Scratch for Neurones

1. Single Neurone


Instructions:

• Set the inputs by pressing the buttons marked input 1 and input 2 (Red is off(False or 0) and Green is on(True or 1))
• Change the weights by changing weights 1 to 3, wx goes with input x and weight 3 is the bias.
• To activate the neuron you need to click on the the yellow ball ('the neuron').

The video below show it in action and explains the code.



To see the code go to https://scratch.mit.edu/projects/131892234/ .

A slight modification click on the bell to change the weights

The code is available at https://scratch.mit.edu/projects/171190294/

2. Training a Neurone
In this part, the training of a neuron all written in Scratch is tackled. The video shows it action and you can have a go at using the software yourself at the end of the post. The Scratch code can be found at https://scratch.mit.edu/projects/132915502/






All opinions in this blog are the Author's and should not in any way be seen as reflecting the views of any organisation the Author has any association with. Twitter @scottturneruon

Lego Robot and Neural Networks

An overview of using Lego RCX  robots for teaching neural networks present at workshop in 2011.

Derby presentation from Scott Turner

The video below shows the robot trying out sets of weights for two neurones, until a set of weights are found that enable the robot to go around the circle.

As a part of a set of tools I have found the following useful for teaching the principles of simple neurones.

 Example code:

import josx.platform.rcx.*;

public class annlf{

 public static void main(String[] args)

 {

  int w[][] ={//put weights here};

  int o[]={1,1};

  int s1,s2,res1,res2;

  int sensor1=0,sensor2=0;

  robot_1 tom=new robot_1();

  Sensor.S1.activate();

  Sensor.S3.activate();

  for(;;){

   sensor1=Sensor.S1.readValue();

   sensor2=Sensor.S3.readValue();

   LCD.showNumber(sensor1);

   if (sensor1<42)

    s1=1;

   else

    s1=0;

   if (sensor2<42)

    s2=1;

   else

    s2=0;

   res1=w[0][1]*s1+w[0][2]*s2+w[0][0];

   if (res1>=0)

    o[0]=1;

   else

    o[0]=0;

   res2=w[1][1]*s1+w[1][2]*s2+w[1][0];

   if (res2>=0)

    o[1]=1;

   else

    o[1]=0;

   if ((o[0]==1)&&(o[1]==1))

    tom.forward1(10);

   if ((o[0]==0)&&(o[1]==0))

    tom.backward1(20);

   if ((o[0]==1)&&(o[1]==0))

    tom.tlturn(20);

   if ((o[0]==0)&&(o[1]==1))

    tom.trturn(20);

   LCD.refresh();

  }

 }

}

The example code uses two neurones to produce a line follower. The nice thing about this though is it easy to adapted this for a single neuron or multiple neuron tasks. For more on this some examples can be found here.

The above approaches used the Mindstorms RCX robots but it can equally be done with the newer NXT robots. 

 All opinions in this blog are the Author's and should not in any way be seen as reflecting the views of any organisation the Author has any association with.