GAN for Recommendation System

Abstract In our day there are many approaches for recommendation generation, however several unsolved problems still exist. In our work we study the potential of Generative Adversarial Networks to generate relevant recommendations and compare the proposed approach with another recommendations system based on neural networks. We found that recommendation system based on GANs can successively compete with state-of-the-art network in the field of recommendation generation and suggest, that GAN recommendation system have high potential.