Link for youtube video:

SNN based on lif model in erlang

This is an erlang implementation of spiking neural network. Our main traget is to develop a network that can determine between X and O. In future work, we want to use Spike-Time Dependent Plasticity (STDP) algorithm in order to train our model to identify more items.

In this branch we will present a full distributed SNN that works with two computers:

One computer would be the input layer of the SNN
Second computer would be the output layer of the SNN

Network Structure

As said before, we used two computers one for every layer.

In the first layer we have 256 neurons and in the second layer we have 4 neurons.

Our work was based on given weights: https://github.com/Shikhargupta/Spiking-Neural-Network

Image Processing

In order to translate the image into spike train, we used receptive field that helped us to encode the image.

In the folder of [Image processing](Image processing) you can see the relevant python files.

Results

After we tested our network we got results that show that we indeed identify between X/O:

Instructions

You need to open 6 different nodes that share the same cookie.

The nodes are: Monitor, resMonitor, Server, Graphics, SNN, outputLayer

Note: The ip address depends on your computers, so you can change it from the file.

Monitor - He is responsible on tracking the other nodes and check if they fell.

erl -name monitorNode@127.0.0.1 -setcookie test
c(monitor).
monitor:init(). % Activates the monitor and starts the program

resMonitor - If the main monitor fell, he replaces him and keep tracking the system.

erl -name resmonitorNode@127.0.0.1 -setcookie test
c(resmonitor).
resmonitor:init(). % Activates the resmonitor

Server - He is responsible on sending messages between the nodes.

erl -name serverNode@127.0.0.1 -setcookie test
c(server).

Graphics - Starts the node that responsible on gui & graphics

erl -name graphicsNode@127.0.0.1 -setcookie test
c(graphics).

SNN - The node of the input layer of the neural network.

erl -name snnNode@127.0.0.1 -setcookie test
c(neuron), c(layer), c(snn).

outLayer - The node of the output layer of the neural network.

erl -name outlayerNode@127.0.0.1 -setcookie test
c(neuron), c(layer), c(outlayer).

amiryt / Erlang-project