Machine Learning Methods Embedded With Domain Knowledge (Part II): Generalization Risk

被引：0

作者：

Shang Y. ^{[1
]}

Guo J. ^{[2
]}

Wu W. ^{[1
]}

Su J. ^{[2
]}

Liu W. ^{[2
]}

Zhuang S. ^{[3
]}

Zhou L. ^{[2
]}

机构：

[1] State Key Lab of Control and Simulation of Power Systems and Generation Equipments, Tsinghua University, Haidian District, Beijing

[2] China Electric Power Research Institute, Haidian District, Beijing

[3] North China Electric Power University, Changping Distirct, Beijing

来源：

Zhongguo Dianji Gongcheng Xuebao/Proceedings of the Chinese Society of Electrical Engineering | 2019年 / 39卷 / 16期

关键词：

Data driven; Generalization risk; Knowledge guiding; Machine learning; Statistical learning theory;

D O I：

10.13334/j.0258-8013.pcsee.190479

中图分类号：

学科分类号：

摘要：

The theoretical achievements of data-driven machine learning models (DDM) were briefly reviewed. Then, the generalization risks of knowledge-guiding & data-driven machine learning model (KDM) in both the local learning space and global learning space were analyzed. The results show that, under certain assumptions, KDM can bound its generalization error in the local learning space approaching probability 1, and bound its generalization error in the global learning space more tightly than the generalization error of DDM with some probability 1-δ. Compared with DDM, KDM is more efficient and robust under the circumstances with limited training samples. © 2019 Chin. Soc. for Elec. Eng.

引用

页码：4641 / 4649

页数：8

共 22 条

[11] Watkins C.J.C.H., Dayan P., Technical note: Q-learning, Machine Learning, 8, 3-4, pp. 279-292, (1992)
[12] Zou B., Zhang H., Xu Z., Learning from uniformly ergodic Markov chains, Journal of Complexity, 25, 2, pp. 188-200, (2009)
[13] Tsitsiklis J.N., Van Roy B., An analysis of temporal- difference learning with function approximation, IEEE Transactions on Automatic Control, 42, 5, pp. 674-690, (1997)
[14] Liu W., Zhuang P., Liang H., Et al., Distributed economic dispatch in microgrids based on cooperative reinforcement learning, IEEE Transactions on Neural Networks and Learning Systems, 29, 6, pp. 2192-2203, (2018)
[15] Wei Q., Liu D., Shi G., A novel dual iterative Q-learning method for optimal battery management in smart residential environments, IEEE Transactions on Industrial Electronics, 62, 4, pp. 2509-2518, (2015)
[16] Mnih V., Kavukcuoglu K., Silver D., Et al., Playing Atari with deep reinforcement learning, (2013)
[17] Silver D., Lever G., Heess N., Et al., Deterministic policy gradient algorithms, Proceedings of the 31st International Conference on Machine Learning, (2014)
[18] Bianchi R.A.C., Ribeiro C.H.C., Costa A.H.R., Accelerating autonomous learning by using heuristic selection of actions, Journal of Heuristics, 14, 2, pp. 135-168, (2008)
[19] Ziebart B.D., Maas A., Bagnell J.A., Et al., Maximum entropy inverse reinforcement learning, Proceedings of the 23rd National Conference on Artificial intelligence, pp. 1433-1438, (2008)
[20] Li C., Cao L., Zhang Y., Et al., knowledge-based deep reinforcement learning: a review, Systems Engineering and Electronics, 39, 11, pp. 2603-2613, (2017)

← 1 2 3 →