@article{visiontokenreductionviaattentiondrivense,
  title = {Vision Token Reduction via Attention-Driven Self-Compression for Efficient Multimodal Large Language Models},
  author = {Omer Faruk Deniz and Ruiyu Mao and Ruochen Li and Yapeng Tian and Latifur Khan},
  year = {2026},
  eprint = {2602.12618},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2602.12618},
}