Future Anchored Verification and Online Recovery for World Action Models
cs.RO, cs.AI
Submitted: 2026-10-05
Updated: 2026-10-05
Comments: 18 pages, 5 figures, 3 tables
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Qwen3-VL Technical Report
- Zero-Shot Robotic Manipulation with Pretrained Image-Editing Diffusion Models
- WorldVLA: Towards Autoregressive Action World Model
- RACER: Rich Language-Guided Failure Recovery Policies for Imitation Learning
- Vision-Language Models as Success Detectors
- LIBERO-Plus: In-depth Robustness Analysis of Vision-Language-Action Models
- HarnessWAM: Bridging Prediction and Deliberation in World Action Models
- DoReMi: Grounding Language Model by Detecting and Recovering from Plan-Execution Misalignment
- SG-WAM: Text-Grounded and Spatial-aware Semantic Guidance for World-Action Models
- Classifier-Free Diffusion Guidance
- Cosmos Policy: Fine-Tuning Video Models for Visuomotor Control and Planning
- A Self-Correcting Vision-Language-Action Model for Fast and Slow System Manipulation
- Causal World Modeling for Robot Control
- BadWAM: When World-Action Models Dream Right but Act Wrong
- Vision-Language-Action Safety: Threats, Challenges, Evaluations, and Mechanisms
- Video Generators are Robot Policies
- Interactive Robot Learning from Verbal Correction
- World Action Verifier: Self-Improving World Models via Forward-Inverse Asymmetry
Related papers
- FMT x: An Efficient and Asymptotically Optimal Extension of the Fast Marching Tree for Dynamic Replanning
- MPCFormer: A physics-informed data-driven approach for explainable socially-aware autonomous driving
- RoboLab: A High-Fidelity Simulation Benchmark for Analysis of Task Generalist Policies
- ARC: A Reasoning Recipe for Robot Foundation Models
- HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments
- APT: Action Expert Pretraining Improves Instruction Generalization of Vision-Language-Action Policies