Learning Safe Flow Matching Robot Policies from Binary Feasibility Oracles
Overview
This research proposes an oracle-guided differentiable safeguard for safe flow-matching robot policies, using binary feasibility queries to correct unsafe actions during forward passes and extract constraint-aware signals during backward passes. The method enables forward-backward aligned policy learning without explicit constraint models, achieving 100% constraint satisfaction in obstacle-aware manipulation tasks.
From arXiv Robotics — research abstracts
Event evidence and corrections
0 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
No current evidence-backed claims. Missing information remains not reported.
Evidence dependency changes (1)
- 2026-10-31T17:55:46.983Z · evidence updated · source revision 1. Evidence extraction was updated; current source attributions are shown above.
Report timeline
Follow attributed reports and material updates.
- arXiv Robotics — research abstractsSignalLearning Safe Flow Matching Robot Policies from Binary Feasibility Oracles
This research proposes an oracle-guided differentiable safeguard for safe flow-matching robot policies, using binary feasibility queries to correct unsafe actions during forward passes and extract constraint-aware signals during backward passes. The method enables forward-backward aligned policy learning without explicit constraint models, achieving 100% constraint satisfaction in obstacle-aware manipulation tasks.
Event coverage history
There is not enough continuous observation data to show a trend.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).