Event
Remote Ph.D. Dissertation Defense: Heidi Komkov
Friday, May 14, 2021
1:00 p.m.
https://zoom.us/j/6179028404
Maria Hoo
301 405 3681
mch@umd.edu
ANNOUNCEMENT: Remote Ph.D. Dissertation Defense
Name: Heidi Komkov
Committee members:
Professor Daniel Lathrop (Chair)
Professor Timothy Horiuchi
Professor Pamela Abshire
Professor Brian Beaudoin
Professor Timothy Koeth
Professor Vedran Lekic (Dean's representative)
Date/Time: Friday, May 14th, 2021 at 1PM
Location: https://zoom.us/j/6179028404
Title: Reservoir Computing with Boolean Logic Network Circuits
Abstract: To push the frontiers of machine learning, completely new computing architectures must be explored which efficiently use hardware resources. We test an unconventional use of digital logic gate circuits for reservoir computing, a machine learning algorithm that is used for rapid time series processing. In our approach, logic gates are configured into networks that can exhibit complex dynamics. Rather than the gates explicitly computing pre-programmed instructions, they are used collectively as a dynamical system that transforms input data into a higher dimensional representation. We probe the dynamics of such circuits using discrete components on a circuit board as well as an FPGA implementation. We show favorable machine learning performance, including radiofrequency classification accuracy comparable to a state of the art convolutional neural network with a fraction of the trainable parameters. Finally, we discuss the design and fabrication of a reservoir computing ASIC for high-speed time series processing.
