Adversarial Personalized Ranking for Recommendation

Abstract

Item recommendation is a personalized ranking task. To this end, many recommender systems optimize models with pairwise ranking objectives, such as the Bayesian Personalized Ranking (BPR). Using matrix Factorization (MF) --- the most widely used model in recommendation --- as a demonstration, we show that optimizing it with BPR leads to a recommender model that is not robust. In particular, we find that the resultant model is highly vulnerable to adversarial perturbations on its model parameters, which implies the possibly large error in generalization. To enhance the robustness of a recommender model and thus improve its generalization performance, we propose a new optimization framework, namely Adversarial Personalized Ranking (APR). In short, our APR enhances the pairwise ranking method BPR by performing adversarial training. It can be interpreted as playing a minimax game, where the minimization of the BPR objective function meanwhile defends an adversary, which adds adversarial perturbations on model parameters to maximize the BPR objective function. To illustrate how it works, we implement APR on MF by adding adversarial perturbations on the embedding vectors of users and items. Extensive experiments on three public real-world datasets demonstrate the effectiveness of APR --- by optimizing MF with APR, it outperforms BPR with a relative improvement of 11.2% on average and achieves state-of-the-art performance for item recommendation. Our implementation is available at: https://github.com/hexiangnan/adversarial_personalized_ranking.

SIGIR 2018

References in corpus (7)

• Explaining and Harnessing Adversarial Examples
• Intriguing properties of neural networks
• BPR: Bayesian Personalized Ranking from Implicit Feedback
• IRGAN: A Minimax Game for Unifying Generative and Discriminative Information Retrieval Models
• Item Silk Road: Recommending Items from Information Domains to Social Users
• Adversarial Training Methods for Semi-Supervised Text Classification
• Aesthetic-based Clothing Recommendation

Cited by in corpus (59)

• Neural Graph Collaborative Filtering
• Deep Learning based Recommender System: A Survey and New Perspectives
• Unifying Knowledge Graph Learning and Recommendation: Towards a Better Understanding of User Preferences
• NAIS: Neural Attentive Item Similarity Model for Recommendation
• Large Language Models as Zero-Shot Conversational Recommenders
• Session-based Recommendation with Hypergraph Attention Networks
• Adversarial Sampling and Training for Semi-Supervised Information Retrieval
• Explainable Reasoning over Knowledge Graphs for Recommendation
• Generating Reliable Friends via Adversarial Training to Improve Social Recommendation
• Deep Learning for Sequential Recommendation: Algorithms, Influential Factors, and Evaluations
• A Survey of Adversarial Learning on Graphs
• Graph Adversarial Training: Dynamically Regularizing Based on Graph Structure
• A survey on Adversarial Recommender Systems: from Attack/Defense strategies to Generative Adversarial Networks
• LambdaOpt: Learn to Regularize Recommender Models in Finer Levels
• DaisyRec 2.0: Benchmarking Recommendation for Rigorous Evaluation
• Relational Collaborative Filtering:Modeling Multiple Item Relations for Recommendation
• Lagrangian Inference for Ranking Problems
• Learning Explicit User Interest Boundary for Recommendation
• Enhancing Social Recommendation with Adversarial Graph Convolutional Networks
• RepeatNet: A Repeat Aware Neural Recommendation Machine for Session-based Recommendation
• Content-Based Collaborative Generation for Recommender Systems
• Boosting Item-based Collaborative Filtering via Nearly Uncoupled Random Walks
• Learning to Hash with Graph Neural Networks for Recommender Systems
• Subverting Fair Image Search with Generative Adversarial Perturbations
• HAM: Hybrid Associations Models for Sequential Recommendation
• From Intrinsic to Counterfactual: On the Explainability of Contextualized Recommender Systems
• Adversarial Training Towards Robust Multimedia Recommender System
• Fast Matrix Factorization with Non-Uniform Weights on Missing Data
• A Generic Network Compression Framework for Sequential Recommender Systems
• Future Data Helps Training: Modeling Future Contexts for Session-based Recommendation
• Hierarchical Attention Network for Visually-aware Food Recommendation
• Adversarial Collaborative Filtering for Free
• Adversarial Attack and Defense in Deep Ranking
• Recommender Systems Based on Generative Adversarial Networks: A Problem-Driven Perspective
• Assessing the Impact of a User-Item Collaborative Attack on Class of Users
• MAWSEO: Adversarial Wiki Search Poisoning for Illicit Online Promotion
• Automatic Feature Fairness in Recommendation via Adversaries
• Domain-to-Domain Translation Model for Recommender System
• Adversarial and Contrastive Variational Autoencoder for Sequential Recommendation
• Towards Amortized Ranking-Critical Training for Collaborative Filtering
• Attentive Representation Learning with Adversarial Training for Short Text Clustering
• Path-Based Reasoning over Heterogeneous Networks for Recommendation via Bidirectional Modeling
• Top-N Recommendation with Counterfactual User Preference Simulation
• Multi-Step Adversarial Perturbations on Recommender Systems Embeddings
• Fairness-aware Personalized Ranking Recommendation via Adversarial Learning
• Personalized News Recommendation with Knowledge-aware Interactive Matching
• An Empirical Study of DNNs Robustification Inefficacy in Protecting Visual Recommenders
• Personalized Hashtag Recommendation for Micro-videos
• Evaluating a Generative Adversarial Framework for Information Retrieval
• Extraction of Hierarchical Functional Connectivity Components in human brain using Adversarial Learning
• Multi-Facet Recommender Networks with Spherical Optimization
• Securing Visually-Aware Recommender Systems: An Adversarial Image Reconstruction and Detection Framework
• Adversarial Training Methods for Network Embedding
• Learning Robust Hierarchical Patterns of Human Brain across Many fMRI Studies
• Understanding the Effects of Adversarial Personalized Ranking Optimization Method on Recommendation Quality
• PT-Ranking: A Benchmarking Platform for Neural Learning-to-Rank
• Ready for Emerging Threats to Recommender Systems? A Graph Convolution-based Generative Shilling Attack
• Dual Adversarial Variational Embedding for Robust Recommendation
• Adversarial Ranking Attack and Defense