Click is Not Equal to Purchase: Multi-task Reinforcement Learning for Multi-behavior Recommendation

Huiwang Zhang, Pengpeng Zhao, Xuefeng Xian, Victor S. Sheng, Yongjing Hao, Zhiming Cui

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Reinforcement learning (RL) has achieved ideal performance in recommendation systems (RS) by taking care of both immediate and future rewards from users. However, the existing RL-based recommendation methods assume that only a single type of interaction behavior (e.g., clicking) exists between user and item, whereas practical recommendation scenarios involve multiple types of user interaction behaviors (e.g., adding to cart, purchasing). In this paper, we propose a Multi-Task Reinforcement Learning model for multi-behavior Recommendation (MTRL4Rec), which gives different actions for users’ different behaviors with a single agent. Specifically, we first introduce a modular network in which modules can be reused or isolated from each other to model the commonalities and differences between users’ behaviors. Then a task routing network is used to generate routes in the modular network for each behavior task. Finally, we adopt a hierarchical reinforcement learning architecture to improve the efficiency of MTRL4Rec. Experiments on two public datasets validated the effectiveness of MTRL4Rec.

Original languageEnglish
Title of host publicationWeb Information Systems Engineering – WISE 2022 - 23rd International Conference, Proceedings
EditorsRichard Chbeir, Helen Huang, Fabrizio Silvestri, Yannis Manolopoulos, Yanchun Zhang, Yanchun Zhang
PublisherSpringer Science and Business Media Deutschland GmbH
Pages443-459
Number of pages17
ISBN (Print)9783031208904
DOIs
StatePublished - 2022
Event23rd International Conference on Web Information Systems Engineering, WISE 2021 - Biarritz, France
Duration: Nov 1 2022Nov 3 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13724 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference23rd International Conference on Web Information Systems Engineering, WISE 2021
Country/TerritoryFrance
CityBiarritz
Period11/1/2211/3/22

Keywords

  • Multi-behavior
  • Multi-task
  • Recommendation system
  • Reinforcement learning

Fingerprint

Dive into the research topics of 'Click is Not Equal to Purchase: Multi-task Reinforcement Learning for Multi-behavior Recommendation'. Together they form a unique fingerprint.

Cite this