A Minimax Optimization Framework for Robust Deep Learning under Adversarial Perturbations. JMTA 2026, 11 (3s), 1-14.