MiniReranker: Efficient Multimodal Reranking Through Visual Cache Reuse and Interaction Sparsity

MCML Authors

Yunpu Ma

Dr.

→ Group Volker Tresp
 Database Systems, Data Mining and AI
→ Co-Group Hinrich Schütze

Abstract

Multimodal large language models (MLLMs) have recently shown strong potential as point-wise rerankers by directly modeling query--document relevance through next-token prediction. However, point-wise reranking suffers from substantial repeated computation across query--document pairs, while the causal structure of transformers allows only prefix segments to be reused via pre-caching. To address the misalignment of existing query-first and document-first formats with both VQA-style prompting and computation-aware reuse, we propose a vision-first formulation that improves both cache reuse efficiency and reranking performance. However, the remaining cost is still considerable and stems from three main sources: (1) model depth, for which we reduce active parameters via early exit; (2) cross-segment attention, which we restrict to a narrow interaction band across a few layers; and (3) visual tokens, where we reduce the number of tokens via embedder-guided pruning. Together, these designs form miniReranker, which reduces reranking runtime to <1% of the dense implementation under high-reuse settings for a single query, while preserving >96% of the dense model performance.

misc FLZ+26

---

Preprint

Jun. 2026

---

Authors

Y. Fan • X. Lu • A. Zhao • J. Tong • P. Nie • K. Zou • Y. Ma • W. Zhang • X. Shen

---

Links

arXiv

---

Research Area

 A3 | Computational Models