Streaming Inference for Infinite Non-Stationary Clustering

TL;DR

This paper introduces a novel probabilistic clustering algorithm for streaming, non-stationary data that dynamically creates new clusters and employs a new stochastic process called the Dynamical CRP, enabling efficient inference.

Contribution

The paper presents the Dynamical CRP, a new non-exchangeable stochastic process, and a streaming variational inference algorithm for non-stationary clustering.

Findings

Effective on synthetic and real data

Handles Gaussian and non-Gaussian likelihoods

Creates clusters online in a principled manner

Abstract

Learning from a continuous stream of non-stationary data in an unsupervised manner is arguably one of the most common and most challenging settings facing intelligent agents. Here, we attack learning under all three conditions (unsupervised, streaming, non-stationary) in the context of clustering, also known as mixture modeling. We introduce a novel clustering algorithm that endows mixture models with the ability to create new clusters online, as demanded by the data, in a probabilistic, time-varying, and principled manner. To achieve this, we first define a novel stochastic process called the Dynamical Chinese Restaurant Process (Dynamical CRP), which is a non-exchangeable distribution over partitions of a set; next, we show that the Dynamical CRP provides a non-stationary prior over cluster assignments and yields an efficient streaming variational inference algorithm. We conclude with experiments showing that the Dynamical CRP can be applied on diverse synthetic and real data with Gaussian and non-Gaussian likelihoods.