Very Efficient Listwise Multimodal Reranking for Long Documents

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sun, Yiqun, Wei, Pengfei, Hsieh, Lawrence B.
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910211911974912
author Sun, Yiqun
Wei, Pengfei
Hsieh, Lawrence B.
author_facet Sun, Yiqun
Wei, Pengfei
Hsieh, Lawrence B.
contents Listwise reranking is a key yet computationally expensive component in vision-centric retrieval and multimodal retrieval-augmented generation (M-RAG) over long documents. While recent VLM-based rerankers achieve strong accuracy, their practicality is often limited by long visual-token sequences and multi-step autoregressive decoding. We propose ZipRerank, a highly efficient listwise multimodal reranker that directly addresses both bottlenecks. It reduces input length via a lightweight query-image early interaction mechanism and eliminates autoregressive decoding by scoring all candidates in a single forward pass. To enable effective learning, ZipRerank adopts a two-stage training strategy: (i) listwise pretraining on large-scale text data rendered as images, and (ii) multimodal finetuning with VLM-teacher-distilled soft-ranking supervision. Extensive experiments on the MMDocIR benchmark show that ZipRerank matches or surpasses state-of-the-art multimodal rerankers while reducing LLM inference latency by up to an order of magnitude, making it well-suited for latency-sensitive real-world systems. The code is available at https://github.com/dukesun99/ZipRerank.
format Preprint
id arxiv_https___arxiv_org_abs_2605_11864
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Very Efficient Listwise Multimodal Reranking for Long Documents
Sun, Yiqun
Wei, Pengfei
Hsieh, Lawrence B.
Information Retrieval
Artificial Intelligence
Computer Vision and Pattern Recognition
Multimedia
Listwise reranking is a key yet computationally expensive component in vision-centric retrieval and multimodal retrieval-augmented generation (M-RAG) over long documents. While recent VLM-based rerankers achieve strong accuracy, their practicality is often limited by long visual-token sequences and multi-step autoregressive decoding. We propose ZipRerank, a highly efficient listwise multimodal reranker that directly addresses both bottlenecks. It reduces input length via a lightweight query-image early interaction mechanism and eliminates autoregressive decoding by scoring all candidates in a single forward pass. To enable effective learning, ZipRerank adopts a two-stage training strategy: (i) listwise pretraining on large-scale text data rendered as images, and (ii) multimodal finetuning with VLM-teacher-distilled soft-ranking supervision. Extensive experiments on the MMDocIR benchmark show that ZipRerank matches or surpasses state-of-the-art multimodal rerankers while reducing LLM inference latency by up to an order of magnitude, making it well-suited for latency-sensitive real-world systems. The code is available at https://github.com/dukesun99/ZipRerank.
title Very Efficient Listwise Multimodal Reranking for Long Documents
topic Information Retrieval
Artificial Intelligence
Computer Vision and Pattern Recognition
Multimedia
url https://arxiv.org/abs/2605.11864