Title: Physics-Guided Machine Learning for Wireless System Design
报告人：Hao Ye, Assistant Professor, University of California, Santa Cruz
Abstract: Deep learning is revolutionizing wireless communication by offering advanced solutions for complex signal processing tasks. However, scalability and generalization remain challenging without leveraging the underlying physical properties of wireless channels. This talk presents two works that integrate channel knowledge into neural network design. First, we introduce ChannelNet, a purely data-driven massive MIMO detector designed to overcome the challenges of high-dimensional systems. By embedding channel layers and employing antenna-wise shared feature processors, ChannelNet maintains equivariance to antenna permutations, enabling efficient scaling to large antenna arrays. With a computational complexity of O(N_tN_r), it offers both theoretical guarantees, such as universal approximation of permutation-symmetric and maximum likelihood detection functions, and practical advantages, including superior or comparable performance to state-of-the-art methods across diverse scenarios. Second, we propose a neural surrogate model for wireless signal propagation, offering a differentiable, continuous representation of the environment. This model facilitates accurate predictions and enables applications such as localization and network planning. Together, these works demonstrate the potential of physics-guided learning to address key challenges in next-generation wireless systems.
Bio: Hao Ye is currently an assistant professor in Electrical and Computer Engineering at University of California, Santa Cruz. He received the B.E. degree in electrical engineering from Southeast University, Nanjing, China, in 2012, the M. E. degree in electrical engineering from Tsinghua University, Beijing, China, in 2015, and the Ph.D. degree in electrical and computer engineering from the Georgia Institute of Technology, Atlanta, GA, USA, in 2020. From 2021 to 2023, he was a Researcher with Qualcomm AI Research, San Diego, CA, USA. His research interests include machine learning and wireless communications. He was the recipient of Fred W. Ellersick Prize Paper Award in 2022.
