Adversarial Point-of-Interest Recommendation

Point-of-interest (POI) recommendation is essential to a variety of services for both users and business. An extensive number of models have been developed to improve the recommendation performance by exploiting various characteristics and relations among POIs (e.g., spatio-temporal, social, etc.). However, very few studies closely look into the underlying mechanism accounting for why users prefer certain POIs to others. In this work, we initiate the first attempt to learn the distribution of user latent preference by proposing an Adversarial POI Recommendation (APOIR) model, consisting of two major components: (1) the recommender (R) which suggests POIs based on the learned distribution by maximizing the probabilities that these POIs are predicted as unvisited and potentially interested; and (2) the discriminator (D) which distinguishes the recommended POIs from the true check-ins and provides gradients as the guidance to improve R in a rewarding framework. Two components are co-trained by playing a minimax game towards improving itself while pushing the other to the boundary. By further integrating geographical and social relations among POIs into the reward function as well as optimizing R in a reinforcement learning manner, APOIR obtains significant performance improvement in four standard metrics compared to the state of the art methods.

[1]  Chunyan Miao,et al.  Exploiting Geographical Neighborhood Characteristics for Location Recommendation , 2014, CIKM.

[2]  Deng Cai,et al.  What to Do Next: Modeling User Behaviors by Time-LSTM , 2017, IJCAI.

[3]  Chi-Yin Chow,et al.  iGSLR: personalized geo-social location recommendation: a kernel density estimation approach , 2013, SIGSPATIAL/GIS.

[4]  Lantao Yu,et al.  SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient , 2016, AAAI.

[5]  Yehuda Koren,et al.  Matrix Factorization Techniques for Recommender Systems , 2009, Computer.

[6]  Chi-Yin Chow,et al.  GeoSoCa: Exploiting Geographical, Social and Categorical Correlations for Point-of-Interest Recommendations , 2015, SIGIR.

[7]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[8]  Michael R. Lyu,et al.  Geo-Teaser: Geo-Temporal Sequential Embedding Rank for Point-of-interest Recommendation , 2016, WWW.

[9]  Qiang Gao,et al.  Trajectory-User Linking via Variational AutoEncoder , 2018, IJCAI.

[10]  Craig MacDonald,et al.  A Contextual Attention Recurrent Architecture for Context-Aware Venue Recommendation , 2018, SIGIR.

[11]  Qiang Gao,et al.  Trajectory-based social circle inference , 2018, SIGSPATIAL/GIS.

[12]  Xin Liu,et al.  Exploring the Context of Locations for Personalized Location Recommendations , 2016, IJCAI.

[13]  Hao Wang,et al.  Location recommendation in location-based social networks using user check-in data , 2013, SIGSPATIAL/GIS.

[14]  Hao Liu,et al.  Learning User's Intrinsic and Extrinsic Interests for Point-of-Interest Recommendation: A Unified Approach , 2017, IJCAI.

[15]  Peng Zhang,et al.  IRGAN: A Minimax Game for Unifying Generative and Discriminative Information Retrieval Models , 2017, SIGIR.

[16]  Min Yang,et al.  Leveraging Long and Short-Term Information in Content-Aware Movie Recommendation via Adversarial Training , 2017, IEEE Transactions on Cybernetics.

[17]  Xiaoli Li,et al.  Rank-GeoFM: A Ranking based Geographical Factorization Method for Point of Interest Recommendation , 2015, SIGIR.

[18]  Qiang Gao,et al.  Identifying Human Mobility via Trajectory Embeddings , 2017, IJCAI.

[19]  Lejian Liao,et al.  Category-aware Next Point-of-Interest Recommendation via Listwise Bayesian Personalized Ranking , 2017, IJCAI.

[20]  Jure Leskovec,et al.  Friendship and mobility: user movement in location-based social networks , 2011, KDD.

[21]  Yishay Mansour,et al.  Policy Gradient Methods for Reinforcement Learning with Function Approximation , 1999, NIPS.

[22]  Michael R. Lyu,et al.  Fused Matrix Factorization with Geographical and Social Influence in Location-Based Social Networks , 2012, AAAI.

[23]  Kunpeng Zhang,et al.  vec2Link: Unifying Heterogeneous Data for Social Link Prediction , 2018, CIKM.

[24]  Mao Ye,et al.  Exploiting geographical influence for collaborative point-of-interest recommendation , 2011, SIGIR.

[25]  Chi-Yin Chow,et al.  LORE: exploiting sequential influence for location recommendations , 2014, SIGSPATIAL/GIS.

[26]  Yifeng Zeng,et al.  Personalized Ranking Metric Embedding for Next New POI Recommendation , 2015, IJCAI.

[27]  Craig MacDonald,et al.  A Deep Recurrent Collaborative Filtering Framework for Venue Recommendation , 2017, CIKM.

[28]  Richang Hong,et al.  Point-of-Interest Recommendations: Learning Potential Check-ins from Friends , 2016, KDD.

[29]  Lars Schmidt-Thieme,et al.  BPR: Bayesian Personalized Ranking from Implicit Feedback , 2009, UAI.

[30]  Xing Xie,et al.  GeoMF: joint geographical modeling and matrix factorization for point-of-interest recommendation , 2014, KDD.

[31]  Qiang Gao,et al.  Predicting Human Mobility via Variational Attention , 2019, WWW.

[32]  Jeffrey Dean,et al.  Efficient Estimation of Word Representations in Vector Space , 2013, ICLR.

[33]  Gao Cong,et al.  A General Model for Out-of-town Region Recommendation , 2017, WWW.

[34]  Gao Cong,et al.  An Experimental Evaluation of Point-of-interest Recommendation in Location-based Social Networks , 2017, Proc. VLDB Endow..

[35]  Jiawei Han,et al.  Bridging Collaborative Filtering and Semi-Supervised Learning: A Neural Approach for POI Recommendation , 2017, KDD.

[36]  Ronald J. Williams,et al.  Simple Statistical Gradient-Following Algorithms for Connectionist Reinforcement Learning , 2004, Machine Learning.

[37]  Yoshua Bengio,et al.  Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling , 2014, ArXiv.