Saved in:
Bibliographic Details
Main Authors: Wang, Xinfeng, Fukumoto, Fumiyo, Cui, Jin, Suzuki, Yoshimi, Li, Jiyi, Yu, Dongjin
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2404.06895
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909303846207488
author Wang, Xinfeng
Fukumoto, Fumiyo
Cui, Jin
Suzuki, Yoshimi
Li, Jiyi
Yu, Dongjin
author_facet Wang, Xinfeng
Fukumoto, Fumiyo
Cui, Jin
Suzuki, Yoshimi
Li, Jiyi
Yu, Dongjin
contents Recommender models aimed at mining users' behavioral patterns have raised great attention as one of the essential applications in daily life. Recent work on graph neural networks (GNNs) or debiasing methods has attained remarkable gains. However, they still suffer from (1) over-smoothing node embeddings caused by recursive convolutions with GNNs, and (2) the skewed distribution of interactions due to popularity and user-individual biases. This paper proposes a contextualized and debiased recommender model (CaDRec). To overcome the over-smoothing issue, we explore a novel hypergraph convolution operator that can select effective neighbors during convolution by introducing both structural context and sequential context. To tackle the skewed distribution, we propose two strategies for disentangling interactions: (1) modeling individual biases to learn unbiased item embeddings, and (2) incorporating item popularity with positional encoding. Moreover, we mathematically show that the imbalance of the gradients to update item embeddings exacerbates the popularity bias, thus adopting regularization and weighting schemes as solutions. Extensive experiments on four datasets demonstrate the superiority of the CaDRec against state-of-the-art (SOTA) methods. Our source code and data are released at https://github.com/WangXFng/CaDRec.
format Preprint
id arxiv_https___arxiv_org_abs_2404_06895
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle CaDRec: Contextualized and Debiased Recommender Model
Wang, Xinfeng
Fukumoto, Fumiyo
Cui, Jin
Suzuki, Yoshimi
Li, Jiyi
Yu, Dongjin
Information Retrieval
Recommender models aimed at mining users' behavioral patterns have raised great attention as one of the essential applications in daily life. Recent work on graph neural networks (GNNs) or debiasing methods has attained remarkable gains. However, they still suffer from (1) over-smoothing node embeddings caused by recursive convolutions with GNNs, and (2) the skewed distribution of interactions due to popularity and user-individual biases. This paper proposes a contextualized and debiased recommender model (CaDRec). To overcome the over-smoothing issue, we explore a novel hypergraph convolution operator that can select effective neighbors during convolution by introducing both structural context and sequential context. To tackle the skewed distribution, we propose two strategies for disentangling interactions: (1) modeling individual biases to learn unbiased item embeddings, and (2) incorporating item popularity with positional encoding. Moreover, we mathematically show that the imbalance of the gradients to update item embeddings exacerbates the popularity bias, thus adopting regularization and weighting schemes as solutions. Extensive experiments on four datasets demonstrate the superiority of the CaDRec against state-of-the-art (SOTA) methods. Our source code and data are released at https://github.com/WangXFng/CaDRec.
title CaDRec: Contextualized and Debiased Recommender Model
topic Information Retrieval
url https://arxiv.org/abs/2404.06895