SOURCE-LINKED INTELLIGENCE

Where Success Breaks: Failure-Boundary Learning for Robust Vision-Language-Action Models

arXiv · AI, language, vision and robotics · article · Sep 5, 2026 · UTC

Vision-language-action (VLA) models adapted through supervised fine-tuning (SFT) inherit a structural asymmetry: expert demonstrations teach the policy where success behavior lies, but provide no signal about where it ceases to be reliable. We argue that robust VLA adaptation should therefore be viewed not as further demonstration fitting, but as **Failure-Boundary Learning**---the problem of *Discovering*, *Localizing*, and *Shaping* the boundary between recoverable deviations and task failure. To instantiate this view, we propose **DLS**: built on a **real-grounded behavioral prior** from fe

Read original source ↗ Open in workspace

recordType

paper

region

Global