So I wrote bucket sort implementation (I use rnd.NextDouble() to fill my array, so all the elements are in the range between 0 and 1). And I have number of experiments (10 experiments for each array size) and different array sizes (I start with 1000 and then +1000, etc. till 300.000). And sometimes it works fine, but sometimes it gives me OutOfRange exception:
public static void BucketSortImplement(ref int countOfElements, ref float[] array)
{
List<float>[] buckets = new List<float>[countOfElements];
for (int i = 0; i < countOfElements; i++)
{
buckets[i] = new List<float>();
}
for (int i = 0; i < countOfElements; i++)
{
float indexOfElement = array[i] * countOfElements;
buckets[(int)indexOfElement].Add(array[i]); // right here
}
for (int i = 0; i < countOfElements; i++)
{
for (int j = 0; j < buckets[i].Count; j++)
{
float keyElement = buckets[i][j];
int k = j - 1;
while (k >= 0 && buckets[i][k] > keyElement)
{
buckets[i][k + 1] = buckets[i][k];
k -= 1;
}
buckets[i][k + 1] = keyElement;
}
}
int arrayIndex = 0;
for (int i = 0; i < countOfElements; i++)
{
for (int j = 0; j < buckets[i].Count; j++)
{
array[arrayIndex++] = buckets[i][j];
}
}
}
I am a bit confused, because the algorithm itself looks fine, that is I have to calculate array[i] * countOfElements to get the index where I can put my element. Could you please direct me?
If you array contain value '1' it will lead to OutOfRange, because if you have a size of the list equal to 3 (for example) then valid indexes will be in the range [0, 2].
I'm coming back to programming after having done none for several years, and created a Sudoku game to get my feet wet again. I've written a recursive function to brute-force a solution, and it will do so for simple board states, but runs into a stack overflow most of the time for more complicated board states. I know that this could be averted using loops or a more efficient algorithm, but since this is for the sake of learning I want to know how to allocate more memory to the stack (acknowledging this isn't the best solution to the problem).
An answer here: How to change stack size for a .NET program? that recommends using a thread with more memory allocated isn't working, as from what I can read .NET 4.0 will not let you increase a thread's maximum allocated memory beyond the default.
Another solution recommends using EDITBIN.EXE, but I am new to Visual Studio and have not found an explanation I understand of how to do so.
Similarly, I've found that you can use a .def file to indicate a larger default stack size, but have not found an explanation I understand of how to create/implement one.
Can anyone offer newbie-level explanations of either EDITBIN.EXE or .def file implementations, or offer another way to increase the stack size?
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
using System.IO;
namespace Sudoku
{
//object keeps track of the value of all numbers currently on the board using an array
class BoardState
{
int testcount = 1;
//3d array of ints representing all values on the board, represted as region, column, row
int[,,] boardVals;
//3d array of bools representing if a number is immutable (true cannot be changed, false can be)
bool[,,] fixedVals;
//create a blank board if no initial values are provided
public BoardState()
{
boardVals = new int[9, 3, 3];
fixedVals = new bool[9, 3, 3];
}
//create a board with the listed initial values as immutable
public BoardState(int[,,] inputVals)
{
boardVals = inputVals;
fixedVals = new bool[9,3,3];
for (int i = 0; i < 9; ++i)
for (int j = 0; j < 3; ++j)
for (int k = 0; k < 3; ++k)
if (boardVals[i, j, k] > 0) fixedVals[i, j, k] = true;
}
//update the state of the board using the coordinates of a single value
//**note** method has not been implemented and tested yet
public void updateState(int region, int column, int row, int val)
{
if (!fixedVals[region, column, row])
{
boardVals[region, column, row] = val;
}
}
//update the state of the board to match the state of another board
public void updateState(int[,,] newState)
{
boardVals = newState;
}
public int[,,] getVals()
{
return boardVals;
}
public bool[,,] getFixed()
{
return fixedVals;
}
//set all non-zero values to be immutable
public void setFixed()
{
for (int i = 0; i < 9; i++)
for (int j = 0; j < 3; j++)
for (int k = 0; k < 3; k++) {
if (boardVals[i, j, k] != 0)
fixedVals[i, j, k] = true;
else
fixedVals[i, j, k] = false;
}
}
//test method
public void testState()
{
for (int i = 0; i < 9; i++)
for (int j = 0; j < 3; j++)
for (int k = 0; k < 3; k++)
Console.WriteLine(boardVals[i, k, j]);
}
//accepts a 3d array representing the current board state.
//returns a 3d bool array denoting whether any region, row, or column is invalid (true=invalid)
//first value of array designates the region, row, or column respectively
//second value designates which of those is invalid
public bool[,] validateBoard()
{
bool[,] valid = new bool[3, 9];
int[,] rows = makeRows(boardVals);
int[,] cols = makeCols(boardVals);
//compare each value in each row to each other value in that row
for (int i = 0; i < 9; i++)
{
for (int j = 0; j < 9; j++)
{
//only validate an entry if it has been assigned a value
if (rows[i, j] != 0)
{
for (int k = 0; k < 9; k++)
{
//if two values are not the same entry and are equal, set that entry to invalid
if (j != k && rows[i, j] == rows[i, k])
valid[1, i] = true;
}
}
}
}
//compare each value in each column to each other value in that column
for (int i = 0; i < 9; i++)
{
for (int j = 0; j < 9; j++)
{
//only validate an entry if it has been assigned a value
if (cols[i, j] != 0)
{
for (int k = 0; k < 9; k++)
{
//if two values are not the same entry and are equal, set that entry to invalid
if (j != k && cols[i, j] == cols[i, k])
valid[2, i] = true;
}
}
}
}
//compare each value in each region to each other value in that region
for (int i = 0; i < 9; i++)
{
for (int j = 0; j < 3; j++)
{
for (int k = 0; k < 3; k++)
{
//only validate an entry if it has been assigned a value
if (boardVals[i, j, k] != 0)
{
for (int l = 0; l < 3; l++)
{
for (int m = 0; m < 3; m++)
{
//if two values are not the same entry and are equal, set that entry to invalid
if (!(l == j && m == k) && boardVals[i, j, k] == boardVals[i, l, m])
valid[0, i] = true;
}
}
}
}
}
}
return valid;
}
public bool isValid()
{
bool retFlag = true;
bool[,] valid = new bool[3, 9];
int[,] rows = makeRows(boardVals);
int[,] cols = makeCols(boardVals);
//for (int i = 0; i < 9; i++)
// for (int j = 0; j < 3; j++)
// for (int k = 0; k < 3; k++)
// Console.Write(boardVals[i, j, k]);
//compare each value in each row to each other value in that row
for (int i = 0; i < 9; i++)
{
for (int j = 0; j < 9; j++)
{
//only validate an entry if it has been assigned a value
if (rows[i, j] != 0)
{
for (int k = 0; k < 9; k++)
{
//if two values are not the same entry and are equal, set that entry to invalid
if (j != k && rows[i, j] == rows[i, k])
{
retFlag = false;
}
}
}
}
}
//compare each value in each column to each other value in that column
for (int i = 0; i < 9; i++)
{
for (int j = 0; j < 9; j++)
{
//only validate an entry if it has been assigned a value
if (cols[i, j] != 0)
{
for (int k = 0; k < 9; k++)
{
//if two values are not the same entry and are equal, set that entry to invalid
if (j != k && cols[i, j] == cols[i, k])
{
retFlag = false;
}
}
}
}
}
//compare each value in each region to each other value in that region
for (int i = 0; i < 9; i++)
{
for (int j = 0; j < 3; j++)
{
for (int k = 0; k < 3; k++)
{
//only validate an entry if it has been assigned a value
if (boardVals[i, j, k] != 0)
{
for (int l = 0; l < 3; l++)
{
for (int m = 0; m < 3; m++)
{
//if two values are not the same entry and are equal, set that entry to invalid
if (!(l == j && m == k) && boardVals[i, j, k] == boardVals[i, l, m])
{
retFlag = false;
}
}
}
}
}
}
}
return retFlag;
}
//returns an array of all the values in each row
public int[,] makeRows(int[,,] boardState)
{
int[,] rows = new int[9, 9];
for (int i = 0; i < 9; i++)
for (int j = 0; j < 9; j++)
rows[i, j] = boardState[j / 3 + ((i / 3) * 3), j % 3, i - ((i / 3) * 3)];
return rows;
}
//returns an array of all values in each column
public int[,] makeCols(int[,,] boardState)
{
int[,] cols = new int[9, 9];
for (int i = 0; i < 9; i++)
for (int j = 0; j < 9; j++)
cols[i, j] = boardState[((j / 3) * 3) + (i / 3), i - ((i / 3) * 3), j % 3];
return cols;
}
//update the board state to a state read in from a file
public void updateFromFile(Stream update)
{
int[,,] newVals = new int[9, 3, 3];
int[,,] newFixed = new int[9, 3, 3];
StreamReader file = new StreamReader(update);
for (int i = 0; i < 9; i++){
for (int j = 0; j < 3; j++){
for (int k = 0; k < 3; k++){
boardVals[i, j, k] = (int)char.GetNumericValue((char)file.Read());
}
}
}
for (int i = 0; i < 9; i++){
for (int j = 0; j < 3; j++){
for (int k = 0; k < 3; k++){
fixedVals[i, j, k] = (0 != ((int)char.GetNumericValue((char)file.Read())));
}
}
}
file.Close();
}
public void Solve(int entry, int val)
{
Console.WriteLine("This is call number " + ++testcount);
//returns if all values are filled and valid
if (entry == 81)
{
Console.WriteLine("Solved!");
return;
}
//creating reference coordinates based on entry value
int reg = entry / 9;
int col = (entry - (reg * 9)) % 3;
int row = (entry - (reg * 9)) / 3;
//if current entry being checked is a fixed value, go the next value
if (!fixedVals[reg, row, col])
{
Console.WriteLine("");
Console.WriteLine("Making an attempt at entry " + entry + " using value " + val);
Console.WriteLine("This entry is at region " + reg + ", col " + col + ", row " + row);
//assign entry the value to be tested
boardVals[reg, row, col] = val;
//if the value is valid, go to the next entry
if (isValid())
{
Console.WriteLine("Entry Valid at " + entry);
val = 1;
entry++;
Console.WriteLine("Trying the next entry at " + entry);
Solve(entry, val);
}
//if the value is invlid and all 9 values have not been tried,
//increment value and call again at same entry
if (!isValid() && val < 9)
{
Console.WriteLine("Entry Invalid at " + entry + " with value " + val);
++val;
Console.WriteLine("Trying again with value " + val);
Solve(entry, val);
}
//if the value in invalid and all 9 values have been tried,
//zero out the entry and go back to the previous non-fixed entry
if (!isValid() && val == 9)
{
do
{
boardVals[reg, row, col] = 0;
Console.WriteLine("Reached Value 9 and was still invalid");
--entry;
Console.WriteLine("Trying again at entry " + entry);
Console.WriteLine("The value at that entry is " + boardVals[reg, row, col]);
reg = entry / 9;
col = (entry - reg * 9) % 3;
row = (entry - reg * 9) / 3;
if (fixedVals[reg, row, col])
Console.WriteLine("But that's a fixed value, so I'll go back one more");
Console.WriteLine("");
} while (boardVals[reg, row, col] == 9 || fixedVals[reg, row, col]);
val = boardVals[reg, row, col] + 1;
Solve(entry, val);
}
}
else Solve(++entry, val);
}
}
}
The big bad warning
If you use recursion in a program and reach a point where having a StackOverflowException is an actual threat, please do not consider increasing the stack size as a valid solution.
If you encounter a StackOverflowException you are doing something very wrong; you should instead be using a Stack<T> for depth-first processing, or a Queue<T> for breadth-first processing. Example.
The solution
This can be achieved by using editbin.exe, which is installed with this package;
Find the location of editbin.exe, mine was located at C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.14.26428\bin\Hostx64\x64\editbin.exe, I would suggest using Everything by voidtools in lieu of Microsoft's awful search to find this.
Set the stack size manually
Navigate to your bin folder and execute the following:
"<full path of editbin.exe>" /stack:<stack size in bytes, decimal> <your executable name>
For example I executed this:
"C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.14.26428\bin\Hostx64\x64\EDITBIN.EXE" /stack:2097152 ExampleProgram.exe
Which set the stack reserve size to 2MB.
With this I was capable of reaching twice the recursion level; (1MB stack reserve on left, 2MB stack reserve on right).
Set the stack size automatically
Right click on your project and select 'Options', then click on 'Build Events' and add the following to your post-build events:
"<full path of editbin.exe>" /stack:<stack size in bytes, decimal> "$(TargetPath)"
For example I added
"C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.14.26428\bin\Hostx64\x64\EDITBIN.EXE" /stack:2097152 "$(TargetPath)"
This will run editbin.exe every time you build your executable.
Note: You will see a lot lower level of recursion reached when running your program from Visual Studio as you will from running it explicitly via explorer or cmd. You will still however see a 2x increase in the level of recursion met if moving from a 1MB stack reserve to a 2MB stack reserve.
Perhaps set the Stack Reserve Size in Visual Studio:
Project -> Properties -> Configuration Properties -> Linker -> System -> Stack Reserve Size
This can also be done via the command-line or programatically when a thread is created:
https://learn.microsoft.com/en-us/cpp/build/reference/stack-stack-allocations?view=vs-2017
I come looking for general tips about the program I'm writing now.
The goal is:
Use neural network program to recognize 3 letters [D,O,M] (or display "nothing is recognized" if i input anything other than those 3).
Here's what I have so far:
A class for my single neuron
public class neuron
{
double[] weights;
public neuron()
{
weights = null;
}
public neuron(int size)
{
weights = new double[size + 1];
Random r = new Random();
for (int i = 0; i <= size; i++)
{
weights[i] = r.NextDouble() / 5 - 0.1;
}
}
public double output(double[] wej)
{
double s = 0.0;
for (int i = 0; i < weights.Length; i++) s += weights[i] * wej[i];
s = 1 / (1 + Math.Exp(s));
return s;
}
}
A class for a layer:
public class layer
{
neuron[] tab;
public layer()
{
tab = null;
}
public layer(int numNeurons, int numInputs)
{
tab = new neuron[numNeurons];
for (int i = 0; i < numNeurons; i++)
{
tab[i] = new neuron(numInputs);
}
}
public double[] compute(double[] wejscia)
{
double[] output = new double[tab.Length + 1];
output[0] = 1;
for (int i = 1; i <= tab.Length; i++)
{
output[i] = tab[i - 1].output(wejscia);
}
return output;
}
}
And finally a class for a network
public class network
{
layer[] layers = null;
public network(int numLayers, int numInputs, int[] npl)
{
layers = new layer[numLayers];
for (int i = 0; i < numLayers; i++)
{
layers[i] = new layer(npl[i], (i == 0) ? numInputs : (npl[i - 1]));
}
}
double[] compute(double[] inputs)
{
double[] output = layers[0].compute(inputs);
for (int i = 1; i < layers.Length; i++)
{
output = layers[i].compute(output);
}
return output;
}
}
Now for the algorythm I chose:
I have a picture box, size 200x200, where you can draw a letter (or read one from jpg file).
I then convert it to my first array(get the whole picture) and 2nd one(cut the non relevant background around it) like so:
Bitmap bmp2 = new Bitmap(this.pictureBox1.Image);
int[,] binaryfrom = new int[bmp2.Width, bmp2.Height];
int minrow=0, maxrow=0, mincol=0, maxcol=0;
for (int i = 0; i < bmp2.Height; i++)
{
for (int j = 0; j < bmp2.Width; j++)
{
if (bmp2.GetPixel(j, i).R == 0)
{
binaryfrom[i, j] = 1;
if (minrow == 0) minrow = i;
if (maxrow < i) maxrow = i;
if (mincol == 0) mincol = j;
else if (mincol > j) mincol = j;
if (maxcol < j) maxcol = j;
}
else
{
binaryfrom[i, j] = 0;
}
}
}
int[,] boundaries = new int[binaryfrom.GetLength(0)-minrow-(binaryfrom.GetLength(0)-(maxrow+1)),binaryfrom.GetLength(1)-mincol-(binaryfrom.GetLength(1)-(maxcol+1))];
for(int i = 0; i < boundaries.GetLength(0); i++)
{
for(int j = 0; j < boundaries.GetLength(1); j++)
{
boundaries[i, j] = binaryfrom[i + minrow, j + mincol];
}
}
And convert it to my final array of 12x8 like so (i know I could shorten this a fair bit, but wanted to have every step in different loop so I can see what went wrong easier[if anything did]):
int[,] finalnet = new int[12, 8];
int k = 1;
int l = 1;
for (int i = 0; i < finalnet.GetLength(0); i++)
{
for (int j = 0; j < finalnet.GetLength(1); j++)
{
finalnet[i, j] = 0;
}
}
while (k <= finalnet.GetLength(0))
{
while (l <= finalnet.GetLength(1))
{
for (int i = (int)(boundaries.GetLength(0) / finalnet.GetLength(0)) * (k - 1); i < (int)(boundaries.GetLength(0) / finalnet.GetLength(0)) * k; i++)
{
for (int j = (int)(boundaries.GetLength(1) / finalnet.GetLength(1)) * (l - 1); j < (int)(boundaries.GetLength(1) / finalnet.GetLength(1)) * l; j++)
{
if (boundaries[i, j] == 1) finalnet[k-1, l-1] = 1;
}
}
l++;
}
l = 1;
k++;
}
int a = boundaries.GetLength(0);
int b = finalnet.GetLength(1);
if((a%b) != 0){
k = 1;
while (k <= finalnet.GetLength(1))
{
for (int i = (int)(boundaries.GetLength(0) / finalnet.GetLength(0)) * finalnet.GetLength(0); i < boundaries.GetLength(0); i++)
{
for (int j = (int)(boundaries.GetLength(1) / finalnet.GetLength(1)) * (k - 1); j < (int)(boundaries.GetLength(1) / finalnet.GetLength(1)) * k; j++)
{
if (boundaries[i, j] == 1) finalnet[finalnet.GetLength(0) - 1, k - 1] = 1;
}
}
k++;
}
}
if (boundaries.GetLength(1) % finalnet.GetLength(1) != 0)
{
k = 1;
while (k <= finalnet.GetLength(0))
{
for (int i = (int)(boundaries.GetLength(0) / finalnet.GetLength(0)) * (k - 1); i < (int)(boundaries.GetLength(0) / finalnet.GetLength(0)) * k; i++)
{
for (int j = (int)(boundaries.GetLength(1) / finalnet.GetLength(1)) * finalnet.GetLength(1); j < boundaries.GetLength(1); j++)
{
if (boundaries[i, j] == 1) finalnet[k - 1, finalnet.GetLength(1) - 1] = 1;
}
}
k++;
}
for (int i = (int)(boundaries.GetLength(0) / finalnet.GetLength(0)) * finalnet.GetLength(0); i < boundaries.GetLength(0); i++)
{
for (int j = (int)(boundaries.GetLength(1) / finalnet.GetLength(1)) * finalnet.GetLength(1); j < boundaries.GetLength(1); j++)
{
if (boundaries[i, j] == 1) finalnet[finalnet.GetLength(0) - 1, finalnet.GetLength(1) - 1] = 1;
}
}
}
The result is a 12x8 (I can change it in the code to get it from some form controls) array of 0 and 1, where 1 form the rough shape of a letter you drawn.
Now my questions are:
Is this a correct algorythm?
Is my function
1/(1+Math.Exp(x))
good one to use here?
What should be the topology? 2 or 3 layers, and if 3, how many neurons in hidden layer? I have 96 inputs (every field of the finalnet array), so should I also take 96 neurons in the first layer? Should I have 3 neurons in the final layer or 4(to take into account the "not recognized" case), or is it not necessary?
Thank you for your help.
EDIT: Oh, and I forgot to add, I'm gonna train my network using Backpropagation algorythm.
You may need 4 layers at least to get accurate results using back propagation method. 1 input, 2 middle layers, and an output layer.
12 * 8 matrix is too small(and you may end up in data loss which will result in total failure) - try some thing 16 * 16. If you want to reduce the size then you have to peel out the outer layers of black pixels further.
Think about training the network with your reference characters.
Remember that you have to feed back the output back to the input layer again and iterate it multiple times.
A while back I created a neural net to recognize digits 0-9 (python, sorry), so based on my (short) experience, 3 layers are ok and 96/50/3 topology will probably do a good job. As for the output layer, it's your choice; you can either backpropagate all 0s when the input image is not a D, O or M or use the fourth output neuron to indicate that the letter was not recognized. I think that the first option would be the best one because it's simpler (shorter training time, less problems debugging the net...), you just need to apply a threshold under which you classify the image as 'not recognized'.
I also used the sigmoid as activation function, I didn't try others but it worked :)
I'm trying to make a program that: When given a 4 digit number (for example, 1001) it sums the digits of that number, for 1001 this is 1 + 0 + 0 + 1 = 2, than it finds all sequences of 6 numbers from 1 to 6 (including permutations, i.e. 1*1*1*1*1*2 is different than 2*1*1*1*1*1) whose product is that number.
The result should be printed on the console in the following format: each sequence of 6 numbers, with their Morse representation, separated with a single pipe: 1 is .---- , 2 is ..---: .----|.----|.----|.----|..---|, on a new row the next permutation: .----|.----|.----|..---|.----| and so on.
The problem is, my solution doesn't show the correct answers, not even the correct number of them.
Here's my code (and please, if possible, tell me where my mistake is, and not some one line hack solutions to the problem with LINQ and regex and God knows what):
string n = Console.ReadLine();
string[] digitsChar = new string[n.Length];
for (int i = 0; i < 4; i++)
{
digitsChar[i] = n[i].ToString();
}
int[] digits = new int[digitsChar.Length];
for (int i = 0; i < 4; i++)
{
digits[i] = Convert.ToInt32(digitsChar[i]);
}
int morseProduct = digits.Sum();
Console.WriteLine(morseProduct);
List<int> morseCodeNumbers = new List<int>();
for (int i = 1; i < 6; i++)
{
for (int j = 1; i < 6; i++)
{
for (int k = 1; i < 6; i++)
{
for (int l = 1; i < 6; i++)
{
for (int m = 1; i < 6; i++)
{
for (int o = 1; o < 6; o++)
{
int product = i * j * k * l * m * o;
if (product == morseProduct)
{
morseCodeNumbers.Add(i);
morseCodeNumbers.Add(j);
morseCodeNumbers.Add(k);
morseCodeNumbers.Add(l);
morseCodeNumbers.Add(m);
morseCodeNumbers.Add(o);
}
}
}
}
}
}
}
int numberOfNumbers = morseCodeNumbers.Count;
string[] morseCodes = new string[] { "-----", ".----", "..---", "...--", "....-", "....." };
for (int i = 0; i < numberOfNumbers; i++)
{
int counter = 0;
if (i % 5 == 0)
{
Console.WriteLine();
counter = 0;
}
if (counter < 5)
{
int index = morseCodeNumbers[i];
Console.Write(morseCodes[index] + "|");
counter++;
}
A lot of your for-loop conditions refer to i instead of j,k,l and m. The same for the increment part. For example:
for (int j = 1; i < 6; i++)
should be
for (int j = 1; j < 6; j++)
Furthermore if the range is from 1 to 6 you need to change < to <=, see:
for (int i = 1; i <= 6; i++)
You don't need to convert the string to a string array to get the int array of digits btw, so while this is correct:
for (int i = 0; i < 4; i++)
{
digitsChar[i] = n[i].ToString();
}
int[] digits = new int[digitsChar.Length];
for (int i = 0; i < 4; i++)
{
digits[i] = Convert.ToInt32(digitsChar[i]);
}
you could it do like that (sry for LINQ):
var digits = n.Select(c=>(int)char.GetNumericValue(c) ).ToArray();
Is there any relation between number of neurons and ability of Hopfield network to recognize patterns?
I write neural network program in C# to recognize patterns with Hopfield network. My network has 64 neurons. When I train network for 2 patterns, every things work nice and easy, but when I train network for more patterns, Hopfield can't find answer!
So, according to my code, how can I use Hopfield network to learn more patterns?
Should I make changes in this code?
There is my train() function:
public void Train(bool[,] pattern)
{
//N is number of rows in our square matrix
//Convert input pattern to bipolar
int[,] PatternBipolar = new int[N, N];
for (int i = 0; i < N; i++)
for (int j = 0; j < N; j++)
{
if (pattern[i, j] == true)
{
PatternBipolar[i, j] = 1;
}
else
{
PatternBipolar[i, j] = -1;
}
}
//convert to row matrix
int count1 = 0;
int[] RowMatrix = new int[(int)Math.Pow(N, 2)];
for (int j = 0; j < N; j++)
for (int i = 0; i < N; i++)
{
RowMatrix[count1] = PatternBipolar[i, j];
count1++;
}
//convert to column matrix
int count2 = 0;
int[] ColMatrix = new int[(int)Math.Pow(N, 2)];
for (int j = 0; j < N; j++)
for (int i = 0; i < N; i++)
{
ColMatrix[count2] = PatternBipolar[i, j];
count2++;
}
//multiplication
int[,] MultipliedMatrix = new int[(int)Math.Pow(N, 2), (int)Math.Pow(N, 2)];
for (int i = 0; i < (int)Math.Pow(N, 2); i++)
for (int j = 0; j < (int)Math.Pow(N, 2); j++)
{
MultipliedMatrix[i, j] = ColMatrix[i] * RowMatrix[j];
}
//cells in the northwest diagonal get set to zero
for (int i = 0; i < (int)Math.Pow(N, 2); i++)
MultipliedMatrix[i, i] = 0;
// WightMatrix + MultipliedMatrix
for (int i = 0; i < (int)Math.Pow(N, 2); i++)
for (int j = 0; j < (int)Math.Pow(N, 2); j++)
{
WeightMatrix[i, j] += MultipliedMatrix[i, j];
}
And there is Present() function (this function is used to return answer for a given pattern):
public void Present(bool[,] Pattern)
{
int[] output = new int[(int)(int)Math.Pow(N, 2)];
for (int j = 0; j < N; j++)
for (int i = 0; i < N; i++)
{
OutputShowMatrix[i, j] = 0;
}
//convert bool to binary
int[] PatternBinary = new int[(int)Math.Pow(N, 2)];
int count = 0;
for (int j = 0; j < N; j++)
for (int i = 0; i < N; i++)
{
if (Pattern[i, j] == true)
{
PatternBinary[count] = 1;
}
else
{
PatternBinary[count] = 0;
}
count++;
}
count = 0;
int activation = 0;
for (int j = 0; j < (int)Math.Pow(N, 2); j++)
{
for (int i = 0; i < (int)Math.Pow(N, 2); i++)
{
activation = activation + (PatternBinary[i] * WeightMatrix[i, j]);
}
if (activation > 0)
{
output[count] = 1;
}
else
{
output[count] = 0;
}
count++;
activation = 0;
}
count = 0;
for (int j = 0; j < N; j++)
for (int i = 0; i < N; i++)
{
OutputShowMatrix[i, j] = output[count++];
}
In below images I trained Hopfield for characters A and P and when input patterns are like A or P, network recognize them in true way
Then I train it for character C:
This is where every things go wrong!
Now if I enter pattern like C, this issue happen:
And if enter pattern like A, see what happen:
And if train more patterns, whole of grid become black!
I've spotted only one mistake in your code: you perform only one iteration of node value calculation, without verifying if the values have converged. I've fixed this method like this:
public bool[,] Present(bool[,] pattern)
{
bool[,] result = new bool[N, N];
int[] activation = new int[N * N];
int count = 0;
for (int i = 0; i < N; i++)
for (int j = 0; j < N; j++)
{
activation[count++] = pattern[i, j] ? 1 : 0;
}
bool convergence = false;
while (!convergence)
{
convergence = true;
var previousActivation = (int[])activation.Clone();
for (int i = 0; i < N * N; i++)
{
activation[i] = 0;
for (int j = 0; j < N * N; j++)
{
activation[i] += (previousActivation[j] * WeightMatrix[i, j]);
}
activation[i] = activation[i] > 0 ? 1 : 0;
if (activation[i] != previousActivation[i])
{
convergence = false;
}
}
}
count = 0;
for (int i = 0; i < N; i++)
for (int j = 0; j < N; j++)
{
result[i, j] = activation[count++] == 1;
}
return result;
}
This slightly improves the results, however probably should also be improved to calculate the values asynchronously to avoid cycles.
Unfortunately, this still introduces the behaviour you've described. This is results from the phenomena called spurious patterns. For the network to learn more than one pattern consider training it with a Hebb rule. You can read about the spurious patterns, stability and learning of the Hopfield network here and here.