Online MAP Inference and Learning for Nonsymmetric Determinantal Point Processes

TL;DR

This paper develops online algorithms for MAP inference and learning in Non-symmetric Determinantal Point Processes, enabling real-time processing of streaming data with theoretical guarantees and competitive performance.

Contribution

It introduces the first online algorithms for MAP inference and learning in NDPPs, suitable for streaming data with single-pass and limited memory constraints.

Findings

Algorithms have theoretical guarantees.

Comparable performance to offline methods.

Effective on real-world datasets.

Abstract

In this paper, we introduce the online and streaming MAP inference and learning problems for Non-symmetric Determinantal Point Processes (NDPPs) where data points arrive in an arbitrary order and the algorithms are constrained to use a single-pass over the data as well as sub-linear memory. The online setting has an additional requirement of maintaining a valid solution at any point in time. For solving these new problems, we propose algorithms with theoretical guarantees, evaluate them on several real-world datasets, and show that they give comparable performance to state-of-the-art offline algorithms that store the entire data in memory and take multiple passes over it.