TY - GEN
T1 - Heterogeneous transfer learning on power systems
T2 - 20th IEEE International Conference on Data Mining, ICDM 2020
AU - Li, Haoran
AU - Weng, Yang
AU - Tong, Hanghang
N1 - Publisher Copyright:
© 2020 IEEE.
PY - 2020/11
Y1 - 2020/11
N2 - Machine Learning (ML) is gaining increasing popularity to tackle uncertainty in physical systems, such as modern power systems. However, ML models can be hardly trained for newly-built power grids with limited data, especially when different power grids have different dimensionalities and distributions for measurement data. To tackle this problem, we propose a novel Heterogeneous Transfer Learning (HTL)-based method to boost the data volume of the target grid. Specifically, we propose a Merged Multi-Modal Gaussian Graphical Model (M{3}G{2}M) with a physical data merging process for knowledge transfer. To solve the maximum likelihood estimation of M{3}G{2}M with imbalanced data from two grids, we propose a novel Expectation-Maximization algorithm. Finally, we quantify the negative transfer via the KL-Divergence to measure the distribution similarity between the source grid and the target grid for the transferring confidence. We demonstrate the advantages and the generalizability of our proposed models in diversified data sets for power systems and human action-sensing systems.
AB - Machine Learning (ML) is gaining increasing popularity to tackle uncertainty in physical systems, such as modern power systems. However, ML models can be hardly trained for newly-built power grids with limited data, especially when different power grids have different dimensionalities and distributions for measurement data. To tackle this problem, we propose a novel Heterogeneous Transfer Learning (HTL)-based method to boost the data volume of the target grid. Specifically, we propose a Merged Multi-Modal Gaussian Graphical Model (M{3}G{2}M) with a physical data merging process for knowledge transfer. To solve the maximum likelihood estimation of M{3}G{2}M with imbalanced data from two grids, we propose a novel Expectation-Maximization algorithm. Finally, we quantify the negative transfer via the KL-Divergence to measure the distribution similarity between the source grid and the target grid for the transferring confidence. We demonstrate the advantages and the generalizability of our proposed models in diversified data sets for power systems and human action-sensing systems.
KW - Expectation-Maximization Algorithm
KW - Heterogeneous Transfer Learning
KW - Maximum Likelihood Estimation
KW - Merged Multi-Modal Gaussian Graphical Model
UR - http://www.scopus.com/inward/record.url?scp=85100891825&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85100891825&partnerID=8YFLogxK
U2 - 10.1109/ICDM50108.2020.00130
DO - 10.1109/ICDM50108.2020.00130
M3 - Conference contribution
AN - SCOPUS:85100891825
T3 - Proceedings - IEEE International Conference on Data Mining, ICDM
SP - 1088
EP - 1093
BT - Proceedings - 20th IEEE International Conference on Data Mining, ICDM 2020
A2 - Plant, Claudia
A2 - Wang, Haixun
A2 - Cuzzocrea, Alfredo
A2 - Zaniolo, Carlo
A2 - Wu, Xindong
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 17 November 2020 through 20 November 2020
ER -