SSE Talk | 下周讲座预告

Dear all,

You are cordially invited to two academic seminars. Please find the details as follows.

Seminar 1

✦

✦

✦

✦

✦

Topic: Deep Learning Meets Sparse Regularization

Time & Date: 2:00-3:00 PM, January 10 (Wednesday)

Venue: Room 208, Cheng Dao Building

Speaker: Dr. Rahul Parhi, École Polytechnique Fédérale de Lausanne

Host: Prof. Tom LUO and Prof. Feng YIN, The Chinese University of Hong Kong, Shenzhen 

Language: English

Abstract：

Deep learning has been wildly successful in practice and most state-of-the-art artificial intelligence systems are based on neural networks. Lacking, however, is a rigorous mathematical theory that adequately explains the amazing performance of deep neural networks.

In this talk, I present a new mathematical framework that provides the beginning of a deeper understanding of deep learning. This framework precisely characterizes the functional properties of trained neural networks. The key mathematical tools which support this framework include transform-domain sparse regularization, the Radon transform of computed tomography, and approximation theory. This framework explains the effect of weight decay regularization in neural network training, the importance of skip connections and low-rank weight matrices in network architectures, the role of sparsity in neural networks, and explains why neural networks can perform well in high-dimensional problems.

Bios：

Rahul Parhi is currently a postdoctoral researcher with the Biomedical Imaging Group at the École Polytechnique Fédérale de Lausanne (EPFL).

He completed his Ph.D. in electrical engineering at the University of Wisconsin-Madison in 2022, where he was supported by an NSF graduate research fellowship. His research interests include applications of functional and harmonic analysis to problems in signal processing and data science, in particular, the mathematical foundations of neural networks and deep learning.

Seminar 2

✦

✦

✦

✦

✦

Topic: Mechanism Designs for Efficiency and Incentive Compatibility in Electricity Markets: Emerging Technology and Strategic Participation

Time & Date: 3:00-4:00 PM, January 12 (Friday)

Venue: Room 108, Teaching D

Speaker: Prof. Pengcheng YOU, Peking University

Host: Prof. Chenye WU, The Chinese University of Hong Kong, Shenzhen 

Language: English

Abstract：

This talk discusses several challenges faced by deregulated electricity markets, and we provide insights into how some of them could be potentially tackled via analytical market mechanism designs. First, emerging technologies such as energy storage have been entitled to participate and compete in markets. However, they may differ drastically from traditional thermal generation in cost structures and operational constraints, which were not accounted for when existing markets were designed. This poses challenges to the basic principles under which electricity markets worldwide operate. We use energy storage as an example and show that a minimalist market modification on their bid structure could allow the precise incentive to be communicated. The resulting market-clearing process yields both efficient dispatch and incentive-compatible prices for cyclical operation of energy storage. Second, price manipulation (market power) has been a long-standing problem in electricity markets and even beyond. To alleviate the loss in social welfare caused by strategic bidding behavior of individual participants, we study a supply-demand balancing market with linear supply function bidding, and propose to exploit the role of subsidy. We design a subsidy mechanism that is able to capture and compensate the mismatch between individual incentive and social welfare, and as a result, fully recovers equilibrium social optimum. In particular, the supply function equilibrium achieves the underlying efficient supply dispatch and the market clearing price that reflects the truthful system marginal production cost.

Bios：

Dr. Pengcheng YOU is an Assistant Professor in the Department of Industrial Engineering and Management, College of Engineering, at Peking University, starting from Spring 2022. Dr. YOU was a postdoctoral fellow in the Department of Electrical and Computer Engineering and also a lecturer in the Engineering for Professionals Program at Johns Hopkins University. His research interests span optimization, control, learning and game theory, with a particular focus on interconnected engineering and economic systems. Rooted in energy, his major work broadly involves power systems, transportation networks, electric vehicles and electricity markets, and addresses the problems of model- and data-driven market designs, saddle flow dynamics, and management of distributed energy resources. He earned his B.S. in EE and Ph.D. in Control, both from Zhejiang University.

【END】

点击以下链接，进入理工时刻：