Deep anomaly detection for time-series data in industrial IoT: a communication-efficient on-device federated learning approach
Since edge device failures (i.e., anomalies) seriously affect the production of industrial products in Industrial IoT (IIoT), accurately and timely detecting anomalies is becoming increasingly important. Furthermore, data collected by the edge device may contain the user's private data, whic...
| Main Authors: | , , , , , , |
|---|---|
| Other Authors: | |
| Format: | Journal Article |
| Language: | English |
| Published: |
2022
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/159853 |
| _version_ | 1826119526718111744 |
|---|---|
| author | Liu, Yi Garg, Sahil Nie, Jiangtian Zhang, Yang Xiong, Zehui Kang, Jiawen Hossain, M. Shamim |
| author2 | School of Computer Science and Engineering |
| author_facet | School of Computer Science and Engineering Liu, Yi Garg, Sahil Nie, Jiangtian Zhang, Yang Xiong, Zehui Kang, Jiawen Hossain, M. Shamim |
| author_sort | Liu, Yi |
| collection | NTU |
| description | Since edge device failures (i.e., anomalies) seriously affect the production
of industrial products in Industrial IoT (IIoT), accurately and timely
detecting anomalies is becoming increasingly important. Furthermore, data
collected by the edge device may contain the user's private data, which is
challenging the current detection approaches as user privacy is calling for the
public concern in recent years. With this focus, this paper proposes a new
communication-efficient on-device federated learning (FL)-based deep anomaly
detection framework for sensing time-series data in IIoT. Specifically, we
first introduce a FL framework to enable decentralized edge devices to
collaboratively train an anomaly detection model, which can improve its
generalization ability. Second, we propose an Attention Mechanism-based
Convolutional Neural Network-Long Short Term Memory (AMCNN-LSTM) model to
accurately detect anomalies. The AMCNN-LSTM model uses attention
mechanism-based CNN units to capture important fine-grained features, thereby
preventing memory loss and gradient dispersion problems. Furthermore, this
model retains the advantages of LSTM unit in predicting time series data.
Third, to adapt the proposed framework to the timeliness of industrial anomaly
detection, we propose a gradient compression mechanism based on Top-k
selection to improve communication efficiency. Extensive experiment studies on
four real-world datasets demonstrate that the proposed framework can accurately
and timely detect anomalies and also reduce the communication overhead by 50%
compared to the federated learning framework that does not use a gradient
compression scheme. |
| first_indexed | 2024-10-01T05:01:28Z |
| format | Journal Article |
| id | ntu-10356/159853 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T05:01:28Z |
| publishDate | 2022 |
| record_format | dspace |
| spelling | ntu-10356/1598532022-07-04T08:23:11Z Deep anomaly detection for time-series data in industrial IoT: a communication-efficient on-device federated learning approach Liu, Yi Garg, Sahil Nie, Jiangtian Zhang, Yang Xiong, Zehui Kang, Jiawen Hossain, M. Shamim School of Computer Science and Engineering Interdisciplinary Graduate School (IGS) Energy Research Institute @ NTU (ERI@N) Alibaba-NTU Joint Research Institute Engineering::Computer science and engineering Deep Anomaly Detection Federated Learning Since edge device failures (i.e., anomalies) seriously affect the production of industrial products in Industrial IoT (IIoT), accurately and timely detecting anomalies is becoming increasingly important. Furthermore, data collected by the edge device may contain the user's private data, which is challenging the current detection approaches as user privacy is calling for the public concern in recent years. With this focus, this paper proposes a new communication-efficient on-device federated learning (FL)-based deep anomaly detection framework for sensing time-series data in IIoT. Specifically, we first introduce a FL framework to enable decentralized edge devices to collaboratively train an anomaly detection model, which can improve its generalization ability. Second, we propose an Attention Mechanism-based Convolutional Neural Network-Long Short Term Memory (AMCNN-LSTM) model to accurately detect anomalies. The AMCNN-LSTM model uses attention mechanism-based CNN units to capture important fine-grained features, thereby preventing memory loss and gradient dispersion problems. Furthermore, this model retains the advantages of LSTM unit in predicting time series data. Third, to adapt the proposed framework to the timeliness of industrial anomaly detection, we propose a gradient compression mechanism based on Top-k selection to improve communication efficiency. Extensive experiment studies on four real-world datasets demonstrate that the proposed framework can accurately and timely detect anomalies and also reduce the communication overhead by 50% compared to the federated learning framework that does not use a gradient compression scheme. Nanyang Technological University This work was supported in part by the Alibaba Group through the Alibaba Innovative Research (AIR) Program and the Alibaba-NTU Singapore Joint Research Institute (JRI), NTU, Singapore; in part by the Open Research Project of the State Key Laboratory of Industrial Control Technology, Zhejiang University, China, under Grant ICT20044; in part by the National Natural Science Foundation of China under Grant 51806157; in part by the Young Innovation Talents Project in Higher Education of Guangdong Province, China, under Grant 2018KQNCX333; and in part by the Researchers Supporting Project under Grant RSP-2020/32, King Saud University, Riyadh, Saudi Arabia. 2022-07-04T08:23:11Z 2022-07-04T08:23:11Z 2020 Journal Article Liu, Y., Garg, S., Nie, J., Zhang, Y., Xiong, Z., Kang, J. & Hossain, M. S. (2020). Deep anomaly detection for time-series data in industrial IoT: a communication-efficient on-device federated learning approach. IEEE Internet of Things Journal, 8(8), 6348-6358. https://dx.doi.org/10.1109/JIOT.2020.3011726 2327-4662 https://hdl.handle.net/10356/159853 10.1109/JIOT.2020.3011726 2-s2.0-85104070584 8 8 6348 6358 en IEEE Internet of Things Journal © 2020 IEEE. All rights reserved. |
| spellingShingle | Engineering::Computer science and engineering Deep Anomaly Detection Federated Learning Liu, Yi Garg, Sahil Nie, Jiangtian Zhang, Yang Xiong, Zehui Kang, Jiawen Hossain, M. Shamim Deep anomaly detection for time-series data in industrial IoT: a communication-efficient on-device federated learning approach |
| title | Deep anomaly detection for time-series data in industrial IoT: a communication-efficient on-device federated learning approach |
| title_full | Deep anomaly detection for time-series data in industrial IoT: a communication-efficient on-device federated learning approach |
| title_fullStr | Deep anomaly detection for time-series data in industrial IoT: a communication-efficient on-device federated learning approach |
| title_full_unstemmed | Deep anomaly detection for time-series data in industrial IoT: a communication-efficient on-device federated learning approach |
| title_short | Deep anomaly detection for time-series data in industrial IoT: a communication-efficient on-device federated learning approach |
| title_sort | deep anomaly detection for time series data in industrial iot a communication efficient on device federated learning approach |
| topic | Engineering::Computer science and engineering Deep Anomaly Detection Federated Learning |
| url | https://hdl.handle.net/10356/159853 |
| work_keys_str_mv | AT liuyi deepanomalydetectionfortimeseriesdatainindustrialiotacommunicationefficientondevicefederatedlearningapproach AT gargsahil deepanomalydetectionfortimeseriesdatainindustrialiotacommunicationefficientondevicefederatedlearningapproach AT niejiangtian deepanomalydetectionfortimeseriesdatainindustrialiotacommunicationefficientondevicefederatedlearningapproach AT zhangyang deepanomalydetectionfortimeseriesdatainindustrialiotacommunicationefficientondevicefederatedlearningapproach AT xiongzehui deepanomalydetectionfortimeseriesdatainindustrialiotacommunicationefficientondevicefederatedlearningapproach AT kangjiawen deepanomalydetectionfortimeseriesdatainindustrialiotacommunicationefficientondevicefederatedlearningapproach AT hossainmshamim deepanomalydetectionfortimeseriesdatainindustrialiotacommunicationefficientondevicefederatedlearningapproach |