一作ACL 2026Findings· Findings of ACL
M i MIC : Mitigating Visual Modality Collapse in Universal Multimodal Retrieval While Avoiding Semantic Misalignment
Li Juan · Chuanghao Ding · Xujie Zhang · Cam-Tu Nguyen
摘要
Universal Multimodal Retrieval (UMR) aims to map different modalities (e.g., visual and textual) into a shared embedding space for multi-modal retrieval. Existing UMR methods can be broadly divided into two categories: early-fusion approaches, such as Marvel, which projects visual features into t...