Saved in:
Bibliographic Details
Main Author: Stein, Annika
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2303.14511
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913127579254784
author Stein, Annika
author_facet Stein, Annika
contents In the field of high-energy physics, deep learning algorithms continue to gain in relevance and provide performance improvements over traditional methods, for example when identifying rare signals or finding complex patterns. From an analyst's perspective, obtaining highest possible performance is desirable, but recently, some attention has been shifted towards studying robustness of models to investigate how well these perform under slight distortions of input features. Especially for tasks that involve many (low-level) inputs, the application of deep neural networks brings new challenges. In the context of jet flavor tagging, adversarial attacks are used to probe a typical classifier's vulnerability and can be understood as a model for systematic uncertainties. A corresponding defense strategy, adversarial training, improves robustness, while maintaining high performance. Investigating the loss surface corresponding to the inputs and models in question reveals geometric interpretations of robustness, taking correlations into account.
format Preprint
id arxiv_https___arxiv_org_abs_2303_14511
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Improving robustness of jet tagging algorithms with adversarial training: exploring the loss surface
Stein, Annika
High Energy Physics - Experiment
Artificial Intelligence
Machine Learning
High Energy Physics - Phenomenology
Data Analysis, Statistics and Probability
In the field of high-energy physics, deep learning algorithms continue to gain in relevance and provide performance improvements over traditional methods, for example when identifying rare signals or finding complex patterns. From an analyst's perspective, obtaining highest possible performance is desirable, but recently, some attention has been shifted towards studying robustness of models to investigate how well these perform under slight distortions of input features. Especially for tasks that involve many (low-level) inputs, the application of deep neural networks brings new challenges. In the context of jet flavor tagging, adversarial attacks are used to probe a typical classifier's vulnerability and can be understood as a model for systematic uncertainties. A corresponding defense strategy, adversarial training, improves robustness, while maintaining high performance. Investigating the loss surface corresponding to the inputs and models in question reveals geometric interpretations of robustness, taking correlations into account.
title Improving robustness of jet tagging algorithms with adversarial training: exploring the loss surface
topic High Energy Physics - Experiment
Artificial Intelligence
Machine Learning
High Energy Physics - Phenomenology
Data Analysis, Statistics and Probability
url https://arxiv.org/abs/2303.14511