TAROT: Towards Essentially Domain-Invariant Robustness with Theoretical Justification

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Yang, Dongyoon, Lee, Jihu, Kim, Yongdai
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908357547261952
author Yang, Dongyoon
Lee, Jihu
Kim, Yongdai
author_facet Yang, Dongyoon
Lee, Jihu
Kim, Yongdai
contents Robust domain adaptation against adversarial attacks is a critical research area that aims to develop models capable of maintaining consistent performance across diverse and challenging domains. In this paper, we derive a new generalization bound for robust risk on the target domain using a novel divergence measure specifically designed for robust domain adaptation. Building upon this, we propose a new algorithm named TAROT, which is designed to enhance both domain adaptability and robustness. Through extensive experiments, TAROT not only surpasses state-of-the-art methods in accuracy and robustness but also significantly enhances domain generalization and scalability by effectively learning domain-invariant features. In particular, TAROT achieves superior performance on the challenging DomainNet dataset, demonstrating its ability to learn domain-invariant representations that generalize well across different domains, including unseen ones. These results highlight the broader applicability of our approach in real-world domain adaptation scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2505_06580
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle TAROT: Towards Essentially Domain-Invariant Robustness with Theoretical Justification
Yang, Dongyoon
Lee, Jihu
Kim, Yongdai
Artificial Intelligence
Machine Learning
Robust domain adaptation against adversarial attacks is a critical research area that aims to develop models capable of maintaining consistent performance across diverse and challenging domains. In this paper, we derive a new generalization bound for robust risk on the target domain using a novel divergence measure specifically designed for robust domain adaptation. Building upon this, we propose a new algorithm named TAROT, which is designed to enhance both domain adaptability and robustness. Through extensive experiments, TAROT not only surpasses state-of-the-art methods in accuracy and robustness but also significantly enhances domain generalization and scalability by effectively learning domain-invariant features. In particular, TAROT achieves superior performance on the challenging DomainNet dataset, demonstrating its ability to learn domain-invariant representations that generalize well across different domains, including unseen ones. These results highlight the broader applicability of our approach in real-world domain adaptation scenarios.
title TAROT: Towards Essentially Domain-Invariant Robustness with Theoretical Justification
topic Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2505.06580