Low-Rank Factorization of Determinantal Point Processes for Recommendation
This work addresses efficiency bottlenecks in DPP-based recommendation systems, offering significant speed improvements for practical applications, though it is incremental as it builds on existing DPP frameworks.
The paper tackles the problem of learning Determinantal Point Processes (DPPs) for product recommendation by introducing a low-rank factorization of the kernel, resulting in a learning algorithm that is nearly an order of magnitude faster and prediction computations up to 20x faster for large catalogs, while maintaining equivalent or better predictive performance compared to prior methods.
Determinantal point processes (DPPs) have garnered attention as an elegant probabilistic model of set diversity. They are useful for a number of subset selection tasks, including product recommendation. DPPs are parametrized by a positive semi-definite kernel matrix. In this work we present a new method for learning the DPP kernel from observed data using a low-rank factorization of this kernel. We show that this low-rank factorization enables a learning algorithm that is nearly an order of magnitude faster than previous approaches, while also providing for a method for computing product recommendation predictions that is far faster (up to 20x faster or more for large item catalogs) than previous techniques that involve a full-rank DPP kernel. Furthermore, we show that our method provides equivalent or sometimes better predictive performance than prior full-rank DPP approaches, and better performance than several other competing recommendation methods in many cases. We conduct an extensive experimental evaluation using several real-world datasets in the domain of product recommendation to demonstrate the utility of our method, along with its limitations.