Generative Adversarial Active Learning for Unsupervised Outlier Detection

TL;DR

This paper introduces a novel generative adversarial active learning framework for outlier detection, using multiple generators to produce informative potential outliers and outperform existing methods on diverse datasets.

Contribution

The paper proposes MO-GAAL, a multi-generator extension of SO-GAAL, to effectively generate reference distributions for outlier detection without prior information.

Findings

MO-GAAL outperforms state-of-the-art methods on synthetic and real datasets.

The approach is especially effective for datasets with diverse cluster types.

MO-GAAL handles high irrelevant variable ratios well.

Abstract

Outlier detection is an important topic in machine learning and has been used in a wide range of applications. In this paper, we approach outlier detection as a binary-classification issue by sampling potential outliers from a uniform reference distribution. However, due to the sparsity of data in high-dimensional space, a limited number of potential outliers may fail to provide sufficient information to assist the classifier in describing a boundary that can separate outliers from normal data effectively. To address this, we propose a novel Single-Objective Generative Adversarial Active Learning (SO-GAAL) method for outlier detection, which can directly generate informative potential outliers based on the mini-max game between a generator and a discriminator. Moreover, to prevent the generator from falling into the mode collapsing problem, the stop node of training should be determined when SO-GAAL is able to provide sufficient information. But without any prior information, it is extremely difficult for SO-GAAL. Therefore, we expand the network structure of SO-GAAL from a single generator to multiple generators with different objectives (MO-GAAL), which can generate a reasonable reference distribution for the whole dataset. We empirically compare the proposed approach with several state-of-the-art outlier detection methods on both synthetic and real-world datasets. The results show that MO-GAAL outperforms its competitors in the majority of cases, especially for datasets with various cluster types or high irrelevant variable ratio.