SOURCE-LINKED INTELLIGENCE

MARS: What Retrieval Signals Are Hidden in Multimodal Large Language Models for Text-Video Retrieval?

arXiv · AI, language, vision and robotics · article · Sep 2, 2026 · UTC

Text-video retrieval requires representations that can distinguish videos with similar scenes, actions, and temporal patterns. Recent multimodal large language models have been adapted as embedding models, but they often represent each input using a single token from the final layer. This can compress diverse video-text cues into a single vector and limit fine-grained retrieval. To address this limitation, we propose MARS, a multi-layer and multi-slot embedding framework for text-video retrieval. MARS constructs multiple adaptive representation slots by combining hidden states from different d

Read original source ↗ Open in workspace

recordType

paper

region

Global