StackVAE-G: An efficient and interpretable model for time series anomaly detection
Recent studies have shown that autoencoder-based models can achieve superior performance on anomaly detection tasks due to their excellent ability to fit complex data in an unsupervised manner. In this work, we propose a novel autoencoder-based model, named StackVAE-G that can significantly bring th...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
KeAi Communications Co. Ltd.
2022-01-01
|
| Series: | AI Open |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2666651022000110 |
| _version_ | 1797977726856986624 |
|---|---|
| author | Wenkai Li Wenbo Hu Ting Chen Ning Chen Cheng Feng |
| author_facet | Wenkai Li Wenbo Hu Ting Chen Ning Chen Cheng Feng |
| author_sort | Wenkai Li |
| collection | DOAJ |
| description | Recent studies have shown that autoencoder-based models can achieve superior performance on anomaly detection tasks due to their excellent ability to fit complex data in an unsupervised manner. In this work, we propose a novel autoencoder-based model, named StackVAE-G that can significantly bring the efficiency and interpretability to multivariate time series anomaly detection. Specifically, we utilize the similarities across the time series channels by the stacking block-wise reconstruction with a weight-sharing scheme to reduce the size of learned models and also relieve the overfitting to unknown noises in the training data. We also leverage a graph learning module to learn a sparse adjacency matrix to explicitly capture the stable interrelation structure among multiple time series channels for the interpretable pattern reconstruction of interrelated channels. Combining these two modules, we introduce the stacking block-wise VAE (variational autoencoder) with GNN (graph neural network) model for multivariate time series anomaly detection. We conduct extensive experiments on three commonly used public datasets, showing that our model achieves comparable (even better) performance with the state-of-the-art models and meanwhile requires much less computation and memory cost. Furthermore, we demonstrate that the adjacency matrix learned by our model accurately captures the interrelation among multiple channels, and can provide valuable information for failure diagnosis applications. |
| first_indexed | 2024-04-11T05:11:34Z |
| format | Article |
| id | doaj.art-21dbb25eaba54978bad4ed8a98a65f49 |
| institution | Directory Open Access Journal |
| issn | 2666-6510 |
| language | English |
| last_indexed | 2024-04-11T05:11:34Z |
| publishDate | 2022-01-01 |
| publisher | KeAi Communications Co. Ltd. |
| record_format | Article |
| series | AI Open |
| spelling | doaj.art-21dbb25eaba54978bad4ed8a98a65f492022-12-25T04:19:28ZengKeAi Communications Co. Ltd.AI Open2666-65102022-01-013101110StackVAE-G: An efficient and interpretable model for time series anomaly detectionWenkai Li0Wenbo Hu1Ting Chen2Ning Chen3Cheng Feng4High Performance Computing Center, Department of Computer Science and Technology, BNRist Center, Institute for AI, Tsinghua-BOSCH Joint ML Center, THBI Lab, Tsinghua University, Beijing, China; THU-Siemens Joint Research Center for Industrial Intelligence and Internet of Things, Beijing, ChinaSchool of Computer and Information, Hefei University of Technology, Hefei, China; Corresponding authors.High Performance Computing Center, Department of Computer Science and Technology, BNRist Center, Institute for AI, Tsinghua-BOSCH Joint ML Center, THBI Lab, Tsinghua University, Beijing, China; Corresponding authors.High Performance Computing Center, Department of Computer Science and Technology, BNRist Center, Institute for AI, Tsinghua-BOSCH Joint ML Center, THBI Lab, Tsinghua University, Beijing, China; Corresponding authors.Siemens AG, Beijing, China; THU-Siemens Joint Research Center for Industrial Intelligence and Internet of Things, Beijing, ChinaRecent studies have shown that autoencoder-based models can achieve superior performance on anomaly detection tasks due to their excellent ability to fit complex data in an unsupervised manner. In this work, we propose a novel autoencoder-based model, named StackVAE-G that can significantly bring the efficiency and interpretability to multivariate time series anomaly detection. Specifically, we utilize the similarities across the time series channels by the stacking block-wise reconstruction with a weight-sharing scheme to reduce the size of learned models and also relieve the overfitting to unknown noises in the training data. We also leverage a graph learning module to learn a sparse adjacency matrix to explicitly capture the stable interrelation structure among multiple time series channels for the interpretable pattern reconstruction of interrelated channels. Combining these two modules, we introduce the stacking block-wise VAE (variational autoencoder) with GNN (graph neural network) model for multivariate time series anomaly detection. We conduct extensive experiments on three commonly used public datasets, showing that our model achieves comparable (even better) performance with the state-of-the-art models and meanwhile requires much less computation and memory cost. Furthermore, we demonstrate that the adjacency matrix learned by our model accurately captures the interrelation among multiple channels, and can provide valuable information for failure diagnosis applications.http://www.sciencedirect.com/science/article/pii/S2666651022000110Time-seriesAnomaly detectionAutoencodersGraph neural network |
| spellingShingle | Wenkai Li Wenbo Hu Ting Chen Ning Chen Cheng Feng StackVAE-G: An efficient and interpretable model for time series anomaly detection AI Open Time-series Anomaly detection Autoencoders Graph neural network |
| title | StackVAE-G: An efficient and interpretable model for time series anomaly detection |
| title_full | StackVAE-G: An efficient and interpretable model for time series anomaly detection |
| title_fullStr | StackVAE-G: An efficient and interpretable model for time series anomaly detection |
| title_full_unstemmed | StackVAE-G: An efficient and interpretable model for time series anomaly detection |
| title_short | StackVAE-G: An efficient and interpretable model for time series anomaly detection |
| title_sort | stackvae g an efficient and interpretable model for time series anomaly detection |
| topic | Time-series Anomaly detection Autoencoders Graph neural network |
| url | http://www.sciencedirect.com/science/article/pii/S2666651022000110 |
| work_keys_str_mv | AT wenkaili stackvaeganefficientandinterpretablemodelfortimeseriesanomalydetection AT wenbohu stackvaeganefficientandinterpretablemodelfortimeseriesanomalydetection AT tingchen stackvaeganefficientandinterpretablemodelfortimeseriesanomalydetection AT ningchen stackvaeganefficientandinterpretablemodelfortimeseriesanomalydetection AT chengfeng stackvaeganefficientandinterpretablemodelfortimeseriesanomalydetection |