@misc{indiciaea1a989f9b520,
  title = {InduceKV: Fixed-Footprint Continual Adaptation of Multimodal LLMs via Inducing KV Memories},
  author = {Qianyu Chen and Ziteng Feng and Canran Xiao and Runxuan Tang},
  year = {2026},
  url = {https://arxiv.org/abs/2607.02010},
  note = {Source identifier: 2607.02010}
}