March 22, 2023
Undergraduate students interested in either of the following for EE 499 credit can reach out to: arpan99@uw.edu
Problem Statement 1: NVM-based Crossbar architecture.
Description:
Emerging memory technology is a trending research domain. Crossbar architecture is crucial for future memory technologies. This project is to implement a set of equations (KCL & KVL) governing the voltage and current distribution of a crossbar array in MATLAB/Python. A research paper will be given to the student containing the basic theory and the set of equations. The main job is to create a function for current and voltage distribution in the crossbar architecture with a set of parameters. The code will then be improved to incorporate various non-idealities associated with a crossbar system.
Qualifications Required:
Suited for students with an interest in scientific computation and memory technology. Prior exposure to MATLAB/Python is strongly recommended.
Problem Statement 2: Spiking Neural Network (SNN)
Description:
Neuromorphic computing is on the rise. Inspired by our brain’s neural net, a neural network has been developed which takes inputs in form of spikes and the neurons are governed by a set of equations. In this work, a student has to develop a simple SNN model and understand the impact of spike time correlation on SNN accuracy. No prior knowledge of SNN is required. This problem can be approached from a mathematical standpoint. This will be a code-heavy project.
Qualifications Required:
Suited for students with an interest in-memory technology and machine learning. Prior exposure to MATLAB/Python is strongly recommended. A good understanding of mathematics and neural networks is needed.