Towards Deterministic Diverse Subset Sampling

TL;DR

This paper introduces a deterministic greedy algorithm for k-DPPs, enhancing interpretability and reliability in diverse subset selection tasks such as image search and summarization.

Contribution

It proposes a novel deterministic adaptation of k-DPPs, providing consistent results and improved low-rank kernel approximations over traditional stochastic methods.

Findings

Achieves low-rank kernel approximations with high accuracy

Demonstrates effectiveness in image search applications

Provides deterministic, interpretable subset selection

Abstract

Determinantal point processes (DPPs) are well known models for diverse subset selection problems, including recommendation tasks, document summarization and image search. In this paper, we discuss a greedy deterministic adaptation of k-DPP. Deterministic algorithms are interesting for many applications, as they provide interpretability to the user by having no failure probability and always returning the same results. First, the ability of the method to yield low-rank approximations of kernel matrices is evaluated by comparing the accuracy of the Nystr\"om approximation on multiple datasets. Afterwards, we demonstrate the usefulness of the model on an image search task.