Optimized Item Selection to Boost Exploration for Recommender Systems

Serdar Kadioglu, Bernard Kleynhans, Xin Wang

Abstract: Recommender Systems have become the backbone of personalized services that provide tailored experiences to individual users. Still, data sparsity remains a common challenging problem, especially for new applications where training data is limited or not available. In this paper, we formalize a combinatorial problem that is concerned with selecting the universe of items for experimentation with recommender systems. On one hand, a large set of items is desirable to increase the diversity of items. On the other hand, a smaller set of items enable rapid experimentation and minimize the time and the amount of data required to train machine learning models. We show how to optimize for such conflicting criteria using a multi-level optimization framework. Our approach integrates techniques from discrete optimization, unsupervised clustering, and latent text embeddings. Experimental results on well-known movie and book recommendation benchmarks demonstrate the benefits of optimized item selection.