[1]郭世圆,汪佳茵,孙培杰,等.协同信号增强的大模型用户画像生成与推荐[J].智能系统学报,2026,21(2):487-497.[doi:10.11992/tis.202506031]
GUO Shiyuan,WANG Jiayin,SUN Peijie,et al.Collaborative signal enhanced LLM user profiling and recommendation[J].CAAI Transactions on Intelligent Systems,2026,21(2):487-497.[doi:10.11992/tis.202506031]
点击复制
GUO Shiyuan,WANG Jiayin,SUN Peijie,et al.Collaborative signal enhanced LLM user profiling and recommendation[J].CAAI Transactions on Intelligent Systems,2026,21(2):487-497.[doi:10.11992/tis.202506031]
协同信号增强的大模型用户画像生成与推荐
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
21
期数:
2026年第2期
页码:
487-497
栏目:
学术论文—智能系统
出版日期:
2026-03-05
- Title:
- Collaborative signal enhanced LLM user profiling and recommendation
- Keywords:
- information retrieval; recommender system; large language model; user profile; user modeling; contrastive learning; collaborative filtering; user representation; feature embedding
- 分类号:
- TP311.13；TP309
- 摘要:
- 用户画像的质量直接影响推荐系统的表现。传统的推荐系统通过建模用户与物品间的协同信息来获得用户画像，无法充分利用用户与物品的文本描述信息。大模型处理文本信息、常识推理的能力及其拥有的世界知识，为用户画像建模提供了新的机会。二者的结合可充分发挥彼此优点，共同提高表现。本文提出了将用户的潜在兴趣与协同等级这两个来自推荐系统的协同信号引入大模型，增强其生成用户画像的方法。用户画像生成通过与大模型进行多次交互的方式，生成的用户画像进一步转换为特征向量，通过对比学习与推荐系统内的用户表征相融合，以增强个性化推荐表现。在两个数据集、多个推荐模型上的实验结果表明，本文方法能够显著提升推荐模型的表现。本文方法弥合了大模型与推荐系统间的鸿沟，为后续类似研究工作启发了新思路。
- Abstract:
- The quality of the user profile directly affects the performance of the recommender system. User profile in a traditional recommender system can be derived through modeling the collaborative information between users and items, but is unable to fully utilize the text description information of users and items. The textual information processing and commonsense reasoning capabilities of LLMs, combined with their world knowledge, provide new opportunities for user profiling. The combination of a recommender system and LLM can give full play to the advantages of each other, and improve each other’s performance mutually. This paper proposes a method to introduce two collaborative signals, named potential interest and collaborative scale, from a recommender system into LLM to further enhance the user profile generation of LLM. The user profile is generated through multiple times of interaction with LLM, and is further transformed into an embedding, fusing with the user representation in the recommender system through contrastive learning to improve recommendation performance. Experimental results on two datasets and multiple recommender models show that the proposed method can significantly improve the performance of the recommender model. The proposed method bridges the gap between LLM and recommender system, and sheds light on further similar research work.
- 参考文献/References:
-
[1] CHEN Hao, BEI Yuanchen, SHEN Qijie, et al. Macro graph neural networks for online billion-scale recommender systems[C]//Proceedings of the ACM Web Conference 2024. Singapore: ACM, 2024: 3598-3608.
[2] FENG Zhichao, XIE Junjie, LI Kaiyuan, et al. Context-based fast recommendation strategy for long user behavior sequence in Meituan Waimai[C]//Companion Proceedings of the ACM Web Conference 2024. Singapore: ACM, 2024: 355-363.
[3] DONG Zhikang, LIU Xiulong, CHEN Bin, et al. MuseChat: a conversational music recommendation system for videos[C]//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle: IEEE, 2024: 12775-12785.
[4] HASAN M R, FERDOUS J. Dominance of AI and machine learning techniques in hybrid movie recommendation system applying text-to-number conversion and cosine similarity approaches[J]. Journal of computer science and technology studies, 2024, 6(1): 94-102
[5] YANG Xinran, ZHANG Liaoniao, FENG Zixin. Personalized tourism recommendations and the E-tourism user experience[J]. Journal of travel research, 2024, 63(5): 1183-1200
[6] BILLSUS D, PAZZANI M J. Learning collaborative information filters[C]//International Conference on Machine Learning. Madison: PMLR, 1998: 46-54.
[7] SARWAR B, KARYPIS G, KONSTAN J, et al. Item-based collaborative filtering recommendation algorithms[C]//Proceedings of the 10th International Conference on World Wide Web. Hong Kong: ACM, 2001: 285-295.
[8] PURIFICATO E, BORATTO L, DE LUCA E W. User modeling and user profiling: a comprehensive survey[EB/OL]. (2024-02-15)[2025-06-25]. https://arxiv.org/abs/2402.09660.
[9] ADOMAVICIUS G, TUZHILIN A. Expert-driven validation of rule-based user models in personalization applications[J]. Data mining and knowledge discovery, 2001, 5(1): 33-58
[10] CHEUNG CHIU B, WEBB G I. Using decision trees for agent modeling: improving prediction performance[J]. User modeling and user-adapted interaction, 1998, 8(1): 131-152
[11] KIM H N, ALKHALDI A, EL SADDIK A, et al. Collaborative user modeling with user-generated tags for social recommender systems[J]. Expert systems with applications, 2011, 38(7): 8488-8496
[12] S?NCHEZ P, BELLOG?N A. Building user profiles based on sequences for content and collaborative filtering[J]. Information processing & management, 2019, 56(1): 192-211
[13] P N, SAITEJA K, RAM K K, et al. University recommender system based on student profile using feature weighted algorithm and KNN[C]//2022 International Conference on Sustainable Computing and Data Communication Systems. Erode: IEEE, 2022: 479-484.
[14] ZIGORIS P, ZHANG Yi. Bayesian adaptive user profiling with explicit & implicit feedback[C]//Proceedings of the 15th ACM International Conference on Information and Knowledge Management. Arlington: ACM, 2006: 397.
[15] ZHANG Yi, KOREN J. Efficient Bayesian hierarchical user modeling for recommendation system[C]//Proceedings of the 30th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval. Amsterdam: ACM, 2007: 47-54.
[16] XIE Yi, YU Shunzheng. A large-scale hidden semi-Markov model for anomaly detection on user browsing behaviors[J]. IEEE/ACM transactions on networking, 2009, 17(1): 54-65
[17] TANG Jie, YAO Limin, ZHANG Duo, et al. A combination approach to web user profiling[J]. ACM transactions on knowledge discovery from data, 2010, 5(1): 1-44
[18] AHMAD S B S, RAFIE M, GHORABIE S M. Spam detection on Twitter using a support vector machine and users’ features by identifying their interactions[J]. Multimedia tools and applications, 2021, 80(8): 11583-11605
[19] CHEN Tinggui, YIN Xiaohua, PENG Lijuan, et al. Monitoring and recognizing enterprise public opinion from high-risk users based on user portrait and random forest algorithm[J]. Axioms, 2021, 10(2): 106
[20] TRIPATHI A, T S A, GUDDETI R M R. A reinforcement learning and recurrent neural network based dynamic user modeling system[C]//2018 IEEE 18th International Conference on Advanced Learning Technologies. Mumbai: IEEE, 2018: 411-415.
[21] YU Zeping, LIAN Jianxun, MAHMOODY A, et al. Adaptive user modeling with long and short-term preferences for personalized recommendation[C]//Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence. Macao: IJCAI, 2019: 4213-4219.
[22] GE Songwei, DOU Zhicheng, JIANG Zhengbao, et al. Personalizing search results using hierarchical RNN with query-aware attention[C]//Proceedings of the 27th ACM International Conference on Information and Knowledge Management. Torino: ACM, 2018: 347-356.
[23] ZHU Y, LI Hao, LIAO Yikang, et al. What to do next: modeling user behaviors by time-LSTM[C]//International Joint Conference on Artificial Intelligence. Melbourne: IJCAI, 2017: 3602-3608.
[24] MA Zhengyi, DOU Zhicheng, BIAN Guanyue, et al. PSTIE: time information enhanced personalized search[C]//Proceedings of the 29th ACM International Conference on Information & Knowledge Management. Virtual Event: ACM, 2020: 1075-1084.
[25] GU Yulong, DING Zhuoye, WANG Shuaiqiang, et al. Hierarchical user profiling for E-commerce recommender systems[C]//Proceedings of the 13th International Conference on Web Search and Data Mining. Houston: ACM, 2020: 223-231.
[26] QI Tao, WU Fangzhao, WU Chuhan, et al. News recommendation with candidate-aware user modeling[C]//Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval. Madrid: ACM, 2022: 1917-1921.
[27] AVNY BROSH T, LIVNE A, SAR SHALOM O, et al. BRUCE: bundle recommendation using contextualized item embeddings[C]//Proceedings of the 16th ACM Conference on Recommender Systems. Seattle: ACM, 2022: 237-245.
[28] CHENG Zhiyong, HAN Sai, LIU Fan, et al. Multi-behavior recommendation with cascading graph convolution networks[C]//Proceedings of the ACM Web Conference 2023. Austin: ACM, 2023: 1181-1189.
[29] HAN Jinkun, LI Wei, CAI Zhipeng, et al. Multi-aggregator time-warping heterogeneous graph neural network for personalized micro-video recommendation[C]//Proceedings of the 31st ACM International Conference on Information & Knowledge Management. Atlanta: ACM, 2022: 676-685.
[30] WANG Xiaolin, SUN Guohao, FANG Xiu, et al. Modeling spatio-temporal neighbourhood for personalized point-of-interest recommendation[C]//Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence. Vienna: IJCAI, 2022: 3530-3536.
[31] TANG Xiangru, ZHANG Xingyao, SHAO Yanjun, et al. Step-back profiling: distilling user history for personalized scientific writing[EB/OL]. (2024-06-20)[2025-06-25]. https://arxiv.org/abs/2406.14275.
[32] ZHANG Jiarui. Guided profile generation improves personalization with LLMs[EB/OL]. (2024-09-19)[2025-06-25]. https://arxiv.org/abs/2409.13093.
[33] SUN Zhongxiang, SI Zihua, ZANG Xiaoxue, et al. Large language models enhanced collaborative filtering[C]//Proceedings of the 33rd ACM International Conference on Information and Knowledge Management. Boise: ACM, 2024: 2178-2188.
[34] ZHOU J, DAI Yijia, JOACHIMS T. Language-based user profiles for recommendation[EB/OL]. (2024-02-23)[2025-06-25]. https://arxiv.org/abs/2402.15623.
[35] JALAN R, PRAKASH T, PEDANEKAR N. LLM-BRec: personalizing session-based social recommendation with LLM-BERT fusion framework[C]//The Second Workshop on Generative Information Retrieval. Washington: ACM, 2024: 1-10.
[36] XI Yunjia, LIU Weiwen, LIN Jianghao, et al. Towards open-world recommendation with knowledge augmentation from large language models[C]//18th ACM Conference on Recommender Systems. Bari: ACM, 2024: 12-22.
[37] WEI Wei, REN Xubin, TANG Jiabin, et al. LLMRec: large language models with graph augmentation for recommendation[C]//Proceedings of the 17th ACM International Conference on Web Search and Data Mining. Merida: ACM, 2024: 806-815.
[38] REN Xubin, WEI Wei, XIA Lianghao, et al. Representation learning with large language models for recommendation[C]//Proceedings of the ACM Web Conference 2024. Singapore: ACM, 2024: 3464-3475.
[39] TANG Zuoli, HUAN Zhaoxin, LI Zihao, et al. One model for all: large language models are domain-agnostic recommendation systems[J]. ACM transactions on information systems, 2025, 43(5): 1-27
[40] DU Yingpeng, WANG Ziyan, SUN Zhu, et al. Large language model with graph convolution for recommendation[EB/OL]. (2024-02-14)[2025-06-25]. https://arxiv.org/abs/2402.08859.
[41] ZHENG Zhi, CHAO Wenshuo, QIU Zhaopeng, et al. Harnessing large language models for text-rich sequential recommendation[C]//Proceedings of the ACM Web Conference 2024. Singapore: ACM, 2024: 3207-3216.
[42] EBRAT D, RUEDA L. Lusifer: LLM-based user simulated feedback environment for online recommender systems[C]//2025 IEEE 4th International Conference on Computing and Machine Intelligence. MI: IEEE, 2025: 1-6.
[43] ZHANG Yang, FENG Fuli, ZHANG Jizhi, et al. CoLLM: integrating collaborative embeddings into large language models for recommendation[J]. IEEE transactions on knowledge and data engineering, 2025, 37(5): 2329-2340
[44] KIM S, KANG H, CHOI S, et al. Large language models meet collaborative filtering: an efficient all-round LLM-based recommender system[C]//Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. Barcelona: ACM, 2024: 1395-1406.
[45] LIU Zhongzhou, ZHANG Hao, DONG Kuicai, et al. Collaborative cross-modal fusion with large language model for recommendation[C]//Proceedings of the 33rd ACM International Conference on Information and Knowledge Management. Boise: ACM, 2024: 1565-1574.
[46] ZHAO Chuang, SU Xing, HE Ming, et al. Collaborative knowledge fusion: a novel method for multi-task recommender systems via LLMs[J]. IEEE transactions on knowledge and data engineering, 2025, 37(9): 5017-5033
[47] VAN DEN OORD A, LI Yazhe, VINYALS O. Representation learning with contrastive predictive coding[EB/OL]. (2018-07-10)[2025-06-25]. https://arxiv.org/abs/1807.03748.
[48] HARPER F M, KONSTAN J A. The MovieLens datasets: history and context[J]. ACM transactions on interactive intelligent systems, 2016, 5(4): 1-19
[49] NI Jianmo, LI Jiacheng, MCAULEY J. Justifying recommendations using distantly-labeled reviews and fine-grained aspects[C]//Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing. Hong Kong: Association for Computational Linguistics, 2019: 188-197.
[50] J?RVELIN K, KEK?L?INEN J. Cumulated gain-based evaluation of IR techniques[J]. ACM transactions on information systems, 2002, 20(4): 422-446
[51] BUCKLEY C, VOORHEES E M. Retrieval evaluation with incomplete information[C]//Proceedings of the 27th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval. Sheffield: ACM, 2004: 25-32.
[52] RENDLE S. Factorization machines[C]//2010 IEEE International Conference on Data Mining. Sydney: IEEE, 2011: 995-1000.
[53] MAO Kelong, ZHU Jieming, WANG Jinpeng, et al. SimpleX: a simple and strong baseline for collaborative filtering[C]//Proceedings of the 30th ACM International Conference on Information & Knowledge Management. Virtual Event: ACM, 2021: 1243-1252.
[54] HE Xiangnan, DENG Kuan, WANG Xiang, et al. LightGCN: simplifying and powering graph convolution network for recommendation[C]//Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. Virtual Event: ACM, 2020: 639-648.
[55] YU Junliang, XIA Xin, CHEN Tong, et al. XSimGCL: towards extremely simple graph contrastive learning for recommendation[J]. IEEE transactions on knowledge and data engineering, 2024, 36(2): 913-926
[56] CAI Xuheng, HUANG Chao, XIA Lianghao, et al. LightGCL: simple yet effective graph contrastive learning for recommendation[EB/OL]. (2023-02-16)[2025-06-25]. https://arxiv.org/abs/2302.08191.
- 相似文献/References:
-
[1]李绍滋,曹 阳,周昌乐.基于非结构化的P2P信息检索关键技术研究[J].智能系统学报,2006,1(2):74.
LI Shao-zi,CAO Yang,ZHOU Chang-le.Research on key techniques about unstructured P2P information retrieval[J].CAAI Transactions on Intelligent Systems,2006,1():74.
[2]黄立威,李德毅.社交媒体中的信息推荐[J].智能系统学报,2012,7(1):1.
HUANG Liwei,LI Deyi.A review of information recommendation in social media[J].CAAI Transactions on Intelligent Systems,2012,7():1.
[3]王超,刘奕群,马少平.搜索引擎点击模型综述[J].智能系统学报,2016,11(6):711.[doi:10.11992/tis.201605023]
WANG Chao,LIU Yiqun,MA Shaoping.A survey of click models for Web browsing[J].CAAI Transactions on Intelligent Systems,2016,11():711.[doi:10.11992/tis.201605023]
[4]郭少成,陈松灿.稀疏化的因子分解机[J].智能系统学报,2017,12(6):816.[doi:10.11992/tis.201706030]
GUO Shaocheng,CHEN Songcan.Sparsified factorization machine[J].CAAI Transactions on Intelligent Systems,2017,12():816.[doi:10.11992/tis.201706030]
[5]马钰,张岩,王宏志,等.面对智能导诊的个性化推荐算法[J].智能系统学报,2018,13(3):352.[doi:10.11992/tis.201711036]
MA Yu,ZHANG Yan,WANG Hongzhi,et al.A personalized recommendation algorithm for intelligent guidance[J].CAAI Transactions on Intelligent Systems,2018,13():352.[doi:10.11992/tis.201711036]
[6]常亮,张伟涛,古天龙,等.知识图谱的推荐系统综述[J].智能系统学报,2019,14(2):207.[doi:10.11992/tis.201805001]
CHANG Liang,ZHANG Weitao,GU Tianlong,et al.Review of recommendation systems based on knowledge graph[J].CAAI Transactions on Intelligent Systems,2019,14():207.[doi:10.11992/tis.201805001]
[7]吴国栋,查志康,涂立静,等.图神经网络推荐研究进展[J].智能系统学报,2020,15(1):14.[doi:10.11992/tis.201908034]
WU Guodong,ZHA Zhikang,TU Lijing,et al.Research advances in graph neural network recommendation[J].CAAI Transactions on Intelligent Systems,2020,15():14.[doi:10.11992/tis.201908034]
[8]梁丽君,李业刚,张娜娜,等.融合用户特征优化聚类的协同过滤算法[J].智能系统学报,2020,15(6):1091.[doi:10.11992/tis.201710024]
LIANG Lijun,LI Yegang,ZHANG Nana,et al.Collaborative filtering algorithm combining user features and preferences in optimized clustering[J].CAAI Transactions on Intelligent Systems,2020,15():1091.[doi:10.11992/tis.201710024]
[9]王健宗,肖京,朱星华,等.联邦推荐系统的协同过滤冷启动解决方法[J].智能系统学报,2021,16(1):178.[doi:10.11992/tis.202009032]
WANG Jianzong,XIAO Jing,ZHU Xinghua,et al.Cold starts in collaborative filtering for federated recommender systems[J].CAAI Transactions on Intelligent Systems,2021,16():178.[doi:10.11992/tis.202009032]
[10]鲍维克,袁春.面向推荐系统的分期序列自注意力网络[J].智能系统学报,2021,16(2):353.[doi:10.11992/tis.202005028]
BAO Weike,YUAN Chun.Recommendation system with long-term and short-term sequential self-attention network[J].CAAI Transactions on Intelligent Systems,2021,16():353.[doi:10.11992/tis.202005028]
备注/Memo
收稿日期:2025-6-25。
作者简介:郭世圆，硕士研究生，主要研究方向为推荐与用户建模，获国家发明专利授权2项，申请软件著作权1项，发表学术论文2篇。E-mail：gsy22@tsinghua.org.cn。;孙培杰，副教授，主要研究方向为推荐算法、用户行为建模。获2024年度“钱伟长中文信息处理科学技术奖”一等奖、自然语言处理实证方法会议（EMNLP）资源奖（Resource Award），获安徽省2023年度优秀博士论文，发表学术论文31篇。E-mail：peijiesun@ njupt.edu.cn。;张敏，教授，国际计算机学会出版委员会（ACM Publications Board）理事，国际计算机学会信息系统期刊（ACM TOIS）首位亚洲主编，担任国际计算机学会2026年信息检索大会（SIGIR）、2025年多媒体大会（MM）、2024年人机信息交互或检索会议（CHIIR）等会议主席或程序主席，主要研究方向为网络检索、个性化推荐与用户建模。曾获评IBM全球杰出学者奖（Global Faculty Award）、中国高校计算机专业优秀教师奖励计划、信息检索大会（SIGIR）时间检验奖、北京科学技术奖一等奖、“钱伟长中文信息处理科学技术奖”一等奖等。发表学术论文300余篇。E-mail：z-m@tsinghua.edu.cn。
通讯作者:张敏. E-mail：z-m@tsinghua.edu.cn
更新日期/Last Update:
1900-01-01