Pailitao-VL: Unified Embedding and Reranker for Real-Time Multi-Modal Industrial Search

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Lei, Ju, Chen, Chen, Xu, Wang, Zhicheng, Jiao, Yuheng, Zhan, Hongfeng, Li, Zhaoyang, Xu, Shihao, Zhao, Zhixiang, Jia, Tong, Li, Lin, Gao, Yuan, Song, Jun, Lan, Jinsong, Zhu, Xiaoyong, Zheng, Bo
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914371423174656
author Chen, Lei
Ju, Chen
Chen, Xu
Wang, Zhicheng
Jiao, Yuheng
Zhan, Hongfeng
Li, Zhaoyang
Xu, Shihao
Zhao, Zhixiang
Jia, Tong
Li, Lin
Gao, Yuan
Song, Jun
Lan, Jinsong
Zhu, Xiaoyong
Zheng, Bo
author_facet Chen, Lei
Ju, Chen
Chen, Xu
Wang, Zhicheng
Jiao, Yuheng
Zhan, Hongfeng
Li, Zhaoyang
Xu, Shihao
Zhao, Zhixiang
Jia, Tong
Li, Lin
Gao, Yuan
Song, Jun
Lan, Jinsong
Zhu, Xiaoyong
Zheng, Bo
contents In this work, we presented Pailitao-VL, a comprehensive multi-modal retrieval system engineered for high-precision, real-time industrial search. We here address three critical challenges in the current SOTA solution: insufficient retrieval granularity, vulnerability to environmental noise, and prohibitive efficiency-performance gap. Our primary contribution lies in two fundamental paradigm shifts. First, we transitioned the embedding paradigm from traditional contrastive learning to an absolute ID-recognition task. Through anchoring instances to a globally consistent latent space defined by billions of semantic prototypes, we successfully overcome the stochasticity and granularity bottlenecks inherent in existing embedding solutions. Second, we evolved the generative reranker from isolated pointwise evaluation to the compare-and-calibrate listwise policy. By synergizing chunk-based comparative reasoning with calibrated absolute relevance scoring, the system achieves nuanced discriminative resolution while circumventing the prohibitive latency typically associated with conventional reranking methods. Extensive offline benchmarks and online A/B tests on Alibaba e-commerce platform confirm that Pailitao-VL achieves state-of-the-art performance and delivers substantial business impact. This work demonstrates a robust and scalable path for deploying advanced MLLM-based retrieval architectures in demanding, large-scale production environments.
format Preprint
id arxiv_https___arxiv_org_abs_2602_13704
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Pailitao-VL: Unified Embedding and Reranker for Real-Time Multi-Modal Industrial Search
Chen, Lei
Ju, Chen
Chen, Xu
Wang, Zhicheng
Jiao, Yuheng
Zhan, Hongfeng
Li, Zhaoyang
Xu, Shihao
Zhao, Zhixiang
Jia, Tong
Li, Lin
Gao, Yuan
Song, Jun
Lan, Jinsong
Zhu, Xiaoyong
Zheng, Bo
Information Retrieval
Artificial Intelligence
Computer Vision and Pattern Recognition
In this work, we presented Pailitao-VL, a comprehensive multi-modal retrieval system engineered for high-precision, real-time industrial search. We here address three critical challenges in the current SOTA solution: insufficient retrieval granularity, vulnerability to environmental noise, and prohibitive efficiency-performance gap. Our primary contribution lies in two fundamental paradigm shifts. First, we transitioned the embedding paradigm from traditional contrastive learning to an absolute ID-recognition task. Through anchoring instances to a globally consistent latent space defined by billions of semantic prototypes, we successfully overcome the stochasticity and granularity bottlenecks inherent in existing embedding solutions. Second, we evolved the generative reranker from isolated pointwise evaluation to the compare-and-calibrate listwise policy. By synergizing chunk-based comparative reasoning with calibrated absolute relevance scoring, the system achieves nuanced discriminative resolution while circumventing the prohibitive latency typically associated with conventional reranking methods. Extensive offline benchmarks and online A/B tests on Alibaba e-commerce platform confirm that Pailitao-VL achieves state-of-the-art performance and delivers substantial business impact. This work demonstrates a robust and scalable path for deploying advanced MLLM-based retrieval architectures in demanding, large-scale production environments.
title Pailitao-VL: Unified Embedding and Reranker for Real-Time Multi-Modal Industrial Search
topic Information Retrieval
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2602.13704