Online Data Thinning via Multi-Subspace Tracking

TL;DR

This paper introduces an online data thinning method that uses dynamic low-rank Gaussian mixture models to efficiently identify and preserve salient or anomalous elements in large-scale streaming high-dimensional data in real time.

Contribution

It presents a novel online anomaly detection approach based on low-rank Gaussian mixtures that adapt to dynamic environments and handle missing data, enabling scalable real-time data thinning.

Findings

Effective in real-time anomaly detection

Robust to missing data and high dimensionality

Scalable and efficient for large-scale streaming data

Abstract

In an era of ubiquitous large-scale streaming data, the availability of data far exceeds the capacity of expert human analysts. In many settings, such data is either discarded or stored unprocessed in datacenters. This paper proposes a method of online data thinning, in which large-scale streaming datasets are winnowed to preserve unique, anomalous, or salient elements for timely expert analysis. At the heart of this proposed approach is an online anomaly detection method based on dynamic, low-rank Gaussian mixture models. Specifically, the high-dimensional covariances matrices associated with the Gaussian components are associated with low-rank models. According to this model, most observations lie near a union of subspaces. The low-rank modeling mitigates the curse of dimensionality associated with anomaly detection for high-dimensional data, and recent advances in subspace clustering and subspace tracking allow the proposed method to adapt to dynamic environments. Furthermore, the proposed method allows subsampling, is robust to missing data, and uses a mini-batch online optimization approach. The resulting algorithms are scalable, efficient, and are capable of operating in real time. Experiments on wide-area motion imagery and e-mail databases illustrate the efficacy of the proposed approach.