Learning Determinantal Point Processes

Abstract

Determinantal point processes (DPPs), which arise in random matrix theory and quantum physics, are natural models for subset selection problems where diversity is preferred. Among many remarkable properties, DPPs offer tractable algorithms for exact inference, including computing marginal probabilities and sampling; however, an important open question has been how to learn a DPP from labeled training data. In this paper we propose a natural feature-based parameterization of conditional DPPs, and show how it leads to a convex and efficient learning formulation. We analyze the relationship between our model and binary Markov random fields with repulsive potentials, which are qualitatively similar but computationally intractable. Finally, we apply our approach to the task of extractive summarization, where the goal is to choose a small subset of sentences conveying the most important information from a set of documents. In this task there is a fundamental tradeoff between sentences that are highly relevant to the collection as a whole, and sentences that are diverse and not repetitive. Our parameterization allows us to naturally balance these two characteristics. We evaluate our system on data from the DUC 2003/04 multi-document summarization task, achieving state-of-the-art results.

Cited by in corpus (13)

• Learning Mixtures of Submodular Shells with Application to Document Summarization
• Expectation-Maximization for Learning Determinantal Point Processes
• Large-Margin Determinantal Point Processes
• What comes next? Extractive summarization by next-sentence prediction
• Scoring Sentence Singletons and Pairs for Abstractive Summarization
• Improving the Similarity Measure of Determinantal Point Processes for Extractive Multi-Document Summarization
• Composable Core-sets for Determinant Maximization: A Simple Near-Optimal Algorithm
• Detecting (Un)Important Content for Single-Document News Summarization
• Multi-Document Summarization with Determinantal Point Processes and Contextualized Representations
• Graphical structure of conditional independencies in determinantal point processes
• Video Summarization via Actionness Ranking
• Towards Bursting Filter Bubble via Contextual Risks and Uncertainties
• The Bayesian Low-Rank Determinantal Point Process Mixture Model