Continual Learning Framework for Long-Sequence Video Understanding With Memory-Aware Visual Representations. JAAIR 2026, 1 (1).