ակնարկ

Speaker: Davit Grigoryan
Date: Friday, Aug 26, 2022 
Time: 5:00 PM
Venue: Museum Lecture Hall
Description: As the limit for digital electronics is being reached, neural networks have taken over the research for AI implementation. However, the ending of Moore’s law would soon halt the development of Neural networks.

To overcome this, many research facilities have turned to an analog hardware implementation which uses a crossbar array to construct the network. Flash memory cells and memristors have been used for the construction of a proposed analogue neural network. During this research MOSFETS are used in the place of memristors to set the weight of the neurons. A crossbar array configuration is used during testing as they have been shown to be the most practical approach for commercialization and their wide use in third generation neural networks (spiking neural networks).