Determinantal point processes for machine learning

TL;DR

Determinantal point processes (DPPs) are probabilistic models that efficiently capture diversity and repulsion, enabling various machine learning tasks such as search result diversification, summarization, and modeling non-overlapping objects.

Contribution

This paper introduces DPPs to the machine learning community, explaining their intuition, algorithms, and applications for modeling diversity and negative correlations.

Findings

DPPs enable efficient sampling and inference in diverse applications.

DPPs outperform traditional models in capturing negative correlations.

Real-world applications include search diversification and summarization.

Abstract

Determinantal point processes (DPPs) are elegant probabilistic models of repulsion that arise in quantum physics and random matrix theory. In contrast to traditional structured models like Markov random fields, which become intractable and hard to approximate in the presence of negative correlations, DPPs offer efficient and exact algorithms for sampling, marginalization, conditioning, and other inference tasks. We provide a gentle introduction to DPPs, focusing on the intuitions, algorithms, and extensions that are most relevant to the machine learning community, and show how DPPs can be applied to real-world applications like finding diverse sets of high-quality search results, building informative summaries by selecting diverse sentences from documents, modeling non-overlapping human poses in images or video, and automatically building timelines of important news stories.