State-of-Art Techniques for Deep Learning-Based Recommender Systems

The information present on the web is endless and it keeps growing exponentially day by day. Such scenarios greet the user with countless search results that are partially relevant and mostly irrelevant. Recommender systems act as a military weapon against the issue of overloaded information. The utility of recommender systems is adopted in several web applications along with its capability to mitigate problems arising due to a plethora of retrieved search results. The integration of machine learning and deep learning techniques with recommender systems produces optimized and contextually appropriate information for users within a specific domain. This paper provides a study of various research studies on recommender systems based on deep learning. To facilitate ongoing research in the area, the study also offers a taxonomy of deep-learning-based models.