Formal Reasoning About Confidence and Automated Verification of Neural Networks
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866915797650112512 |
|---|---|
| author | Afzal, Mohammad Akshay, S. Genest, Blaise Gupta, Ashutosh |
| author_facet | Afzal, Mohammad Akshay, S. Genest, Blaise Gupta, Ashutosh |
| contents | In the last decade, a large body of work has emerged on robustness of neural networks, i.e., checking if the decision remains unchanged when the input is slightly perturbed. However, most of these approaches ignore the confidence of a neural network on its output. In this work, we aim to develop a generalized framework for formally reasoning about the confidence along with robustness in neural networks. We propose a simple yet expressive grammar that captures various confidence-based specifications. We develop a novel and unified technique to verify all instances of the grammar in a homogeneous way, viz., by adding a few additional layers to the neural network, which enables the use any state-of-the-art neural network verification tool. We perform an extensive experimental evaluation over a large suite of 8870 benchmarks, where the largest network has 138M parameters, and show that this outperforms ad-hoc encoding approaches by a significant margin. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_07293 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Formal Reasoning About Confidence and Automated Verification of Neural Networks Afzal, Mohammad Akshay, S. Genest, Blaise Gupta, Ashutosh Logic in Computer Science Artificial Intelligence Computer Vision and Pattern Recognition In the last decade, a large body of work has emerged on robustness of neural networks, i.e., checking if the decision remains unchanged when the input is slightly perturbed. However, most of these approaches ignore the confidence of a neural network on its output. In this work, we aim to develop a generalized framework for formally reasoning about the confidence along with robustness in neural networks. We propose a simple yet expressive grammar that captures various confidence-based specifications. We develop a novel and unified technique to verify all instances of the grammar in a homogeneous way, viz., by adding a few additional layers to the neural network, which enables the use any state-of-the-art neural network verification tool. We perform an extensive experimental evaluation over a large suite of 8870 benchmarks, where the largest network has 138M parameters, and show that this outperforms ad-hoc encoding approaches by a significant margin. |
| title | Formal Reasoning About Confidence and Automated Verification of Neural Networks |
| topic | Logic in Computer Science Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2511.07293 |