Research map / Vision-language and multimodal / Robot action and navigation
ZAP: Zero-Shot Assembly Planning with Large Language Models
Linpeng Peng, Yanbo WANG, Chuanjie Lv, Wencan Jiang, Liming Xu, Jianbiao Mei, Xinyue Yao, Yong 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 Iterative Decision Aggregator (propose-and-verify with backtracking) (ablation) | ZAP · ASAP-G (all 150) | PSR (%): 74.00 -> 88.67 | Improved |
| add IDA (ablation) | ZAP · ASAP-G (Hard) | PSR (%): 63.33 -> 86.67 | Improved |
| add IDA (ablation) | ZAP · ASAP-G (Simple) | PSR (%): 81.42 -> 90.00 | Improved |
| add raw image input alongside structured dossiers (ablation) | ZAP · ASAP-G (all 150) | PSR (%): 86.00 -> 88.67 | Improved |
| add raw image input (ablation) | ZAP · ASAP-G (Hard) | PSR (%): 80.00 -> 86.67 | Improved |
| add structured part dossiers (ablation) | ZAP · ASAP-G (all 150) | PSR (%): 86.67 -> 88.67 | Improved |
| add structured part dossiers (ablation) | ZAP · ASAP-G (Medium) | PSR (%): 84.00 -> 88.00 | Improved |
| add VLM semantic pruning over random permutation with verification | Random Permutation · ASAP-G (all 150) | PSR (%): 84.67 -> 88.67 | Improved |
| ZAP versus ASAP on Hard tier | ASAP (Heur./Learn.) · ASAP-G (Hard) | PSR (%): 76.67 -> 86.67 | Improved |
| ZAP versus ASAP on Toy category | ASAP (Heur./Learn.) · ASAP-G (Toy) | PSR (%): 76.19 -> 95.24 | Improved |
| ZAP versus ASAP on Simple tier | ASAP (Heur./Learn.) · ASAP-G (Simple) | PSR (%): 90.00 -> 90.00 | No clear effect |
| ZAP versus ASAP on Furniture category | ASAP (Heur./Learn.) · ASAP-G (Furniture) | PSR (%): 87.14 -> 84.29 | Hurt |
Closest papers by meaning
- Neural Assembler: Learning to Generate Fine-Grained Robotic Assembly Instructions from Multi-View Images (AAAI 2025 · similarity 0.86)
- UniDomain: Pretraining a Unified PDDL Domain from Real-World Demonstrations for Generalizable Robot Task Planning (NeurIPS 2025 · similarity 0.85)
- Robotic Visual Instruction (CVPR 2025 · similarity 0.85)
- Imagine: Image-Guided 3D Part Assembly with Structure Knowledge Graph (AAAI 2025 · similarity 0.84)
- RoboCerebra: A Large-scale Benchmark for Long-horizon Robotic Manipulation Evaluation (NeurIPS 2025 · similarity 0.84)
- GENMANIP: LLM-driven Simulation for Generalizable Instruction-Following Manipulation (CVPR 2025 · similarity 0.84)
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.
