Research

Artificial synapses: We are developing single devices that can emulate the characteristics of a biological synapse. Currently we are using 2D MoS2 as the active layer. We typically use graphene as the electrodes. We are focusing on developing these devices for unsupervised learning. Think of Mars rovers whose obstacles keep changing over time. These need to dynamically adapt to their environment. Our synapses will enable such unsupervised learning.

Artificial neurons: We are developing single devices that show spikes in time, similar to biological neurons. These neurons work in conjunction with artificial synapses to form spiking neural networks.

Photonic/optoelectronic synapses: Using materials that absorb light, we are using electrical controls to operate photodetectors as optoelectronic synapses. These devices respond to light pulses, change their conductance state and can remember that state when light is withdrawn. We want to create devices that can function the same way as the human eye.

Radiation effects and reliability of GaN and Ga2O3 devices: Supported by AFOSR MURI, we are working on the electrical characterization of radiation-induced defects in these devices.