mlsanityData to 14 Oct 2026

Learning Lookahead Lemmas for Neural Network Verification

arXiv 2607.29051 ↗First posted 31 Jul 2026cs.LG cs.AI cs.LOcited by 0

Not on the map. Why?

Abstract (arXiv)

State-of-the-art neural network verifiers use the branch-and-bound procedure as their core solving mechanism. We introduce an inprocessing framework for neural network verification driven by the lookahead procedure. Under this framework, lookahead derives new lemmas over the phases of unstable ReLUs, which are collected into an implication graph that is used to prune the search space and vivify boolean cuts. We instantiate the framework in two state-of-the-art verifiers, Marabou and α-β-CROWN, and demonstrate that it improves performance in both, proving up to 34% more instances unsatisfiable.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 18 of its 35 references that are in the map. First posted on arXiv 31 Jul 2026.

Builds on

18 of 35 references.

References · 18

  1. 2015Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classificationcited by 3,611
  2. 2013Playing Atari with Deep Reinforcement Learningcited by 3,274
  3. 2017Reluplex: An Efficient SMT Solver for Verifying Deep Neural Networkscited by 744
  4. 2018Efficient Neural Network Robustness Certification with General Activation Functionscited by 445
  5. 2017Deep Reinforcement Learning framework for Autonomous Drivingcited by 225
  6. 2020Fast and Complete: Enabling Complete Neural Network Verification with Rapid and Massively Parallel Incomplete Verifierscited by 187
  7. 2019Branch and Bound for Piecewise Linear Neural Network Verificationcited by 138
  8. 2022General Cutting Planes for Bound-Propagation-Based Neural Network Verificationcited by 103
  9. 2024Neural Network Verification with Branch-and-Bound for General Nonlinearitiescited by 47
  10. 2024The Fifth International Verification of Neural Networks Competition (VNN-COMP 2024): Summary and Resultscited by 43
  11. 2023The Fourth International Verification of Neural Networks Competition (VNN-COMP 2023): Summary and Resultscited by 38
  12. 2024Scalable Neural Network Verification with Branch-and-bound Inferred Cutting Planescited by 33
6 more
  1. 2022Efficient Neural Network Analysis with Sum-of-Infeasibilitiescited by 24
  2. 2025Clip-and-Verify: Linear Constraint-Driven Domain Clipping for Accelerating Neural Network Verificationcited by 9
  3. 2018Clause Vivification by Unit Propagation in CDCL SAT Solverscited by 6
  4. 2025PICID: Proof-Driven Clause Learning in Neural Network Verificationcited by 3
  5. 2026Incremental Neural Network Verification via Learned Conflictscited by 2
  6. 2026Lookahead Branching for Neural Network Verificationcited by 1

Built on by

Not cited yet.