Beyond Classification: Dynamic Adapter Routing for Continual Multimodal Retrieval

ArXi:2605.31229v1 Announce Type: cross While retrieval is a core function of vision-language models, continually updating these models for retrieval tasks remains critically underexplored. Existing work often approaches continual retrieval through the lens of class-incremental learning (CIL), evaluating both standard CIL methods and retrieval-oriented adaptations in settings that may not fully capture the retrieval-specific dynamics. To address this, we