Disentangled Contrastive Collaborative Filtering

Abstract

Recent studies show that graph neural networks (GNNs) are prevalent to model high-order relationships for collaborative filtering (CF). Towards this research line, graph contrastive learning (GCL) has exhibited powerful performance in addressing the supervision label shortage issue by learning augmented user and item representations. While many of them show their effectiveness, two key questions still remain unexplored: i) Most existing GCL-based CF models are still limited by ignoring the fact that user-item interaction behaviors are often driven by diverse latent intent factors (e.g., shopping for family party, preferred color or brand of products); ii) Their introduced non-adaptive augmentation techniques are vulnerable to noisy information, which raises concerns about the model's robustness and the risk of incorporating misleading self-supervised signals. In light of these limitations, we propose a Disentangled Contrastive Collaborative Filtering framework (DCCF) to realize intent disentanglement with self-supervised augmentation in an adaptive fashion. With the learned disentangled representations with global context, our DCCF is able to not only distill finer-grained latent factors from the entangled self-supervision signals but also alleviate the augmentation-induced noise. Finally, the cross-view contrastive learning task is introduced to enable adaptive augmentation with our parameterized interaction mask generator. Experiments on various public datasets demonstrate the superiority of our method compared to existing solutions. Our model implementation is released at the link https://github.com/HKUDS/DCCF.

Published as a SIGIR'23 full paper

References in corpus (24)

• Adam: A Method for Stochastic Optimization
• Neural Graph Collaborative Filtering
• KGAT: Knowledge Graph Attention Network for Recommendation
• Self-supervised Graph Learning for Recommendation
• Disentangled Graph Collaborative Filtering
• Learning Intents behind Interactions with Knowledge Graph for Recommendation
• Improving Graph Collaborative Filtering with Neighborhood-enriched Contrastive Learning
• Session-based Social Recommendation via Dynamic Graph Attention Networks
• Hypergraph Contrastive Collaborative Filtering
• LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation
• Intent Contrastive Learning for Sequential Recommendation
• Heterogeneous Graph Contrastive Learning for Recommendation
• Iterative Deep Graph Learning for Graph Neural Networks: Better and Robust Node Embeddings
• Contrastive Meta Learning with Behavior Multiplicity for Recommendation
• Multi-Behavior Hypergraph-Enhanced Transformer for Sequential Recommendation
• DisenHAN: Disentangled Heterogeneous Graph Attention Network for Recommendation
• Deconfounded Recommendation for Alleviating Bias Amplification
• Automated Self-Supervised Learning for Recommendation
• Graph Neural Networks in Recommender Systems: A Survey
• Multi-view Intent Disentangle Graph Networks for Bundle Recommendation
• LightGCL: Simple Yet Effective Graph Contrastive Learning for Recommendation
• Improving Knowledge-aware Recommendation with Multi-level Interactive Contrastive Learning
• Bounds on the Jensen Gap, and Implications for Mean-Concentrated Distributions
• Mutually-Regularized Dual Collaborative Variational Auto-encoder for Recommendation Systems

Cited by in corpus (14)

• Representation Learning with Large Language Models for Recommendation
• A Survey of Large Language Models for Graphs
• SSLRec: A Self-Supervised Learning Framework for Recommendation
• A Comprehensive Survey on Self-Supervised Learning for Recommendation
• ES-GNN: Generalizing Graph Neural Networks Beyond Homophily with Edge Splitting
• Path-Specific Counterfactual Fairness for Recommender Systems
• MixRec: Individual and Collective Mixing Empowers Data Augmentation for Recommender Systems
• Meta-GPS++: Enhancing Graph Meta-Learning with Contrastive Learning and Self-Training
• Intent Propagation Contrastive Collaborative Filtering
• A Worrying Reproducibility Study of Intent-Aware Recommendation Models
• Unified Dual-Intent Translation for Joint Modeling of Search and Recommendation
• Lightweight and Direct Document Relevance Optimization for Generative Information Retrieval
• Reproducibility and Artifact Consistency of the SIGIR 2022 Recommender Systems Papers Based on Message Passing
• Benchmarking Recommendation, Classification, and Tracing Based on Hugging Face Knowledge Graph