Enhancing generalization in high energy physics using white-box adversarial attacks

2024-11-14

Physical Review D

AI Summary

Supervised learning models in high-energy physics suffer from poor generalization to real experimental data due to overreliance on artifacts inherent in Monte Carlo simulations. Method: We propose a white-box adversarial training framework to smooth the loss landscape, thereby improving generalization. For the first time, we systematically incorporate four classes of white-box attacks—operating in both weight and feature spaces—and combine gradient ascent-based perturbation with reduced Hessian eigenvalue analysis to quantitatively characterize and mitigate local minima sharpness. Results: Evaluated on the Higgs boson decay classification task, loss landscape smoothing yields significant gains in real-data generalization, empirically confirming its positive correlation with generalization performance. Computational overhead remains moderate. This work establishes a novel, interpretable, and quantifiable paradigm for enhancing deep learning generalization to bridge the simulation-to-reality gap.

Technology Category

• Machine Learning: Adversarial Learning & Robustness
• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Adversarial Search

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Machine learning is becoming increasingly popular in the context of particle physics. Supervised learning, which uses labeled Monte Carlo (MC) simulations, remains one of the most widely used methods for discriminating signals beyond the Standard Model. However, this paper suggests that supervised models may depend excessively on artifacts and approximations from Monte Carlo simulations, potentially limiting their ability to generalize well to real data. This study aims to enhance the generalization properties of supervised models by reducing the sharpness of local minima. It reviews the application of four distinct white-box adversarial attacks in the context of classifying Higgs boson decay signals. The attacks are divided into weight-space attacks and feature-space attacks. To study and quantify the sharpness of different local minima, this paper presents two analysis methods: gradient ascent and reduced Hessian eigenvalue analysis. The results show that white-box adversarial attacks significantly improve generalization performance, albeit with increased computational complexity. Published by the American Physical Society 2025

Problem

• Improving generalization in high energy physics models
• Reducing dependence on Monte Carlo simulation artifacts
• Enhancing supervised learning via white-box adversarial attacks

Innovation

• Uses white-box adversarial attacks for generalization
• Applies gradient ascent and Hessian analysis
• Improves model performance with computational trade-off