Large-Margin Determinantal Point Processes
This work addresses the need for more flexible and error-aware parameter estimation in DPPs, which are used for diverse subset selection in applications like summarization, offering incremental improvements over existing methods.
The paper tackles the problem of learning parameters for determinantal point processes (DPPs) from labeled training data by introducing a reparameterization with multiple kernel functions for enhanced flexibility and a large-margin estimation technique that explicitly models selection errors and allows customization of precision-recall trade-offs. Empirical results show effectiveness in document and video summarization applications.
Determinantal point processes (DPPs) offer a powerful approach to modeling diversity in many applications where the goal is to select a diverse subset. We study the problem of learning the parameters (the kernel matrix) of a DPP from labeled training data. We make two contributions. First, we show how to reparameterize a DPP's kernel matrix with multiple kernel functions, thus enhancing modeling flexibility. Second, we propose a novel parameter estimation technique based on the principle of large margin separation. In contrast to the state-of-the-art method of maximum likelihood estimation, our large-margin loss function explicitly models errors in selecting the target subsets, and it can be customized to trade off different types of errors (precision vs. recall). Extensive empirical studies validate our contributions, including applications on challenging document and video summarization, where flexibility in modeling the kernel matrix and balancing different errors is indispensable.