@misc{indiciae9c0b5cd4bef4,
  title = {VISA: Group-wise Visual Token Selection and Aggregation via Graph Summarization for Efficient MLLMs Inference},
  author = {Pengfei Jiang and Hanjun Li and Linglan Zhao and Fei Chao and Ke Yan and Shouhong Ding and Rongrong Ji},
  year = {2025},
  doi = {10.1145/3746027.3755792},
  url = {https://arxiv.org/abs/2508.17857},
  note = {Source identifier: 2508.17857}
}