Xiong, Feng
(2020)
Scalable Artificial Synapses with Tunable Spatio-Temporal Dynamics for Neuromorphic Computing.
In: Pitt Momentum Fund 2020, University of Pittsburgh, Pittsburgh, Pa.
(Unpublished)
Abstract
Artificial intelligence (AI) has the potential to transform people’s lives. While recent successes in machine learning (ML) has demonstrated promising applications in industries such as healthcare, transportation, personal assistance, and advanced manufacturing, these ML computations consume huge amounts of energy in conventional computing systems. Inspired by the human brain, which is better at cognitive tasks such as pattern recognition than even supercomputers at a much lower energy efficiency, neuromorphic computing and artificial neural networks have recently attracted immense research interest. In particular, spiking neural network (SNN), which mimics the biological neural network by incorporating the temporal dynamics between stimulations, offers a promising route for energy-efficient computing with high bandwidth. However, it is challenging and expensive to implement the spatio-temporal dynamics (such as short-term and long-term memory) in SNN with existing digital electronics. In this project, the PI will develop a dynamic synapse capable of programmable spatio-temporal dynamics by controlling the charge carrier concentration in two-dimensional devices. By developing this critically missing element, this work will lead to a truly neuro-realistic computing system with orders-of-magnitude improvements in energy efficiency, bandwidth, and cognitive capabilities. This can lead to the wide use of AI and revolutionize the society through technologies such as cognitive computing, self-driving vehicles, smart wearable electronics, and autonomous manufacturing.
Share
Citation/Export: |
|
Social Networking: |
|
Details
Metrics
Monthly Views for the past 3 years
Plum Analytics
Altmetric.com
Actions (login required)
|
View Item |