Student Seminar

Spiking neural networks

Friday, 08 February 2019 01:30PM PST
Facebook
Twitter
LinkedIn
Reddit
SMS
Email
Copy
 
Student Seminar
 
Aleksei Bunevich
Department of Physics, SFU
 
Spiking neural networks
 
Feb 08, 2019 at 1:30PM
 

Synopsis

"Neural network" is a well known title, and it is commonly associated with computer science, as the most frequent applications are data processing, robotics, computer control and machine learning. However as from the word neuron one can imagine that these frameworks are based on real-life cells, whose working patterns were converted into the mathematical models by understanding physical principles. In this talk  two basic neuron models will be introduced: "Hodgkin Huxley" and "Integrate-and-Fire".  Most of modern Spiking Neural Networks (SNN) are based on these models.  Recent SNNs are capable of basic visual recognition and may be a key to understanding the way that the primate visual system learns.