[2603.25266] Probabilistic Abstract Interpretation on Neural Networks via Grids Approximation

Computer Science > Artificial Intelligence arXiv:2603.25266 (cs) [Submitted on 26 Mar 2026] Title:Probabilistic Abstract Interpretation on Neural Networks via Grids Approximation Authors:Zhuofan Zhang, Herbert Wiklicky View a PDF of the paper titled Probabilistic Abstract Interpretation on Neural Networks via Grids Approximation, by Zhuofan Zhang and Herbert Wiklicky View PDF HTML (experimental) Abstract:Probabilistic abstract interpretation is a theory used to extract particular properties of a computer program when it is infeasible to test every single inputs. In this paper we apply the theory on neural networks for the same purpose: to analyse density distribution flow of all possible inputs of a neural network when a network has uncountably many or countable but infinitely many inputs. We show how this theoretical framework works in neural networks and then discuss different abstract domains and corresponding Moore-Penrose pseudo-inverses together with abstract transformers used in the framework. We also present experimental examples to show how this framework helps to analyse real world problems. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2603.25266 [cs.AI]   (or arXiv:2603.25266v1 [cs.AI] for this version)   https://doi.org/10.48550/arXiv.2603.25266 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Zhuofan Zhang [view email] [v1] Thu, 26 Mar 2026 10:04:29 UTC (238 KB) Full-text links: Access Paper: View a ...