Ying Sun
Research Subjects
Ying Sun has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Fei Wu
Nanjing University of Posts and Telecommunications
1 shared publication - Jin Qi
School of Internet of Things Nanjing University of Posts and Telecommunications Nanjing China
1 shared publication - Sanlei Dang
Measurement Center of Guangdong Power Grid Co., Ltd., Guangzhou, Guangdong 510000, China
1 shared publication - Wan Huang
Dongguan Power Supply Bureau, Guangdong Power Grid Co., Ltd., Dongguan, Guangdong 523000, China
1 shared publication - Wenchong Fang
China Southern Power Grid Co., Ltd., Guangzhou, Guangdong 510000, China
1 shared publication - Xiao-Yuan Jing
School of Computer, Guangdong University of Petrochemical Technology, Maoming 525000, China
1 shared publication
Latest Publications
- 11 February 2025
- Published by World Scientific Pub Co Pte Ltd in International Journal of High Speed Electronics and Systems
- Vol. 34 (04)
Design and Analysis of Electricity Anomaly Detection and Diagnosis Models Based on Generative Adversarial Networks
Effective energy management and optimal utilization of assets are crucial for a nation’s progress and prosperity. However, as electricity travels from power plants to end users, it incurs two types of losses: Technical Losses (TL) and Non-Technical Losses (NTL). TLs arise from outdated or inefficient technologies, whereas NTLs result from abnormal electricity usage, including electricity theft, which is often conducted to reduce expenses. These losses pose significant challenges to maintaining grid reliability and result in reduced profits for utility operators. Although the implementation of Automatic Metering Infrastructure (AMI) has improved grid predictability, it has also created new vulnerabilities for NTLs through Cyber-Physical Theft Attacks (CPTA). Machine learning techniques have been applied to identify and mitigate CPTA; however, they often fail to capture comprehensive Energy Consumption Patterns (ECPs), limiting their effectiveness in detecting malicious activities. Generative Adversarial Networks (GANs) have shown remarkable success in various anomaly detection applications. In the context of electricity anomaly identification and diagnosis, GANs are particularly effective in modeling typical energy usage patterns and detecting deviations. Building on this foundation, this study introduces an Adaptive GAN (AGAN) model to detect and diagnose electricity anomalies. The AGAN dynamically optimizes key parameters, such as the number of hidden neurons, epochs and steps per epoch, using the Improved Honey Badger Algorithm (IHBA). This optimization enhances the model’s performance, particularly in terms of Detection Accuracy, Critical Success Index (CSI) and False Omission Rate (FOR). The proposed model demonstrates its capability to accurately identify electricity anomalies and has been rigorously evaluated against existing anomaly detection methods. Results highlight AGAN’s superior effectiveness and robustness, establishing it as a reliable solution for improving the reliability and efficiency of electricity grid systems.
- 11 February 2025
- Published by World Scientific Pub Co Pte Ltd in International Journal of High Speed Electronics and Systems
- Vol. 34 (04)
Design and Analysis of Electricity Anomaly Detection and Diagnosis Models Based on Generative Adversarial Networks
Effective energy management and optimal utilization of assets are crucial for a nation’s progress and prosperity. However, as electricity travels from power plants to end users, it incurs two types of losses: Technical Losses (TL) and Non-Technical Losses (NTL). TLs arise from outdated or inefficient technologies, whereas NTLs result from abnormal electricity usage, including electricity theft, which is often conducted to reduce expenses. These losses pose significant challenges to maintaining grid reliability and result in reduced profits for utility operators. Although the implementation of Automatic Metering Infrastructure (AMI) has improved grid predictability, it has also created new vulnerabilities for NTLs through Cyber-Physical Theft Attacks (CPTA). Machine learning techniques have been applied to identify and mitigate CPTA; however, they often fail to capture comprehensive Energy Consumption Patterns (ECPs), limiting their effectiveness in detecting malicious activities. Generative Adversarial Networks (GANs) have shown remarkable success in various anomaly detection applications. In the context of electricity anomaly identification and diagnosis, GANs are particularly effective in modeling typical energy usage patterns and detecting deviations. Building on this foundation, this study introduces an Adaptive GAN (AGAN) model to detect and diagnose electricity anomalies. The AGAN dynamically optimizes key parameters, such as the number of hidden neurons, epochs and steps per epoch, using the Improved Honey Badger Algorithm (IHBA). This optimization enhances the model’s performance, particularly in terms of Detection Accuracy, Critical Success Index (CSI) and False Omission Rate (FOR). The proposed model demonstrates its capability to accurately identify electricity anomalies and has been rigorously evaluated against existing anomaly detection methods. Results highlight AGAN’s superior effectiveness and robustness, establishing it as a reliable solution for improving the reliability and efficiency of electricity grid systems.
- Open Access
- 1 January 2021
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 68 (3) , 3373-3389
Unsupervised Domain Adaptation Based on Discriminative Subspace Learning for Cross-Project Defect Prediction
Cross-project defect prediction (CPDP) aims to predict the defects on target project by using a prediction model built on source projects. The main problem in CPDP is the huge distribution gap between the source project and the target project, which prevents the prediction model from performing well. Most existing methods overlook the class discrimination of the learned features. Seeking an effective transferable model from the source project to the target project for CPDP is challenging. In this paper, we propose an unsupervised domain adaptation based on the discriminative subspace learning (DSL) approach for CPDP. DSL treats the data from two projects as being from two domains and maps the data into a common feature space. It employs cross-domain alignment with discriminative information from different projects to reduce the distribution difference of the data between different projects and incorporates the class discriminative information. Specifically, DSL first utilizes subspace learning based domain adaptation to reduce the distribution gap of data between different projects. Then, it makes full use of the class label information of the source project and transfers the discrimination ability of the source project to the target project in the common space. Comprehensive experiments on five projects verify that DSL can build an effective prediction model and improve the performance over the related competing methods by at least 7.10% and 11.08% in terms of G-measure and AUC.
- 1 January 2021
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 68 (3) , 3373-3389
Open AccessUnsupervised Domain Adaptation Based on Discriminative Subspace Learning for Cross-Project Defect Prediction
Cross-project defect prediction (CPDP) aims to predict the defects on target project by using a prediction model built on source projects. The main problem in CPDP is the huge distribution gap between the source project and the target project, which prevents the prediction model from performing well. Most existing methods overlook the class discrimination of the learned features. Seeking an effective transferable model from the source project to the target project for CPDP is challenging. In this paper, we propose an unsupervised domain adaptation based on the discriminative subspace learning (DSL) approach for CPDP. DSL treats the data from two projects as being from two domains and maps the data into a common feature space. It employs cross-domain alignment with discriminative information from different projects to reduce the distribution difference of the data between different projects and incorporates the class discriminative information. Specifically, DSL first utilizes subspace learning based domain adaptation to reduce the distribution gap of data between different projects. Then, it makes full use of the class label information of the source project and transfers the discrimination ability of the source project to the target project in the common space. Comprehensive experiments on five projects verify that DSL can build an effective prediction model and improve the performance over the related competing methods by at least 7.10% and 11.08% in terms of G-measure and AUC.