A MEMS-based Analog Computer for Edge AI Computing

A MEMS-based Analog Computer for Edge AI Computing

Thursday, October 17, 2024 - 4:30pm to 5:30pm
Venue: Bldg. 320-109
Speaker: Fadi Alsaleem (UNL) and David Lin (GE Aerospace)

Abstract / Description 

*To receive email announcements and live stream information for upcoming seminars, please subscribe to the SystemX Seminar/EE310 Mailing list here.

NOTE: Dr. David Lin's portion of the talk will NOT be recorded

This talk presents the novel aspects of using MEMS devices as computers. First, Dr. Lin from GE will report findings of a MEMS-based computer that has demonstrated a 300x reduction in power consumption and a 100x improvement in speed compared to a typical digital computer when solving the same machine learning signal de-noising problem. Then, Dr. Alsaleem will showcase another advantage of using MEMS as computers by integrating sensing and computing into a single physical layer. This new computing paradigm has been shown to perform sensing and classification tasks simultaneously with near-zero latency and extremely low power consumption—on the order of 10^-17 kWh per operation.

Bios 

Dr. Alsaleem is an associate professor at the College of Engineering at the University of Nebraska at Lincoln (UNL). He has over six years of industry experience related to IoT devices.  He received over 7.5 million in funding from federal funding agencies such as NSF, DOE, and IARPA to support his research work related to MEMS analog computers.

Dr. David Lin is a Principal Engineer of Microsystems at GE Aerospace Research. His research work focuses on Microsystem design, integration, and reliability. David has 22 years of industrial experience in microsystem development and a proven track record of bringing MEMS devices from concept generation to production. David holds 45 issued US patents.

Calendar