analiza aiqbench
MiniFool introduces a physics-constrained adversarial attack framework designed to rigorously test the robustness of deep neural networks in high-energy particle and astroparticle physics. By integrating experimental uncertainties directly into the cost function, this method allows researchers to quantify classification stability under realistic data perturbations beyond standard MNIST benchmarks. The approach offers a critical tool for validating AI models used in major experiments like IceCube and CMS before deployment on unlabeled scientific datasets.
MiniFool -- Physics-Constraint-Aware Minimizer-Based Adversarial Attacks in Deep Neural Networks
arXiv:2511.01352v2 Announce Type: replace Abstract: In this paper, we present a new algorithm, MiniFool, that implements physics-inspired adversarial attacks for testing neural network-based classification tasks in particle and astroparticle physics. While we initially developed the algorithm for the search for astrophysical tau neutrinos with the IceCube Neutrino Observatory, we apply it to further data from other science domains, thus demonstrating its general applicability. Here, we apply the algorithm to the well-known MNIST data set and furthermore, to Open Data data from the CMS experiment at the Large Hadron Collider. The algorithm is based on minimizing a cost function that combines a $\chi^2$ based test-statistic with the deviation from the desired target score. The test statistic quantifies the probability of the perturbations applied to the data based on the experimental uncertainties. For our studied use cases, we find that the likelihood of a flipped classification differs for both the initially correctly and incorrectly classified events. When testing changes of the classifications as a function of an attack parameter that scales the experimental uncertainties, the robustness of the network decision can be quantified. Furthermore, this allows testing the robustness of the classification of unlabeled experimental data.
Abstract:In this paper, we present a new algorithm, MiniFool, that implements physics-inspired adversarial attacks for testing neural network-based classification tasks in particle and astroparticle physics. While we initially developed the algorithm for the search for astrophysical tau neutrinos with the IceCube Neutrino Observatory, we apply it to further data from other science domains, thus demonstrating its general applicability. Here, we apply the algorithm to the well-known MNIST data set and furthermore, to Open Data data from the CMS experiment at the Large Hadron Collider. The algorithm is based on minimizing a cost function that combines a $\chi^2$ based test-statistic with the deviation from the desired target score. The test statistic quantifies the probability of the perturbations applied to the data based on the experimental uncertainties. For our studied use cases, we find that the likelihood of a flipped classification differs for both the initially correctly and incorrectly classified events. When testing changes of the classifications as a function of an attack parameter that scales the experimental uncertainties, the robustness of the network decision can be quantified. Furthermore, this allows testing the robustness of the classification of unlabeled experimental data.
Machine Learning (cs.LG); High Energy Astrophysical Phenomena (astro-ph.HE); Instrumentation and Methods for Astrophysics (astro-ph.IM); High Energy Physics - Experiment (hep-ex); Data Analysis, Statistics and Probability (physics.data-an)
arXiv:2511.01352 [cs.LG]
(or arXiv:2511.01352v2 [cs.LG] for this version)
Published in: Eur.Phys.J.C 86 (2026) 6, 641
https://doi.org/10.1140/epjc/s10052-026-15773-2 DOI(s) linking to related resources
Text integral sub licența deschisă a publicației · 236 cuvinte