2026· IEEE Transactions on Reliability· Vol 75, pp. 2668-2682· 0 citations· 56 references
Abstract
Pretrained programming language models (PLMs) have shown strong performance in software engineering tasks, but remain vulnerable to adversarial attacks. Traditional methods rely on narrow, static perturbation strategies, leading to poor diversity and a combinatorial explosion of the search space when multiple operations are applied. To address these limitations, we propose CodeRL-IA, an adversarial attack method that integrates reinforcement learning and importance analysis to dynamically optimize semantics-preserving perturbation strategies, achieving high attack effectiveness while maintaining code quality. Extensive experiments on code summarization, code translation, and defect detection tasks demonstrate that CodeRL-IA outperforms the baselines, with average attack success rates of 9.8% and 14.87% higher than those of the baseline methods, while keeping comparable code quality.
Adversarial robustness remains a significant challenge in machine-learning malware detection. We propose a co-evolutionary adversarial training framework that integrates gradient-based adversarial attacks with evolutionary controllers to improve robustness. A hybrid Convolutional Neural Network (CNN-MLP) processes spatial and vector-based features, while evolved controllers adaptively guide Projected Gradient Descent (PGD) using gradient statistics, optimizing adversarial effectiveness while respecting feature constraints. This online, co-evolutionary process exposes the classifier to increasingly adaptive attacks. Experimental results demonstrate that the proposed approach enhances adversarial robustness relative to standard training while maintaining clean accuracy, exemplifying a co-evolutionary arms-race framework for realistic, feature-constrained malware detection.
Sabre Didi, Geoff S. Nitschke· Proceedings of the Genetic a...· 0 citations
This survey provides a comprehensive synthesis of adversarial attacks and defense mechanisms in modern AI security, introducing a structured taxonomy categorizing attacks into evasion, poisoning, and model inversion strategies, evaluated across varying levels of attacker knowledge.
P. Mahalakshmi, D. Jayalakshmi· International Journal of Inn...· 0 citations
Predictive models are widely used in many fields, but are vulnerable to false data injection attacks. To address this, detection schemes and adversarial training have been proposed, but such approaches lack guarantees against stealthy attacks. We therefore propose a detector-based switched model, in which optimal attack strategies are stealthy. For linear prediction models, we derive a convex formulation of the resulting adversarial risk. The model incorporates protected features and introduces a hyperparameter modelling attack probability, enabling an explicit performance trade-off between clean and attacked data regimes. Numerical simulations on real and synthetic data show improved performance on partially attacked data, even for misspecified attack probabilities.
Lovisa Eriksson, Dave Zachariah, A. Teixeira· 0 citations
Gradient-seeded Reinforcement Learning And Stealthy Pruning (GRASP), a three-stage framework that tackles challenges of adversarial attacks on machine learning-based malware detectors, and out-performs baselines, achieving higher attack success with fewer queries and smaller file-size inflation.
Yutong Liu, Jianting Ning, Qi Feng et al.· 0 citations
This work proposes BMAT (Bilevel-Minimax Adversarial Transfer), an integrated bottom-up solver that combines a Soft Weight Modulator and an Implicit Gradient Approximator to enable ternary coupling among initialization, surrogate adaptation, and perturbation optimization.
