Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

End-to-end Learnable Clustering for Intent Learning in Recommendation

Yue Liu, Shihao Zhu, Jun Xia, Yingwei Ma, Jian Ma, Xinwang Liu, Shengju Yu, Kejun Zhang, Wenliang Zhong

TL;DR

This work proposes a novel intent learning method termed ELCRec, which improves NDCG@5 by 8.9\% and reduces computational costs by 22.5\% on the Beauty dataset and deploys this method on the industrial recommendation system with 130 million page views and achieves promising results.

Abstract

Intent learning, which aims to learn users' intents for user understanding and item recommendation, has become a hot research spot in recent years. However, existing methods suffer from complex and cumbersome alternating optimization, limiting performance and scalability. To this end, we propose a novel intent learning method termed \underline{ELCRec}, by unifying behavior representation learning into an \underline{E}nd-to-end \underline{L}earnable \underline{C}lustering framework, for effective and efficient \underline{Rec}ommendation. Concretely, we encode user behavior sequences and initialize the cluster centers (latent intents) as learnable neurons. Then, we design a novel learnable clustering module to separate different cluster centers, thus decoupling users' complex intents. Meanwhile, it guides the network to learn intents from behaviors by forcing behavior embeddings close to cluster centers. This allows simultaneous optimization of recommendation and clustering via mini-batch data. Moreover, we propose intent-assisted contrastive learning by using cluster centers as self-supervision signals, further enhancing mutual promotion. Both experimental results and theoretical analyses demonstrate the superiority of ELCRec from six perspectives. Compared to the runner-up, ELCRec improves NDCG@5 by 8.9\% and reduces computational costs by 22.5\% on the Beauty dataset. Furthermore, due to the scalability and universal applicability, we deploy this method on the industrial recommendation system with 130 million page views and achieve promising results. The codes are available on GitHub (https://github.com/yueliu1999/ELCRec). A collection (papers, codes, datasets) of deep group recommendation/intent learning methods is available on GitHub (https://github.com/yueliu1999/Awesome-Deep-Group-Recommendation).

End-to-end Learnable Clustering for Intent Learning in Recommendation

TL;DR

This work proposes a novel intent learning method termed ELCRec, which improves NDCG@5 by 8.9\% and reduces computational costs by 22.5\% on the Beauty dataset and deploys this method on the industrial recommendation system with 130 million page views and achieves promising results.

Abstract

Intent learning, which aims to learn users' intents for user understanding and item recommendation, has become a hot research spot in recent years. However, existing methods suffer from complex and cumbersome alternating optimization, limiting performance and scalability. To this end, we propose a novel intent learning method termed \underline{ELCRec}, by unifying behavior representation learning into an \underline{E}nd-to-end \underline{L}earnable \underline{C}lustering framework, for effective and efficient \underline{Rec}ommendation. Concretely, we encode user behavior sequences and initialize the cluster centers (latent intents) as learnable neurons. Then, we design a novel learnable clustering module to separate different cluster centers, thus decoupling users' complex intents. Meanwhile, it guides the network to learn intents from behaviors by forcing behavior embeddings close to cluster centers. This allows simultaneous optimization of recommendation and clustering via mini-batch data. Moreover, we propose intent-assisted contrastive learning by using cluster centers as self-supervision signals, further enhancing mutual promotion. Both experimental results and theoretical analyses demonstrate the superiority of ELCRec from six perspectives. Compared to the runner-up, ELCRec improves NDCG@5 by 8.9\% and reduces computational costs by 22.5\% on the Beauty dataset. Furthermore, due to the scalability and universal applicability, we deploy this method on the industrial recommendation system with 130 million page views and achieve promising results. The codes are available on GitHub (https://github.com/yueliu1999/ELCRec). A collection (papers, codes, datasets) of deep group recommendation/intent learning methods is available on GitHub (https://github.com/yueliu1999/Awesome-Deep-Group-Recommendation).
Paper Structure (49 sections, 8 equations, 6 figures, 12 tables, 1 algorithm)

This paper contains 49 sections, 8 equations, 6 figures, 12 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methodology
  4. Basic Notation
  5. Task Definition
  6. Problem Analyses
  7. Proposed Method
  8. Behavior Encoding
  9. End-to-end Learnable Cluster Module
  10. Intent-assisted Contrastive Learning
  11. Overall Objective
  12. Experiment
  13. Superiority
  14. Effectiveness
  15. Efficiency
  16. ...and 34 more sections

Figures (6)

  • Figure 1: Ablation studies of the proposed end-to-end learnable cluster module (ELCM) and the intent-assisted contrastive learning (ICL). The results are the sum of four metrics, including HR@5, HR@20, NDCG@5, and NDCG@20.
  • Figure 2: Sensitivity analyses of ELCRec. The results are evaluated by the HR@5 metric.
  • Figure 3: Convergence analyses. The first and second row denotes HR@5 on the evaluation set and training loss, respectively.
  • Figure 4: Case studies on different user groups. The distributions of different user groups are visualized. "top" denotes the cluster IDs, which have the highest proportion in the user group. "bottom" denotes the cluster IDs, which have the lowest proportion in the user group.
  • Figure 5: Case studies on the learned cluster. We visualize the distribution of the learned clusters. "HA", "MA", "LA", and "N" denote the high-activity, medium-activity, low-activity, and new user groups, respectively.
  • ...and 1 more figures