MemRerank: Preference Memory for Personalized Product Reranking

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Peng, Zhiyuan, Wu, Xuyang, Tou, Huaixiao, Fang, Yi, Gong, Yu
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915908087185408
author Peng, Zhiyuan
Wu, Xuyang
Tou, Huaixiao
Fang, Yi
Gong, Yu
author_facet Peng, Zhiyuan
Wu, Xuyang
Tou, Huaixiao
Fang, Yi
Gong, Yu
contents LLM-based shopping agents increasingly rely on long purchase histories and multi-turn interactions for personalization, yet naively appending raw history to prompts is often ineffective due to noise, length, and relevance mismatch. We propose MemRerank, a preference memory framework that distills user purchase history into concise, query-independent signals for personalized product reranking. To study this problem, we build an end-to-end benchmark and evaluation framework centered on an LLM-based \textbf{1-in-5} selection task, which measures both memory quality and downstream reranking utility. We further train the memory extractor with reinforcement learning (RL), using downstream reranking performance as supervision. Experiments with two LLM-based rerankers show that MemRerank consistently outperforms no-memory, raw-history, and off-the-shelf memory baselines, yielding up to \textbf{+10.61} absolute points in 1-in-5 accuracy. These results suggest that explicit preference memory is a practical and effective building block for personalization in agentic e-commerce systems.
format Preprint
id arxiv_https___arxiv_org_abs_2603_29247
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle MemRerank: Preference Memory for Personalized Product Reranking
Peng, Zhiyuan
Wu, Xuyang
Tou, Huaixiao
Fang, Yi
Gong, Yu
Computation and Language
Artificial Intelligence
Machine Learning
LLM-based shopping agents increasingly rely on long purchase histories and multi-turn interactions for personalization, yet naively appending raw history to prompts is often ineffective due to noise, length, and relevance mismatch. We propose MemRerank, a preference memory framework that distills user purchase history into concise, query-independent signals for personalized product reranking. To study this problem, we build an end-to-end benchmark and evaluation framework centered on an LLM-based \textbf{1-in-5} selection task, which measures both memory quality and downstream reranking utility. We further train the memory extractor with reinforcement learning (RL), using downstream reranking performance as supervision. Experiments with two LLM-based rerankers show that MemRerank consistently outperforms no-memory, raw-history, and off-the-shelf memory baselines, yielding up to \textbf{+10.61} absolute points in 1-in-5 accuracy. These results suggest that explicit preference memory is a practical and effective building block for personalization in agentic e-commerce systems.
title MemRerank: Preference Memory for Personalized Product Reranking
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2603.29247