- This event has passed.
Colloquium – Neal Patwari
April 1 @ 10:00 am - 11:00 am
Colloquium
Neal Patwari
Washington University in St. Louis
April 1, 2024
10:00am
3780 WEB
Better systems models help design wireless and equitable systems
This talk follows the philosophy that models should be as simple as possible “without having to surrender the adequate representation of a single datum of experience” (Einstein). This talk applies this in two parts. First, we address the need for new technologies to enable the spectrum sharing systems, motivated by the debate over spectrum sharing rules in the 6 GHz band. New adaptable and explainable channel prediction models are needed for efficient sharing. In addition, monitoring and control mechanisms are required to deal with the worst-case interference problems. The talk covers methods to address these needs. Second, we address the fairness of AI systems, using accent bias in speech recognition systems as an example. Our models for AI systems’ interaction with societal inequities should represent key context: What assumptions are made about the inputs an AI system operates on? What feedback mechanisms exist between the system and societal inequity? By modeling aspects of this interaction between a system’s inputs and outputs, and society, we can design better systems that lead to more equity over time.
Bio:
Neal Patwari is a Professor at Washington University in St. Louis, jointly appointed in the Department of Electrical and Systems Engineering and the Department of Computer Science and Engineering. He directs the Sensors, People, and Networks (SPAN) Lab, which investigates wireless communications, and equity, within networked engineered systems. He has projects in feedback models for inequity, and biases in pulse oximetry and speech recognition. He develops tools for POWDER, an open city-scale software-defined radio testbed that enables next-generation wireless research. His prior work has improved privacy in wireless networks, and capabilities for using a wireless network as a sensor. He has a BS (1997) and MS (1999) in EE from Virginia Tech, and a Ph.D. in EE from the University of Michigan, Ann Arbor (2005). He received the NSF CAREER Award in 2008, the 2009 IEEE Signal Processing Society Best Magazine Paper Award, and the 2011 University of Utah Early Career Teaching Award. He has co-authored papers with best paper awards at IEEE SenseApp 2012 and at the ACM/IEEE IPSN 2014 conference. Neal served as the TPC co-chair of IPSN 2020 and ACM Sensys 2023, and has served as member of the TPC of conferences such as IPSN, MobiCom, SECON, FAccT, and SenSys.