Adversarial Training

Ask about this topic

Overview

Adversarial training studies how to make models perform reliably when inputs, environments, or data distributions are deliberately or naturally shifted. A common approach alternates between finding difficult examples—using methods such as projected-gradient attacks, distributionally robust optimization, or learned attackers—and training the model to handle them; related methods target latent representations, domains, or interaction policies. These techniques are used in vision, language-model safety, detection, and reinforcement learning, with evaluations ranging from bounded perturbation benchmarks to adaptive attacks and shifted environments. A central tension is that robustness to trained or tested threats may not transfer to unseen attacks, while stronger training can cost accuracy, utility, or computation.

Latest papers

CardsList

187 papers31 papers in the last four weeks, up 121% on the four weeks before. 0.2% of all new papers.

  1. destroR: A Benchmark and Adversarial-Training Defense for Bangla Transfer Models under Meaning-Preserving Attacks

    Nov 13, 2025Saadat Rafid Ahmed, Rubayet Shareen, Radoan Sharkar +3Adversarial Text AttacksAdversarial Robustness of Language Models

    Dept. of Computer Science and Engineering, BRAC University

  2. Solver-Integrated Adversarial Attacking and Training of Neural Operators

    Oct 21, 2025Yifei SunAdversarial TrainingAdversarial Attacks

  3. CogniDir: Combating Cognitive Malicious Comments via Adaptive Distributional Learning for Robust Fake News Detection

    Oct 10, 2025Zhao Tong, Chunlin Gong, Yimeng Gu +5Adversarial TrainingFake News Detection

    Institute of Information Engineering, Chinese Academy of Sciences · School of Cyber Security, University of Chinese Academy of Sciences · University of Minnesota +3

  4. On the Escaping Efficiency of Distributed Adversarial Training Algorithms

    Sep 14, 2025Ying Cao, Kun Yuan, Ali H. SayedAdversarial TrainingAdversarial Robustness

    Institute of Electrical and Micro Engineering in EPFL, Lausanne · Center for Machine Learning Research (CMLR) in Peking University

  5. Catastrophic Overfitting, Entropy Gap and Participation Ratio: A Noiseless lpl^p Norm Solution for Fast Adversarial Training

    May 5, 2025Fares B. Mehouachi, Saif Eddin JabariAdversarial TrainingAdversarial Robustness

    New York University of Abu Dhabi, Saadiyat Island, Abu Dhabi, UAE · Department of Civil and Urban Engineering, NYU Tandon School of Engineering, New York, USA

  6. Unified Enhancement of the Generalization and Robustness of Language Models via Bi-Stage Optimization

    Mar 19, 2025Yudao Sun, Juan Yin, Juan Zhao +3Adversarial TrainingNeural Network Generalization

    Department of New Networks, Peng Cheng Laboratory, Shenzhen 518055, China · Department of Industrial Systems Engineering and Management, National University of Singapore, Singapore 117576 · School of Data Science and Artificial Intelligence, Chang’an University, Shaanxi, 710064, China

  7. How Learning Dynamics Drive Adversarially Robust Generalization?

    Oct 10, 2024Yuelin Xu, Xiao ZhangAdversarial TrainingPAC-Bayesian Generalization Bounds

    CISPA Helmholtz Center for Information Security, Saarbrücken, Germany