Efficient Generative Retrieval for E-commerce Search with Semantic Cluster IDs and Expert-Guided RL

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Zhu, Jianbo, Fang, Xing, Wang, Jing, Jin, Mingmin, Wang, Bokang, Song, Guangxin, Xie, Zhenyu, Bai, Junjie
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866916012062932992
author Zhu, Jianbo
Fang, Xing
Wang, Jing
Jin, Mingmin
Wang, Bokang
Song, Guangxin
Xie, Zhenyu
Bai, Junjie
author_facet Zhu, Jianbo
Fang, Xing
Wang, Jing
Jin, Mingmin
Wang, Bokang
Song, Guangxin
Xie, Zhenyu
Bai, Junjie
contents Generative retrieval offers a promising alternative by unifying the fragmented multi-stage retrieval process into a single end-to-end model. However, its practical adoption in industrial e-commerce search remains challenging, given the massive and dynamic product catalogs, strict latency requirements, and the need to align retrieval with downstream ranking goals. In this work, we propose a retrieval framework tailored for real-world recall scenarios, positioning generative retrieval as a recall-stage supplement rather than an end-to-end replacement. Our method, CQ-SID (Category-and-Query constrained Semantic ID), employs category-aware and query-item contrastive learning along with Residual Quantized VAEs to encode items into hierarchical semantic cluster identifiers, significantly reducing beam search complexity. Additionally, we develop EG-GRPO (Expert-Guided Group Relative Policy Optimization), a reinforcement learning approach that aligns generative recall with downstream ranking under sparse rewards by injecting ground-truth samples to stabilize training. Offline experiments on TmallAPP search logs show that CQ-SID achieves up to 26.76% and 11.11% relative gains in semantic and personalized click hitrate over RQ-VAE baselines, while halving beam search size. EG-GRPO further improves multi-objective performance. Online A/B tests confirm gains in GMV (+1.15%) and UCTCVR (+0.40%). The generative recall channel now contributes substantially in production, accounting for over 50.25% of exposures, 58.96% of clicks, and 72.63% of purchases, demonstrating a viable path for deploying generative retrieval in real-world e-commerce systems.
format Preprint
id arxiv_https___arxiv_org_abs_2605_14434
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Efficient Generative Retrieval for E-commerce Search with Semantic Cluster IDs and Expert-Guided RL
Zhu, Jianbo
Fang, Xing
Wang, Jing
Jin, Mingmin
Wang, Bokang
Song, Guangxin
Xie, Zhenyu
Bai, Junjie
Information Retrieval
Artificial Intelligence
Generative retrieval offers a promising alternative by unifying the fragmented multi-stage retrieval process into a single end-to-end model. However, its practical adoption in industrial e-commerce search remains challenging, given the massive and dynamic product catalogs, strict latency requirements, and the need to align retrieval with downstream ranking goals. In this work, we propose a retrieval framework tailored for real-world recall scenarios, positioning generative retrieval as a recall-stage supplement rather than an end-to-end replacement. Our method, CQ-SID (Category-and-Query constrained Semantic ID), employs category-aware and query-item contrastive learning along with Residual Quantized VAEs to encode items into hierarchical semantic cluster identifiers, significantly reducing beam search complexity. Additionally, we develop EG-GRPO (Expert-Guided Group Relative Policy Optimization), a reinforcement learning approach that aligns generative recall with downstream ranking under sparse rewards by injecting ground-truth samples to stabilize training. Offline experiments on TmallAPP search logs show that CQ-SID achieves up to 26.76% and 11.11% relative gains in semantic and personalized click hitrate over RQ-VAE baselines, while halving beam search size. EG-GRPO further improves multi-objective performance. Online A/B tests confirm gains in GMV (+1.15%) and UCTCVR (+0.40%). The generative recall channel now contributes substantially in production, accounting for over 50.25% of exposures, 58.96% of clicks, and 72.63% of purchases, demonstrating a viable path for deploying generative retrieval in real-world e-commerce systems.
title Efficient Generative Retrieval for E-commerce Search with Semantic Cluster IDs and Expert-Guided RL
topic Information Retrieval
Artificial Intelligence
url https://arxiv.org/abs/2605.14434