Research map / Vision-language and multimodal / Robot action and navigation
World-in-Loop: Online Correction via Event-Triggered World Models for Robust VLA Policies
Shaoqing Xu, Fang Li, Zhi-Xin Yang, Qimao Chen, Yuechen Luo, Zhixiang Duan, Yifan Yang, Long Chen
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 World-in-Loop online correction | CogACT · SIMPLER WidowX | average success (%): 51.1 -> 64.6 | Improved |
| add World-in-Loop online correction | MemoryVLA · SIMPLER WidowX | average success (%): 71.9 -> 78.2 | Improved |
| add World-in-Loop online correction | SpatialVLA (fine-tuned) · SIMPLER WidowX | average success (%): 42.7 -> 51.7 | Improved |
| add World-in-Loop online correction | CogACT · SIMPLER Google Robot Variant Aggregation | average success (%): 61.3 -> 67.0 | Improved |
| add World-in-Loop online correction | MemoryVLA · SIMPLER Google Robot Visual Matching | average success (%): 77.7 -> 78.6 | No clear effect |
| add World-in-Loop online correction | baseline VLA · Real-world robot | average success (%): 63.6 -> 76.1 | Improved |
| add World-in-Loop online correction | baseline VLA · Real-world robot | average grasp (%): 70.9 -> 85.4 | Improved |
| add state rollback of 15 steps before correction (ablation) | CogACT · SIMPLER WidowX rollback ablation | average success (%): 54.2 -> 64.6 | Improved |
| increase rollback depth from 15 to 25 steps (ablation) | CogACT · SIMPLER WidowX rollback ablation | average success (%): 64.6 -> 58.4 | Hurt |
| add world-model correction without rollback (ablation) | CogACT · SIMPLER WidowX rollback ablation | average success (%): 51.1 -> 54.2 | No clear effect |
| raise intervention limit from Max=2 to Max=3 (ablation) | CogACT · SIMPLER WidowX intervention limit | success rate (%): 65.6 -> 65.6 | No clear effect |
| allow unbounded interventions instead of Max=2 (ablation) | CogACT · SIMPLER WidowX intervention limit | success rate (%): 65.6 -> 66.7; latency 15.4s -> 18.1s | Mixed |
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.
- MemoryVLA — MemoryVLA: Perceptual-Cognitive Memory in Vision-Language-Action Models for Robotic Manipulation (ICLR 2026) · better on 4 of 4 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
- VLA-REPLICA: A Low-Cost, Reproducible Benchmark for Real-World Evaluation of Vision-Language-Action Models (NeurIPS 2026 · similarity 0.91)
- ElegantVLA: Learning When to Think for Efficient Vision-Language-Action Models (NeurIPS 2026 · similarity 0.90)
- ActionUNet: Improving Robustness of VLA Models with Efficient Multi-scale Fine-tuning (NeurIPS 2026 · similarity 0.90)
- VLA-Corrector: Lightweight Detect-and-Correct Inference for Adaptive Action Horizon (NeurIPS 2026 · similarity 0.90)
- Failing Forward: Adaptive Failure-Informed Learning for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.90)
- VLA-Cache: Efficient Vision-Language-Action Manipulation via Adaptive Token Caching (NeurIPS 2025 · similarity 0.90)
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.
