| Special Seminar Talks

Name of the Speaker: Mr. Sourodeep (EE19D414
Guide: Dr. Bhaswar Chakrabarti
Venue: ESB-210B (Conference Hall)
Date/Time: 31th July 2024 (Wednesday), 4:00 AM
Title: Spike-Time Dependent Plasticity in HfO2-Based Ferroelectric FET Synapses

Abstract :

Spiking Neural Networks (SNNs), inspired by biological neural networks, offer efficient information processing in terms of energy consumption. They have several advantages over traditional Artificial Neural Networks (ANNs), including the ability to process spatio-temporal information effectively. However, training SNNs in hardware presents challenges due to the non-differentiable nature of spikes. As a result, various event-timing-based methods, such as Spike Timing Dependent Plasticity (STDP), are employed. Emerging non-volatile memory devices, which are scalable, reliable, and energy-efficient, like Ferroelectric Field-Effect Transistors (FeFETs), could be excellent for demonstrating STDP-based learning. In this study, we discuss the learning behavior of an HfO2-based FeFET synaptic device for various spike shapes and the dependence of energy efficiency on spike parameters.