Description | In the context of fast and energy-efficient implementation of AI in edge devices, photonic-crossbar-array-based neuromorphic computing has advantages of higher data bandwidth and simultaneous data processing (due to wavelength division multiplexing), as compared to its electronic counterpart. The objective of this project is to introduce additional novel functionalities in photonic neuromorphic computing through the use of nanomagnetic/ spintronic devices as non-volatile synapses in photonic neural networks with the help of their magneto-optic properties. The project will involve micromagnetics-based simulation of domain-wall motion in the spintronic synapse, followed by electromagnetics-based modeling of the spintronic device integrated into a waveguide for modulation of light's intensity through domain-wall motion (non-volatile property) in the spintronic device. Then large photonic neural networks will be modelled using this device model and wavelength division multiplexing will be implemented in them. |
---|---|
Number of students | 1 |
Year of study | Students in their 3rd year (Semester 5) |
CPI | 8 and above |
Prerequisites | The student must have taken introductory undergrad courses both in Physics and Elec. Engg. So I would only prefer EE and EP undergrad students for this project |
Duration | 6 months (can be extended to BTP or DDP) |
Learning outcome | - |
Weekly time commitment | 7 hours |
General expectations | The student is expected to continue this project as BTP or DDP |
Assignment |
Lectures from my ongoing course this semester at IIT Bombay
(EE501: Spin-Based Computing) will be a good source. The
lecture recordings (through CDEEP) are here: https://www.youtube.com/watch?v=5Z2l82SV-hg&list=PLDREIwGwrHBdMXZxYXDRFKIhjIpMR09G9 |
Instructions for assignment | Go through the videos mentioned above |
Dive deep in the realm of Academic and Industry related research projects
Created with ❤️ by UGAC Web Team, 2023-2024