo
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 constraint solver and visual API tool library (ablation) | VULCAN · VULCAN benchmark (25 scenes, 111 unit tasks) | plausibility: 3.484 -> 3.796 | Improved |
| add constraint solver and visual API tool library (ablation) | VULCAN · VULCAN benchmark (25 scenes, 111 unit tasks) | collision rate: 0.495 -> 0.000 | Improved |
| add adaptive backtracking (ablation) | VULCAN · VULCAN benchmark (25 scenes, 111 unit tasks) | collision rate: 0.036 -> 0.000 | Improved |
| add adaptive backtracking (ablation) | VULCAN · VULCAN benchmark (25 scenes, 111 unit tasks) | floating rate: 0.054 -> 0.000 | Improved |
| add MCP-based tools instead of raw Blender scripts (ablation) | VULCAN · VULCAN benchmark (25 scenes, 111 unit tasks) | collision rate: 0.603 -> 0.000 | Improved |
| add planner target coordinates (ablation) | VULCAN · VULCAN benchmark (25 scenes, 111 unit tasks) | consistency: 3.526 -> 3.592 | Improved |
| add planner target coordinates (ablation) | VULCAN · VULCAN benchmark (25 scenes, 111 unit tasks) | collision rate: 0.000 -> 0.000 | No clear effect |
| use multi-agent planner/executor/evaluator instead of single agent (ablation) | VULCAN · VULCAN benchmark (25 scenes, 111 unit tasks) | consistency: 3.328 -> 3.592 | Improved |
| use multi-agent instead of single agent (ablation) | VULCAN · VULCAN benchmark (25 scenes, 111 unit tasks) | collision rate: 0.000 -> 0.000 | No clear effect |
| VULCAN full system vs best single-step baseline FirePlace | VULCAN · VULCAN benchmark (25 scenes, 111 unit tasks) | consistency: 3.135 -> 3.592 | Improved |
Closest papers by meaning
- Localizing, Structuring, and Rendering: Bridging 3D and 2D Vision-Language-Action Models for Robotic Manipulation (CVPR 2026 · similarity 0.88)
- ManipLLM: Embodied Multimodal Large Language Model for Object-Centric Robotic Manipulation (CVPR 2024 · similarity 0.88)
- OmniManip: Towards General Robotic Manipulation via Object-Centric Interaction Primitives as Spatial Constraints (CVPR 2025 · similarity 0.88)
- D3D-VLP: Dynamic 3D Vision-Language-Planning Model for Embodied Grounding and Navigation (CVPR 2026 · similarity 0.88)
- AffordBot: 3D Fine-grained Embodied Reasoning via Multimodal Large Language Models (NeurIPS 2025 · similarity 0.88)
- 3DCodeBench: Benchmarking Agentic Procedural 3D Modeling Via Code (NeurIPS 2026 · similarity 0.88)
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.
