Explainable Active Learning Strategy For Recommender Systems

Recommender System technologies are witnessing a revolutionary era these days. They are growing more and more important since they help in the discovery and promotion of products, books, news, music, movies, courses, restaurants, etc. One major problem of Recommender Systems is that the most accurate models tend to be black boxes, whose results cannot be explained. This may cause the user to lose trust in the model. Another issue is the cost of acquiring data from users. That is why designing interactive algorithms that can automatically choose only the most helpful training data, also known as active learning approaches, are very useful. The main research question that we try to answer in this thesis is: How to improve the trust of the user in the Recommender System? And how to answer this question using the minimum amount of data?
We Implemented two active learning strategies that improve accuracy and explainability and we delineate a combined strategy that can optimize both metrics at the same time.

Link to Poster