Curriculum-scheduled knowledge distillation from multiple pre-trained teachers for multi-domain sequential recommendation
World Wide Web · 2025-11 · 28 (6)
DOI: 10.1007/s11280-025-01381-9 ↗Kurz gefasst
Fachartikel in der Zeitschrift World Wide Web, online erschienen am 12.11.2025. Verfasst von Wenqi Sun und 5 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 47 Literaturangaben. Gefördert von: Young Elite Scientists Sponsorship Program by China Association for Science and Technology, National Natural Science Foundation of China.
Abstract (Originalsprache)
Pre-trained recommendation models (PRMs) have received increasing interest recently. However, their intrinsically heterogeneous model structure, huge model size and computation cost hinder their adoptions in practical recommender systems. Hence, it is highly essential to explore how to use different pre-trained recommendation models efficiently in real-world systems. In this paper, we propose a novel curriculum-scheduled knowledge distillation from multiple pre-trained teachers for multi-domain sequential recommendation, called CKD-MDSR, which takes full advantages of different PRMs as multiple teacher models to boost a small student recommendation model, integrating the knowledge across multiple domains from PRMs. Specifically, CKD-MDSR first adopts curriculum-scheduled user behavior sequence sampling and distills informative knowledge jointly from the representative PRMs such as UniSRec and Recformer. Then, the knowledge from the above PRMs are selectively integrated into the student model in consideration of their confidence and consistency. Finally, we verify the proposed method on multi-domain sequential recommendation and further demonstrate its universality with multiple types of student models, including feature interaction and graph based recommendation models. Extensive experiments on five real-world datasets demonstrate the effectiveness and efficiency of CKD-MDSR, which can be viewed as an efficient shortcut using PRMs in real-world systems. Our code is available at https://github.com/RUCAIBox/CKD-MDSR.
Förderung
- Young Elite Scientists Sponsorship Program by China Association for Science and Technology
- National Natural Science Foundation of China
Zitieren
APA
Sun, W., Xie, R., Zhang, J., Zhao, W. X., Lin, L., & Wen, J. (2025). Curriculum-scheduled knowledge distillation from multiple pre-trained teachers for multi-domain sequential recommendation. World Wide Web, 28(6). https://doi.org/10.1007/s11280-025-01381-9
Vancouver
Sun W, Xie R, Zhang J, Zhao WX, Lin L, Wen J. Curriculum-scheduled knowledge distillation from multiple pre-trained teachers for multi-domain sequential recommendation. World Wide Web. 2025;28(6). doi:10.1007/s11280-025-01381-9
BibTeX
@article{Sun2025,
author = {Sun, Wenqi and Xie, Ruobing and Zhang, Junjie and Zhao, Wayne Xin and Lin, Leyu and Wen, Ji-Rong},
title = {Curriculum-scheduled knowledge distillation from multiple pre-trained teachers for multi-domain sequential recommendation},
journal = {World Wide Web},
year = {2025},
volume = {28},
number = {6},
doi = {10.1007/s11280-025-01381-9}
}
EUVIMED