1 paper · 1 filter
Siyuan Li, Youyuan Zhang, Ruitong Liu +2
Online multimodal knowledge editing requires injecting a continual stream of visual-textual corrections into multimodal large language models (MLLMs) with bounded overhead and mini…