SCoTER: Structured Chain-of-Thought Transfer for Enhanced Recommendation

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Jiang, Jie, Wu, Yang, Li, Qian, Xiong, Yuling, Tang, Hongbo, Liu, Xun, Wang, Haoze, Zhang, Jun, Yu, Huan, Shi, Hailong
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918354974932992
author Jiang, Jie
Wu, Yang
Li, Qian
Xiong, Yuling
Tang, Hongbo
Liu, Xun
Wang, Haoze
Zhang, Jun
Yu, Huan
Shi, Hailong
author_facet Jiang, Jie
Wu, Yang
Li, Qian
Xiong, Yuling
Tang, Hongbo
Liu, Xun
Wang, Haoze
Zhang, Jun
Yu, Huan
Shi, Hailong
contents Harnessing the reasoning power of Large Language Models (LLMs) for recommender systems is hindered by two fundamental challenges. First, current approaches lack a mechanism for automated, data-driven discovery of effective reasoning patterns, relying instead on brittle manual templates or unstable zero-shot prompting. Second, they employ structure-collapsing integration: direct prompting incurs prohibitive online inference costs, while feature extraction collapses reasoning chains into single vectors, discarding stepwise logic. To address these challenges, we propose SCoTER (Structured Chain-of-Thought Transfer for Enhanced Recommendation), a unified framework that treats pattern discovery and structure-aware transfer as a jointly optimized problem. Specifically, SCoTER operationalizes this through two synergistic components: a Generate-Validate-Mine (GVM) pipeline for automated pattern discovery and a structure-preserving integration architecture that transfers stepwise logic to efficient models. Empirically, experiments on four benchmarks demonstrate consistent improvements across diverse backbones. Moreover, in production deployment on the Tencent Advertising Platform, SCoTER achieved a 2.14\% lift in Gross Merchandise Value (GMV) while eliminating online LLM inference costs. Overall, SCoTER presents a practical and unified framework for integrating structured LLM reasoning into recommender systems, validated by consistent improvements in both offline benchmarks and online production environments.
format Preprint
id arxiv_https___arxiv_org_abs_2511_19514
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SCoTER: Structured Chain-of-Thought Transfer for Enhanced Recommendation
Jiang, Jie
Wu, Yang
Li, Qian
Xiong, Yuling
Tang, Hongbo
Liu, Xun
Wang, Haoze
Zhang, Jun
Yu, Huan
Shi, Hailong
Information Retrieval
Harnessing the reasoning power of Large Language Models (LLMs) for recommender systems is hindered by two fundamental challenges. First, current approaches lack a mechanism for automated, data-driven discovery of effective reasoning patterns, relying instead on brittle manual templates or unstable zero-shot prompting. Second, they employ structure-collapsing integration: direct prompting incurs prohibitive online inference costs, while feature extraction collapses reasoning chains into single vectors, discarding stepwise logic. To address these challenges, we propose SCoTER (Structured Chain-of-Thought Transfer for Enhanced Recommendation), a unified framework that treats pattern discovery and structure-aware transfer as a jointly optimized problem. Specifically, SCoTER operationalizes this through two synergistic components: a Generate-Validate-Mine (GVM) pipeline for automated pattern discovery and a structure-preserving integration architecture that transfers stepwise logic to efficient models. Empirically, experiments on four benchmarks demonstrate consistent improvements across diverse backbones. Moreover, in production deployment on the Tencent Advertising Platform, SCoTER achieved a 2.14\% lift in Gross Merchandise Value (GMV) while eliminating online LLM inference costs. Overall, SCoTER presents a practical and unified framework for integrating structured LLM reasoning into recommender systems, validated by consistent improvements in both offline benchmarks and online production environments.
title SCoTER: Structured Chain-of-Thought Transfer for Enhanced Recommendation
topic Information Retrieval
url https://arxiv.org/abs/2511.19514