报告题目：Machine Learning for Man-Made Systems: From Monitoring to Control
报 告 人：翁阳 助理教授（美国亚利桑那州立大学）
报告时间：2019年11月28日10：00
报告地点：南一楼中311室
Abstract: In this talk, we are going to introduce ideas on the evolution of Artificial Intelligence from 1.0 to 3.0, and from monitoring to control. For this purpose, we will use the world’s largest machine, the electric power grid, as an example to show three examples on this evolution. At the first stage, we will show one example on how to increase forecastability in precursors for contingency, e.g., fault, by using structural and system information. During such a process, we illustrate how big data can reveal pattern that a small-scale data can not illustrate. We will also show two platforms that we speed up the transition from ideas to products. In the second phase, we explain how to create big data monitoring by using public clouds. Three examples will be given on distribution grid topology recovery, charging electric vehicles, and solar panel discovery. Finally, we will explain how to design learning method to make the Artificial Intelligence sustainable, especially in the controls. The example of Alpha Go 1, 0, and 2 based on various reinforcement learning methods will be explained with transferability to any man-made systems.
Bio: Yang Weng is an assistant professor at the School of Electrical, Computer and Energy Engineering (ECEE) of Arizona State University. He received his B.S. in Automation at 2006 with extremely outstanding academic performance (学习特优生) from the department of Control Science and Engineering. The department is now the School of AI and Automation. He obtained his Ph.D. in Electrical and Computer Engineering (ECE) from Carnegie Mellon University, where he also obtained his M.S. degree in Machine Learning from the School of Computer Science. Before joining ASU, Yang was a TomKat postdoctoral fellow at Stanford University. Yang was the Best Paper Award winner of the 2012 International Conference on Smart Grid Communication. In 2013, his paper ranked first in the same conference. In 2014, his paper was among the Best Papers at the IEEE Power and Energy Society General Meeting. In 2016, his paper won the Best of Best Paper Award at the International Conference on Probabilistic Methods Applied to Power Systems. In 2017, He earned the best paper award at the IEEE International Conference on Energy Internet and Energy System Integration. In 2018, his team received the winner award of Chunhui Cup in the International Competition on Innovation and Entrepreneurship: Renewable Energy Division. In 2019, his team won two best paper awards at the IEEE North American Power Symposium and IEEE Sustainable Power & Energy Conference. In the same year, his team won the 2nd place in Accuracy and 1st place in Speed for RTE International Competition on "Learning to Run a Power Network”. Finally, Yang is a task force chair at the IEEE PES Subcommittee on Big Data & Analytics for Power Systems. He is also the co-chair of the IEEE Phoenix Chapter - Conference Division.
学院公众号二维码
