Ishmam, A.M. et al. 2026. LAMP: Learning Universal Adversarial Perturbations for Multi-Image Tasks via Pre-trained Models. Proceedings of the AAAI Conference on Artificial Intelligence. 40, 7 (Mar. 2026), 5267–5275. DOI:https://doi.org/10.1609/aaai.v40i7.37442.