Han, Y., Liu, X., Zhang, Z., Ding, P., Chen, J., Chen, H., … Huang, S. (2026). Filter, Correlate, Compress: Training-Free Token Reduction for MLLM Acceleration. Proceedings of the AAAI Conference on Artificial Intelligence, 40(6), 4601–4609. https://doi.org/10.1609/aaai.v40i6.42460