Giacobbe / Lukina AI Verification

Second International Symposium, SAIV 2025, Zagreb, Croatia, July 21–22, 2025, Proceedings
Erscheinungsjahr 2025
ISBN: 978-3-031-99991-8
Verlag: Springer International Publishing
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, 280 Seiten

Reihe: Computer Science

This LNCS volume constitutes the proceedings of the Second International Symposium, SAIV 2025, in Zagreb, Croatia, during July 2025.

The scope of the topics was broadly categorized into two groups. The first group, formal methods for artificial intelligence, comprised: formal specifications for systems with AI components; formal methods for analyzing systems with AI components; formal synthesis methods of AI components; testing approaches for systems with AI components; statistical approaches for analyzing systems with AI components; and approaches for enhancing the explainability of systems with AI components. The second group, artificial intelligence for formal methods, comprised: AI methods for formal verification; AI methods for formal
synthesis; AI methods for safe control; and AI methods for falsification.

Zielgruppe

Research

Autoren/Hrsg.

Giacobbe, Mirco

Lukina, Anna

Inhaltsverzeichnis

._Technical Program.
._Scenario-based Compositional Verification of Autonomous Systems.
._Robustness Margin: A new measure for the robustness of neural networks.
._GRENA: GPU-aided Abstract Refinement for Neural Network Verification.
._ClassInvGen: Class Invariant Synthesis using Large Language Models.
._Bridging Neural ODE and ResNet: A Formal Error Bound for Safety Verification.
._Probabilistic verification of neural networks with sampling-based Probability Box propagation.
._How to Verify Generalization Capability of a Neural Network with Formal Methods.
._Certified Error Analysis of Homomorphically Encrypted Neural Networks.
._Neural Network Verification for Gliding Drone Control: A Case Study.
._Extended Abstracts.
._Abstraction-Based Proof Production in Formal Verification of Neural  Networks.
._On the Complexity of Formal Reasoning in State Space Models.
._Quantifiers for Di’erentiable Logics in Rocq.
._CTRAIN - A Training Library for Certifiably Robust Neural Networks.
._Competition Contributions.
._NeuralSAT: Scaling Constraint Solving for DNN Verification.
._NNV: a Star Set Reachability Approach.
._PyRAT: Verifying Neural Networks with Abstract Interpretation.
._SobolBox: Boxed Refinement of Sobol Sequence Samples for Neural Network Verification.