MARCO: A Cooperative Knowledge Transfer Framework for Personalized Cross-domain Recommendations
Authors:
- Lili XieLXLili Xie
- Yi ZhangYZYi Zhang
- Ruihong QiuRQRuihong Qiu
- Jiajun LiuJLJiajun Liu
- Sen WangSWSen Wang
Preprint
Abstract
Recommender systems frequently encounter data sparsity issues, particularly when addressing cold-start scenarios involving new users or items. Multi-source cross-domain recommendation (CDR) addresses these challenges by transferring valuable knowledge from multiple source domains to enhance recommendations in a target domain. However, existing reinforcement learning (RL)-based CDR methods typically rely on a single-agent framework, leading to negative transfer issues caused by inconsistent domain contributions and inherent distributional discrepancies among source domains. To overcome these limitations, MARCO, a Multi-Agent Reinforcement Learning-based Cross-Domain recommendation framework, is proposed. It leverages cooperative multi-agent reinforcement learning, where each agent is dedicated to estimating the contribution from an individual source domain, effectively managing credit assignment and mitigating negative transfer. In addition, an entropy-based action diversity penalty is introduced to enhance policy expressiveness and stabilize training by encouraging diverse agents' joint actions. Extensive experiments across four benchmark datasets demonstrate MARCO's superior performance over state-of-the-art methods, highlighting its robustness and strong generalization capabilities. The code is at https://github.com/xiewilliams/MARCO.Keywords
All Related Versions
This publication has 44 references indexed in Scilit:
- A Survey on Negative TransferIEEE/CAA Journal of Automatica Sinica, 2022
- Beyond Double Ascent via Recurrent Neural Tangent Kernel in Sequential Recommendation2022
- Pre-training Graph Neural Network for Cross Domain Recommendation2021
- Memory Augmented Multi-Instance Contrastive Predictive Coding for Sequential Recommendation2021
- A Unified Framework for Cross-Domain and Cross-System RecommendationsIEEE Transactions on Knowledge and Data Engineering, 2021
- Top-aware reinforcement learning based recommendationNeurocomputing, 2020
- Multi-Site User Behavior Modeling and Its Application in Video RecommendationIEEE Transactions on Knowledge and Data Engineering, 2019
- Making recommendations from multiple domains2013
- Personalized recommendation via cross-domain triadic factorization2013
- Transfer learning in heterogeneous collaborative filtering domainsArtificial Intelligence, 2013