[1]
Wang, M. et al. 2026. FantasyTalking2: Timestep-Layer Adaptive Preference Optimization for Audio-Driven Portrait Animation. Proceedings of the AAAI Conference on Artificial Intelligence. 40, 12 (Mar. 2026), 9966–9974. DOI:https://doi.org/10.1609/aaai.v40i12.37962.