Searcharxiv⌕ Search

arXiv · 2611.20781

Verification-Guided Safety-Repair of Learned Controllers

Abstract

Formal verification can certify that a learned controller is safe. But what should be done when verification fails? Simply suppressing actions that lead to unsafe next states is insufficient because the dangerous action that ultimately caused the safety violation might have occurred many steps earlier. Retraining modifies the policy globally, thereby potentially introducing new failures. Instead, we introduce ReachECO, a verification-guided repair procedure that exploits the information of failed multi-step reachability analysis for synthesizing explicit constraints on controller behavior. For Differentiable Weightless Controllers, the repairs compile into exact circuit patches that leave the controller unchanged everywhere outside the targeted policy cells. We prove soundness and termination of the repair procedure for ReachECO, as well as, assuming exhaustive continuation search, completeness relative to the reachability abstraction. Empirically, we demonstrate \method's ability to successfully repair otherwise unsafe controllers in three continuous-control benchmarks: ACC, Docking2D and Car2D.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Fabian Kresse, Sobhan Gharanfoli, Nikolaus Vertovec, Alessandro Abate, Thomas A. Henzinger, Christoph H. Lampert. 2026-11-16. Verification-Guided Safety-Repair of Learned Controllers. https://searcharxiv.com/abs/2611.20781

Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Searcharxiv by Indiciae: BibTeX · RIS.

KEEP EXPLORING

Related papers

Generalizing Unit Commitment Problem Solving via SAT-based Decoupling

As the cornerstone of modern power systems, the Unit Commitment Problem (UC) is critical for ensuring operational security and economic efficiency in the ongoing global energy transition. However, existing UC studies typically propose specialized algorithms for specific variants and operational requirements, tightly coupling the algorithms to their target models and limiting their applicability to other variants. To address this issue, this paper proposes a method that uses SAT-based reduction to decouple the algorithm from the problem, which allows a single algorithm to solve multiple UC variants. By uniformly reducing all UC variants to SAT instances solvable by standard SAT solvers, this method makes the solving algorithm independent of the original UC variant, thus granting it broad applicability across diverse variants. Experimental results show that our method achieves better solution quality than specialized algorithms and demonstrates stronger generalizability. This work offers a fast and flexible framework for addressing newly emerging UC formulations in evolving power systems.

cs.LO↗

DAPS: Dependency-Aware Premise Selection for LLM Theorem Proving

LLM-based theorem proving in Lean 4 has advanced rapidly, but provers are difficult to leverage newly contributed lemmas, because retrieval skews toward popular foundational lemmas, leaving the right premises buried among hundreds of thousands of candidates. Existing neural selectors rank candidates by semantic similarity, with three intrinsic limitations: frequency-skewed retrieval, isolated pairwise ranking, and degradation on out-of-distribution queries such as competition problems and informal language. Given that mathematics is structured by dependencies rather than similarities, we restore this signal to both the training data and the model. We curate two underused dependency resources. We conduct the first systematic mining of 53 Lean Blueprint projects, a previously untapped corpus of 3,805 manually curated informal+formal nodes that capture rare, research-level premises absent from Mathlib4. We further pair it with the extraction of comprehensive, typed, multi-level Mathlib4 dependencies. We then introduce DAPS, a dependency-aware premise selector with a structural neighborhood encoder, a group-level contrastive objective, and a Mathlib-then-Blueprint adaptation procedure. DAPS reaches 78.74 Recall@16 on Mathlib4-Heldout (+11.41 over the strong selector LeanHammer) and widens the margin to 7xrecall on competition-style queries. Downstream, its selected premises raise Goedel-Prover-V2-8B Pass@4 on miniF2F-test from 49.90 to 54.03 and improve all four general-purpose LLMs on the informal IMO-ProofBench (up to +11.4% relative). Project page: https://daps-eth.github.io/.

cs.LO↗

Path Abstraction for Markov Reward Models: An Isabelle Formalisation in a Free Monoid View

The classical definition of discrete-time Markov chains maps pairs of states to transition probabilities. An alternative, proposed in the context of path abstraction in probabilistic counterexample refinement, is to view them as a transition kernel on the free monoid on the state space, mapping paths directly to path probabilities. In this paper, we first extend the free monoid view and path abstraction algorithm to cumulative rewards on Markov reward models. We then use Isabelle/HOL to formalise Markov chains and reward models in this view, exploiting the discrete finite-prefix semantics to work with Isabelle's standard HOL-Analysis library. We mechanise the path abstraction algorithm and its key correctness properties of well-definedness and monotonic absorption, establishing the preservation of reachability probabilities and expected rewards. Proving the convergence of these quantities on relevant path sets is aided by our use of the lightweight free monoid semantics in place of the standard measure-theoretic one used in established Isabelle formalisations of Markov models. We obtain a verified abstract implementation of path abstraction in Isabelle/HOL, providing a foundation for refinement to efficient executable code and integration with probabilistic model checkers.

cs.LO↗