Scalable modeling improves diverse selections from large datasets
Scalable Learning of Determinantal Point Processes with Kernel-Structured Interactions
Machine Learning
Summary
Choosing diverse groups of items from large collections is hard because items might be similar and overlap. The authors offer a new way to model this diversity using a mathematical tool called a determinantal point process, improving how similarities between items are understood. They make this method faster and better by using clever approximations and random features. Their technique works well on real data and uses fewer parameters.
What this means in practice
- •For recommendation system developers: Improve diverse item recommendations by efficiently modeling full interactions among large item sets.
- •For data engineers: Scale diversity-aware subset selection in large datasets with fewer model parameters and faster learning.
Authors
Takahiro Kawashima, Hideitsu Hino
Abstract
Determinantal point processes (DPPs) on a finite set are probability models that exhibit repulsion where inter-item similarities are encoded by a positive definite kernel matrix. Scaling DPP learning to large datasets typically requires imposing structural constraints on the kernel matrix, such as a low-rank-plus-diagonal form. However, structural assumptions introduce a severe trade-off between expressive power and computational scalability. Our study addresses this challenge by (a) modeling the full-rank interactions via a nonlinear kernel function and latent features, (b) designing a surrogate function for the maximum likelihood estimator, and (c) adopting stochastic optimization using random features of the nonlinear kernel. Numerical experiments on real-world datasets demonstrate that our method significantly improves this trade-off, achieving better generalization performance with remarkable parameter efficiency.