{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/mad-gan-multivariate-anomaly-detection-for","title":"MAD-GAN: Multivariate Anomaly Detection for Time Series Data with Generative Adversarial Networks","arxiv_id":"1901.04997","date":"2019-01-15","proceeding":null,"authors":["Dan Li","Dacheng Chen","Lei Shi","Baihong Jin","Jonathan Goh","See-Kiong Ng"],"abstract":"The prevalence of networked sensors and actuators in many real-world systems\nsuch as smart buildings, factories, power plants, and data centers generate\nsubstantial amounts of multivariate time series data for these systems. The\nrich sensor data can be continuously monitored for intrusion events through\nanomaly detection. However, conventional threshold-based anomaly detection\nmethods are inadequate due to the dynamic complexities of these systems, while\nsupervised machine learning methods are unable to exploit the large amounts of\ndata due to the lack of labeled data. On the other hand, current unsupervised\nmachine learning approaches have not fully exploited the spatial-temporal\ncorrelation and other dependencies amongst the multiple variables\n(sensors/actuators) in the system for detecting anomalies. In this work, we\npropose an unsupervised multivariate anomaly detection method based on\nGenerative Adversarial Networks (GANs). Instead of treating each data stream\nindependently, our proposed MAD-GAN framework considers the entire variable set\nconcurrently to capture the latent interactions amongst the variables. We also\nfully exploit both the generator and discriminator produced by the GAN, using a\nnovel anomaly score called DR-score to detect anomalies by discrimination and\nreconstruction. We have tested our proposed MAD-GAN using two recent datasets\ncollected from real-world CPS: the Secure Water Treatment (SWaT) and the Water\nDistribution (WADI) datasets. Our experimental results showed that the proposed\nMAD-GAN is effective in reporting anomalies caused by various cyber-intrusions\ncompared in these complex real-world systems.","url_abs":"http://arxiv.org/abs/1901.04997v1","url_pdf":"http://arxiv.org/pdf/1901.04997v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"mad-gan-multivariate-anomaly-detection-for","repo_url":"https://github.com/LiDan456/MAD-GANs","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"mad-gan-multivariate-anomaly-detection-for","repo_url":"https://github.com/guillem96/madgan-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"anomaly-detection","task_name":"Anomaly Detection"},{"task_slug":"machine-learning","task_name":"BIG-bench Machine Learning"},{"task_slug":"time-series-1","task_name":"Time Series"},{"task_slug":"time-series","task_name":"Time Series Analysis"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1901.04997","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}