TSI-GAN: Unsupervised Time Series Anomaly Detection using Convolutional Cycle-Consistent Generative Adversarial Networks

TL;DR

TSI-GAN is an unsupervised time series anomaly detection model that converts data into images, uses cycle-consistent GANs for learning complex patterns, and outperforms existing methods on diverse datasets.

Contribution

The paper introduces TSI-GAN, a novel unsupervised model that automatically learns complex temporal patterns without dataset-specific tuning or assumptions.

Findings

Outperforms 8 state-of-the-art baselines by 13-31%

Uses image encoding to capture temporal deviance

Employs cycle-consistency loss for accurate inverse mappings

Abstract

Anomaly detection is widely used in network intrusion detection, autonomous driving, medical diagnosis, credit card frauds, etc. However, several key challenges remain open, such as lack of ground truth labels, presence of complex temporal patterns, and generalizing over different datasets. This paper proposes TSI-GAN, an unsupervised anomaly detection model for time-series that can learn complex temporal patterns automatically and generalize well, i.e., no need for choosing dataset-specific parameters, making statistical assumptions about underlying data, or changing model architectures. To achieve these goals, we convert each input time-series into a sequence of 2D images using two encoding techniques with the intent of capturing temporal patterns and various types of deviance. Moreover, we design a reconstructive GAN that uses convolutional layers in an encoder-decoder network and employs cycle-consistency loss during training to ensure that inverse mappings are accurate as well. In addition, we also instrument a Hodrick-Prescott filter in post-processing to mitigate false positives. We evaluate TSI-GAN using 250 well-curated and harder-than-usual datasets and compare with 8 state-of-the-art baseline methods. The results demonstrate the superiority of TSI-GAN to all the baselines, offering an overall performance improvement of 13% and 31% over the second-best performer MERLIN and the third-best performer LSTM-AE, respectively.