TADA: A Transferable Domain-Adversarial Training for Smart Grid Intrusion Detection Based on Ensemble Divergence Metrics and Spatiotemporal Features
For attack detection in the smart grid, transfer learning is a promising solution to tackle data distribution divergence and maintain performance when facing system and attack variations. However, there are still two challenges when introducing transfer learning into intrusion detection: when to app...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-11-01
|
| Series: | Energies |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1996-1073/15/23/8778 |
| _version_ | 1797463319754309632 |
|---|---|
| author | Pengyi Liao Jun Yan Jean Michel Sellier Yongxuan Zhang |
| author_facet | Pengyi Liao Jun Yan Jean Michel Sellier Yongxuan Zhang |
| author_sort | Pengyi Liao |
| collection | DOAJ |
| description | For attack detection in the smart grid, transfer learning is a promising solution to tackle data distribution divergence and maintain performance when facing system and attack variations. However, there are still two challenges when introducing transfer learning into intrusion detection: when to apply transfer learning and how to extract effective features during transfer learning. To address these two challenges, this paper proposes a transferability analysis and domain-adversarial training (TADA) framework. The framework first leverages various data distribution divergence metrics to predict the accuracy drop of a trained model and decides whether one should trigger transfer learning to retain performance. Then, a domain-adversarial training model with CNN and LSTM is developed to extract the spatiotemporal domain-invariant features to reduce distribution divergence and improve detection performance. The TADA framework is evaluated in extensive experiments where false data injection (FDI) attacks are injected at different times and locations. Experiments results show that the framework has high accuracy in accuracy drop prediction, with an RMSE lower than 1.79%. Compared to the state-of-the-art models, TADA demonstrates the highest detection accuracy, achieving an average accuracy of 95.58%. Moreover, the robustness of the framework is validated under different attack data percentages, with an average F1-score of 92.02%. |
| first_indexed | 2024-03-09T17:49:02Z |
| format | Article |
| id | doaj.art-06ace71d0f5d415cb0b320af824cd46e |
| institution | Directory Open Access Journal |
| issn | 1996-1073 |
| language | English |
| last_indexed | 2024-03-09T17:49:02Z |
| publishDate | 2022-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Energies |
| spelling | doaj.art-06ace71d0f5d415cb0b320af824cd46e2023-11-24T10:50:07ZengMDPI AGEnergies1996-10732022-11-011523877810.3390/en15238778TADA: A Transferable Domain-Adversarial Training for Smart Grid Intrusion Detection Based on Ensemble Divergence Metrics and Spatiotemporal FeaturesPengyi Liao0Jun Yan1Jean Michel Sellier2Yongxuan Zhang3Department of Electrical and Computer Engineering (ECE), Concordia University, Montréal, QC H3G 1M8, CanadaConcordia Institute for Information Systems Engineering (CIISE), Concordia University, Montréal, QC H3G 1M8, CanadaEricsson GAIA Montréal, AI hub Canada, Montréal, QC H4S 0B6, CanadaDepartment of Computer Science and Software Engineering (CSSE), Concordia University, Montréal, QC H3G 1M8, CanadaFor attack detection in the smart grid, transfer learning is a promising solution to tackle data distribution divergence and maintain performance when facing system and attack variations. However, there are still two challenges when introducing transfer learning into intrusion detection: when to apply transfer learning and how to extract effective features during transfer learning. To address these two challenges, this paper proposes a transferability analysis and domain-adversarial training (TADA) framework. The framework first leverages various data distribution divergence metrics to predict the accuracy drop of a trained model and decides whether one should trigger transfer learning to retain performance. Then, a domain-adversarial training model with CNN and LSTM is developed to extract the spatiotemporal domain-invariant features to reduce distribution divergence and improve detection performance. The TADA framework is evaluated in extensive experiments where false data injection (FDI) attacks are injected at different times and locations. Experiments results show that the framework has high accuracy in accuracy drop prediction, with an RMSE lower than 1.79%. Compared to the state-of-the-art models, TADA demonstrates the highest detection accuracy, achieving an average accuracy of 95.58%. Moreover, the robustness of the framework is validated under different attack data percentages, with an average F1-score of 92.02%.https://www.mdpi.com/1996-1073/15/23/8778cybersecuritysmart gridtransferability analysisadversarial trainingspatiotemporal featuretransfer learning |
| spellingShingle | Pengyi Liao Jun Yan Jean Michel Sellier Yongxuan Zhang TADA: A Transferable Domain-Adversarial Training for Smart Grid Intrusion Detection Based on Ensemble Divergence Metrics and Spatiotemporal Features Energies cybersecurity smart grid transferability analysis adversarial training spatiotemporal feature transfer learning |
| title | TADA: A Transferable Domain-Adversarial Training for Smart Grid Intrusion Detection Based on Ensemble Divergence Metrics and Spatiotemporal Features |
| title_full | TADA: A Transferable Domain-Adversarial Training for Smart Grid Intrusion Detection Based on Ensemble Divergence Metrics and Spatiotemporal Features |
| title_fullStr | TADA: A Transferable Domain-Adversarial Training for Smart Grid Intrusion Detection Based on Ensemble Divergence Metrics and Spatiotemporal Features |
| title_full_unstemmed | TADA: A Transferable Domain-Adversarial Training for Smart Grid Intrusion Detection Based on Ensemble Divergence Metrics and Spatiotemporal Features |
| title_short | TADA: A Transferable Domain-Adversarial Training for Smart Grid Intrusion Detection Based on Ensemble Divergence Metrics and Spatiotemporal Features |
| title_sort | tada a transferable domain adversarial training for smart grid intrusion detection based on ensemble divergence metrics and spatiotemporal features |
| topic | cybersecurity smart grid transferability analysis adversarial training spatiotemporal feature transfer learning |
| url | https://www.mdpi.com/1996-1073/15/23/8778 |
| work_keys_str_mv | AT pengyiliao tadaatransferabledomainadversarialtrainingforsmartgridintrusiondetectionbasedonensembledivergencemetricsandspatiotemporalfeatures AT junyan tadaatransferabledomainadversarialtrainingforsmartgridintrusiondetectionbasedonensembledivergencemetricsandspatiotemporalfeatures AT jeanmichelsellier tadaatransferabledomainadversarialtrainingforsmartgridintrusiondetectionbasedonensembledivergencemetricsandspatiotemporalfeatures AT yongxuanzhang tadaatransferabledomainadversarialtrainingforsmartgridintrusiondetectionbasedonensembledivergencemetricsandspatiotemporalfeatures |