Ma, Y., Zhou, Q., Wang, Z., Chen, X., Yang, H., Song, J., & Zheng, B. (2026). Contribution-aware Token Compression for Efficient Video Understanding via Reinforcement Learning. Proceedings of the AAAI Conference on Artificial Intelligence, 40(29), 24317–24325. https://doi.org/10.1609/aaai.v40i29.39612