![[2602.19931] Expanding the Role of Diffusion Models for Robust Classifier Training](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.19931] Expanding the Role of Diffusion Models for Robust Classifier Training
Summary
This article explores the use of diffusion models to enhance adversarial training for robust image classifiers, demonstrating improved performance through diverse representations.
Why It Matters
As machine learning models face increasing challenges from adversarial attacks, this research highlights the potential of diffusion models to improve classifier robustness, which is crucial for applications in security-sensitive areas such as autonomous vehicles and healthcare.
Key Takeaways
- Diffusion models can generate synthetic data that enhances adversarial training.
- Incorporating diffusion representations as auxiliary signals improves classifier robustness.
- The study shows that diffusion representations encourage more disentangled feature learning.
- Experiments validate the effectiveness of combining diffusion representations with synthetic data.
- Robustness improvements were observed across multiple datasets, including CIFAR-10 and ImageNet.
Computer Science > Machine Learning arXiv:2602.19931 (cs) [Submitted on 23 Feb 2026] Title:Expanding the Role of Diffusion Models for Robust Classifier Training Authors:Pin-Han Huang, Shang-Tse Chen, Hsuan-Tien Lin View a PDF of the paper titled Expanding the Role of Diffusion Models for Robust Classifier Training, by Pin-Han Huang and 2 other authors View PDF HTML (experimental) Abstract:Incorporating diffusion-generated synthetic data into adversarial training (AT) has been shown to substantially improve the training of robust image classifiers. In this work, we extend the role of diffusion models beyond merely generating synthetic data, examining whether their internal representations, which encode meaningful features of the data, can provide additional benefits for robust classifier training. Through systematic experiments, we show that diffusion models offer representations that are both diverse and partially robust, and that explicitly incorporating diffusion representations as an auxiliary learning signal during AT consistently improves robustness across settings. Furthermore, our representation analysis indicates that incorporating diffusion models into AT encourages more disentangled features, while diffusion representations and diffusion-generated synthetic data play complementary roles in shaping representations. Experiments on CIFAR-10, CIFAR-100, and ImageNet validate these findings, demonstrating the effectiveness of jointly leveraging diffusion representatio...
