[2602.23369] Reason to Contrast: A Cascaded Multimodal Retrieval Framework

Computer Science > Information Retrieval arXiv:2602.23369 (cs) [Submitted on 21 Dec 2025] Title:Reason to Contrast: A Cascaded Multimodal Retrieval Framework Authors:Xuanming Cui, Hong-You Chen, Hao Yu, Hao Yuan, Zihao Wang, Shlok Kumar Mishra, Hanchao Yu, Yonghuan Yang, Jun Xiao, Ser-Nam Lim, Jianpeng Cheng, Qi Guo, Xiangjun Fan View a PDF of the paper titled Reason to Contrast: A Cascaded Multimodal Retrieval Framework, by Xuanming Cui and 12 other authors View PDF HTML (experimental) Abstract:Traditional multimodal retrieval systems rely primarily on bi-encoder architectures, where performance is closely tied to embedding dimensionality. Recent work, Think-Then-Embed (TTE), shows that incorporating multimodal reasoning to elicit additional informative tokens before embedding can further improve retrieval. In this paper, we extend this paradigm with TTE-v2, a hybrid multimodal retrieval framework that introduces reasoning-driven performance scaling based on additional input token budget rather than model or embedding size. Our approach augments the initial multimodal retrieval with additional reasoning steps for reranking, enabling more expressive query-candidate interactions at test time. The reranking stage further provides fine-grained supervision for hard negative mining and false negative filtering, creating a feedback loop that effectively strengthens the upstream retriever. This cascaded design delivers substantial test-time improvements based on intermediate reas...