Analysis of Learning Mechanisms in Spiking Neural Networks with R(T) Elements and Memristive Synapses

ECE Doctoral Candidate Farhana Afrin is scheduled to present her dissertation Analysis of Learning Mechanisms in Spiking Neural Networks with R(T) Elements and Memristive Synapses on Friday, October 4.

Abstract

As Moore’s law approaches its limits, the traditional von Neumann architecture struggles with efficiency due to continuous power consumption. Spiking neural networks (SNNs), inspired by the human brain, offer a more energy-efficient alternative by using power only during spikes. While electronic SNNs show promises, challenges such as circuit complexity and parameter limitations persist. Researchers are addressing these issues by using memristors to emulate synapses, adjusting conductance based on Spike-TimingDependent Plasticity (STDP). This work explores algorithmic approaches to implement memristive SNNs using simpler circuit elements for improved performance.

Biographical Sketch

Afrin earned her B.S. in Electrical Engineering from the Military Institute of Science and Technology in 2015. After teaching in the EEE department at BUBT for three years, she pursued her Master’s in ECE at LSU, starting in Fall 2018. She has been working with Dr. Kurtis Cantley in the ENDS Lab since the fall of 2020. Her research centers on designing neuromorphic circuits and developing algorithms to better mimic biological neural behavior. Outside of work, she enjoys traveling and baking. Afrin is supported by her research advisor, Dr. Kurtis Cantley, and her committee members Drs. Nader Rafla and Benjamin Johnson.