Deep Diffusion Models for Robust Channel Estimation

Presented by: Marius Arvinte Advisor: Prof. Jon Tamir My research deals with machine learning algorithms that work out-of-the-box in new environments, with a focus on physical layer processing and medical imaging. Robustness is important in digital communications, since channel conditions can change rapidly and it's the modem's job to make sure the connection is seamless. In our latest work, we use deep diffusion models for wireless channel estimation and show that this approach can be reliably used in new channel models with zero adaptation required. Find out more about my research: https://arxiv.org/abs/2111.08177 Topics: Machine Learning for Wireless Systems -- The Wireless Networking and Communications Group (WNCG) is an interdisciplinary center for research and education at The University of Texas at Austin with an emphasis on industrial relevance. Founded in 2002, the group includes 25 faculty from the Departments of Electrical and Computer Engineering, Aerospace Engineering, Mathematics, and Computer Sciences. WNCG currently has over 130 graduate and undergraduate students as well as an ever-changing number of research scientists, postdoctoral fellows and industry visitors. https://wncg.org/