Explainable Neural Attention Recommender Systems

Recommender systems, predictive models that provide lists of personalized suggestions, have become increasingly popular in many web-based businesses. By presenting potential items that may interest a user, these systems are able to better monetize and improve users’ satisfaction. In recent years, the most successful approaches rely on capturing what best define users and items in the form of latent vectors, a numeric representation that assumes all instances can be described by their respective affiliation towards a set of hidden features. However, recommendation methods based on latent features still face some realworld limitations. The data sparsity problem originates from the unprecedented variety of available items, making generated suggestions irrelevant to many users. Furthermore, many systems have been recently expected to accompany their suggestions with corresponding reasoning. Users who receive unjustified recommendations they do not agree with are susceptible to stop using the system or ignore its suggestions. In this work we investigate the current trends in the field of recommender systems and focus on two rising areas, deep recommendation and explainable recommender systems. First we present Textual and Contextual Embedding-based Neural Recommender (TCENR), a model that mitigates the data sparsity problem in the area of point-of-interest (POI) recommendation. This method employs different types of deep neural networks to learn varied perspectives of the same user-location interaction, using textual reviews, geographical data and social networks