← Research

Robustness / Adversarial Training

SORA

Free Second Order Attacks in Fast Adversarial Training

Mazdak Teymourian*, Ramtin Moslemi*, Farzan Rahmani, Mohammad Hossein Rohban

* Equal contribution.

Supervised by Mohammad Hossein Rohban

Can useful second-order information be incorporated into fast adversarial training without giving up its computational advantages?

ICML 2026 · Accepted

Figure 1 from the SORA paper, showing the main method and contribution.
Figure 1. Overview of Epsilon Overfitting perspective and SORA method.
01 · Motivation

Fast adversarial training is attractive because it can obtain robustness at a fraction of the cost of multi-step adversarial training. At the same time, it can exhibit catastrophic overfitting: a model may become highly accurate against the attack used during training while failing against stronger evaluations.

02 · Research question

Can fast training use richer attack information without losing efficiency?

This work studies whether efficient approximations to second-order attack information can make fast adversarial training more reliable and mitigate catastrophic overfitting.

03 · Approach

SORA uses an efficient second-order attack approximation inside fast adversarial training. The project combines analysis of catastrophic overfitting with experiments across multiple datasets, architectures, and threat models.

04 · Benchmarking

The accompanying PyTorch codebase reimplements SORA and 14 named comparison methods in a unified framework, including FGSM, FGSM-RS, GradAlign, NuAT, ATAS, NFGSM, ZeroGrad, MultiGrad, AAER, ELLE, FreeAT, PGD-2, PGD-10, and TRADES. It supports training these methods on PreActResNet, ResNet, WideResNet, SENet, and ViT architectures across CIFAR-10/100, TinyImageNet, ImageNet100, PathMNIST, and TissueMNIST.

The repository also includes visualization code used to inspect and compare training and robustness behavior.

My contribution

I conceived the core idea, derived the underlying theory, implemented the method, and co-wrote the manuscript. I also carried out extensive experimental evaluation and robustness benchmarking.

05 · Why this work interests me

Robustness is, to me, more than a desirable property of a machine learning model. It is connected to what I expect from an intelligent system: it should not be easily fooled by superficial changes in its input. At the same time, I am interested in achieving this reliability efficiently rather than simply scaling computation.

06 · Takeaway

This project strengthened my interest in the relationship between robustness and efficiency, and in understanding why a model that appears robust under one evaluation can fail under another.