arXiv Machine Learning October 9, 2026 By Kamil Ciosek, Aleksandr V. Petrov, Nicol\`o Felicioni, Konstantina Palla

Fast Adversarial Attacks with Gradient Prediction

Read the original on arXiv Machine Learning →

The paper introduces a new class of adversarial attacks that eliminate the need for a backward pass by predicting input gradients from forward-pass hidden states using lightweight linear regression. The authors theoretically derive exact affine conditional gradient means, proving optimality in the idealized NTK regime, and empirically demonstrate that their methods recover most of FGSM’s performance while achieving a 532% increase in throughput on practical finite-width models. This approach offers a simple and general way to generate adversarial examples much faster under realistic wall‑clock constraints.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

arXiv AI

Aug 3

Efficient LLM Adversarial Training via Low-Rank Defense and Circuit-Guided Surrogates

arXiv:2607. 28959v1 Announce Type: cross Abstract: Adversarial training is one of the most effective defenses against adversarial attacks, yet the computational cost remains prohibitive at modern scales, especially for large language models (LLMs).

By Weiyi He, Yuping Lin, Jiliang Tang, Yue Xing

llms fine-tuning safety

arXiv Machine Learning

Oct 23

MIG-COW: Transferable Adversarial Attacks on Deepfake Detectors via Gradient Decomposition

MIG‑COW is a new adversarial attack framework that uses momentum‑integrated gradients and a consensus‑orthogonal decomposition to create highly generalizable, visually imperceptible perturbations against deepfake detectors. It achieves a 99.96% white‑box attack success rate on the AADD‑2025 Challenge benchmarks while maintaining high perceptual similarity (SSIM). In black‑box settings, its success rate drops to 7.16%, and adding low‑performing but diverse models to the ensemble can reduce attack effectiveness, highlighting the importance of careful surrogate model selection.

By Wonjune Seo, Joonhyuk Baek, Yeseong Jung, Saerom Park

multimodal benchmarks safety

arXiv AI

Jun 9

Defending Against Malicious Finetuning by Scaling Train-time Adversarial Attacks

arXiv:2606. 07970v1 Announce Type: cross Abstract: Current open-weight large language models (LLMs) are prone to malicious finetuning attacks, which could compromise the safety alignment of LLMs with only a few steps of supervised finetuning (SFT) on poisoned datasets.

By Haoming Wen, Shi Chen, Qingyu Shi, Siyuan Liu, Minrui Luo, Jingzhao Zhang, Tianxing He

llms fine-tuning safety

arXiv Machine Learning

Aug 17

Localization then Neutralization: Gradient-guided Token Suppression against Visual Prompt Injection Attack

arXiv:2605. 25194v2 Announce Type: replace Abstract: Adversarial images pose a severe security threat to multimodal large language models through prompt injection.

By Dongpeng Zhang, Ke Ma, Yangbangyan Jiang, Gaozheng Pei, Longtao Huang, Qianqian Xu, Qingming Huang

llms multimodal safety

arXiv Machine Learning

Aug 14

GENADA: efficient generative time series adversarial attack framework

arXiv:2608. 12535v1 Announce Type: new Abstract: Deep learning models are widely used for time series analysis in domains such as healthcare, finance, energy systems, and environmental monitoring.

By Michael Baronov, Denis Vorobev, Margarita Rusanova, Petr Sokerin, Alexey Zaytsev

safety

arXiv AI

Jun 24

Tuning without Peeking: Provable Generalization Bounds and Robust LLM Post-Training

arXiv:2507. 01752v4 Announce Type: replace-cross Abstract: Gradient-based optimization is the workhorse of deep learning, offering efficient and scalable training via backpropagation.

By Ismail Labiad, Mathurin Videau, Matthieu Kowalski, Marc Schoenauer, Alessandro Leite, Julia Kempe, Olivier Teytaud

llms benchmarks safety

One email a morning, machine-written

Related stories

Efficient LLM Adversarial Training via Low-Rank Defense and Circuit-Guided Surrogates

MIG-COW: Transferable Adversarial Attacks on Deepfake Detectors via Gradient Decomposition

Defending Against Malicious Finetuning by Scaling Train-time Adversarial Attacks

Localization then Neutralization: Gradient-guided Token Suppression against Visual Prompt Injection Attack

GENADA: efficient generative time series adversarial attack framework

Tuning without Peeking: Provable Generalization Bounds and Robust LLM Post-Training