Talking with Verifiers: Automatic Specification Generation for Neural Network Verification
This addresses the limitation for users who need to verify neural networks with high-level, human-understandable requirements, making formal verification more applicable to real-world domains.
The paper tackles the problem that neural network verification tools only support low-level specifications, hindering practical adoption, by introducing a framework that automatically translates natural language specifications into formal verification queries. The results show this approach successfully verifies complex semantic specifications with high fidelity and low computational overhead.
Neural network verification tools currently support only a narrow class of specifications, typically expressed as low-level constraints over raw inputs and outputs. This limitation significantly hinders their adoption and practical applicability across diverse application domains where correctness requirements are naturally expressed at a higher semantic level. This challenge is rooted in the inherent nature of deep neural networks, which learn internal representations that lack an explicit mapping to human-understandable features. To address this, we bridge this gap by introducing a novel component to the verification pipeline, making existing verification tools applicable to a broader range of domains and specification styles. Our framework enables users to formulate specifications in natural language, which are then automatically analyzed and translated into formal verification queries compatible with state-of-the-art neural network verifiers. We evaluate our approach on both structured and unstructured datasets, demonstrating that it successfully verifies complex semantic specifications that were previously inaccessible. Our results show that this translation process maintains high fidelity to user intent while incurring low computational overhead, thereby substantially extending the applicability of formal DNN verification to real-world, high-level requirements.