Multimodal Continual Learning with MLLMs from Multi-scenario Perspectives
多场景视角下的多模态持续学习与MLLMs
AI总结 本文提出UNIFIER框架,通过视觉表征扩展和视觉一致性约束解决多场景下的持续学习问题,提升跨场景视觉问答和F1分数。
Comments 22 pages, 17 figures. This is a preprint version of a paper submitted to ICML 2026