public void LearningInLayerTest(int width, int height, string imageName)
{
// Prepare test output folder
var outFolder = EnsureFolderExist(nameof(ImageEncoderTest));
// Prepare input file for test
string inputImage = Path.Combine("TestFiles", imageName);
// Initialize Image Encoder
ImageEncoder encoder = new ImageEncoder(new BinarizerParams {
ImageWidth = width, ImageHeight = height
});
// Initialize HTMModules
int inputBits = width * height;
int numColumns = 1024;
HtmConfig cfg = new HtmConfig(new int[] { inputBits }, new int[] { numColumns });
var mem = new Connections(cfg);
SpatialPoolerMT sp = new SpatialPoolerMT();
sp.Init(mem);
CortexLayer <object, object> layer1 = new CortexLayer <object, object>("L1");
layer1.HtmModules.Add("encoder", encoder);
layer1.HtmModules.Add("sp", sp);
//Test Compute method
var computeResult = layer1.Compute(inputImage, true) as int[];
var activeCellList = GetActiveCells(computeResult);
Debug.WriteLine($"Active Cells computed from Image {inputImage}: {activeCellList}");
}
public void SpatialSequenceLearningExperiment()
{
var parameters = GetDefaultParams();
parameters.Set(KEY.POTENTIAL_RADIUS, 64 * 64);
parameters.Set(KEY.POTENTIAL_PCT, 1.0);
parameters.Set(KEY.GLOBAL_INHIBITION, false);
parameters.Set(KEY.STIMULUS_THRESHOLD, 0.5);
parameters.Set(KEY.INHIBITION_RADIUS, (int)0.25 * 64 * 64);
parameters.Set(KEY.LOCAL_AREA_DENSITY, -1);
parameters.Set(KEY.NUM_ACTIVE_COLUMNS_PER_INH_AREA, 0.1 * 64 * 64);
parameters.Set(KEY.DUTY_CYCLE_PERIOD, 1000000);
parameters.Set(KEY.MAX_BOOST, 5);
parameters.setInputDimensions(new int[] { 32, 32 });
parameters.setColumnDimensions(new int[] { 64, 64 });
parameters.setNumActiveColumnsPerInhArea(0.02 * 64 * 64);
var sp = new SpatialPoolerMT();
var mem = new Connections();
double[] inputSequence = new double[] { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0 };
var inputVectors = GetEncodedSequence(inputSequence, 0.0, 100.0);
parameters.apply(mem);
sp.Init(mem, UnitTestHelpers.GetMemory());
foreach (var inputVector in inputVectors)
{
for (int i = 0; i < 3; i++)
{
var activeIndicies = sp.Compute(inputVector, true, true) as int[];
var activeArray = sp.Compute(inputVector, true, false) as int[];
Debug.WriteLine(Helpers.StringifyVector(activeArray));
Debug.WriteLine(Helpers.StringifyVector(activeIndicies));
}
}
}
public void TestCortexLayer()
{
int inputBits = 100;
double max = 20;
int numColumns = 2048;
List <double> inputValues = new List <double>(new double[] { 0.0, 1.0, 0.0, 2.0, 3.0, 4.0, 5.0, 6.0, 5.0, 4.0, 3.0, 7.0, 1.0, 9.0, 12.0, 11.0, 12.0, 13.0, 14.0, 11.0, 12.0, 14.0, 5.0, 7.0, 6.0, 9.0, 3.0, 4.0, 3.0, 4.0, 3.0, 4.0 });
int numInputs = inputValues.Distinct().ToList().Count;
var inputs = inputValues.ToArray();
Dictionary <string, object> settings = new Dictionary <string, object>()
{
{ "W", 15 },
{ "N", inputBits },
{ "Radius", -1.0 },
{ "MinVal", 0.0 },
{ "Periodic", false },
{ "Name", "scalar" },
{ "ClipInput", false },
{ "MaxVal", max }
};
EncoderBase encoder = new ScalarEncoder(settings);
HtmConfig htmConfig = new HtmConfig(new int[] { inputBits }, new int[] { numColumns })
{
Random = new ThreadSafeRandom(42),
CellsPerColumn = 25,
GlobalInhibition = true,
LocalAreaDensity = -1,
NumActiveColumnsPerInhArea = 0.02 * numColumns,
PotentialRadius = 50,
InhibitionRadius = 15,
MaxBoost = 10.0,
DutyCyclePeriod = 25,
MinPctOverlapDutyCycles = 0.75,
MaxNewSynapseCount = (int)(0.02 * numColumns),
ActivationThreshold = 15,
ConnectedPermanence = 0.5,
PermanenceDecrement = 0.25,
PermanenceIncrement = 0.15,
PredictedSegmentDecrement = 0.1
};
Connections memory = new Connections(htmConfig);
HomeostaticPlasticityController hpa = new HomeostaticPlasticityController(memory, numInputs * 55, (isStable, numPatterns, actColAvg, seenInputs) =>
{
if (isStable)
{
// Event should be fired when entering the stable state.
Debug.WriteLine($"STABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
else
{
// Ideal SP should never enter unstable state after stable state.
Debug.WriteLine($"INSTABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
}, numOfCyclesToWaitOnChange: 25);
SpatialPoolerMT spatialPooler = new SpatialPoolerMT(hpa);
spatialPooler.Init(memory, UnitTestHelpers.GetMemory());
TemporalMemory temporalMemory = new TemporalMemory();
temporalMemory.Init(memory);
List <CortexRegion> regions = new List <CortexRegion>();
CortexRegion region0 = new CortexRegion("1st Region");
regions.Add(region0);
CortexLayer <object, object> layer1 = new CortexLayer <object, object>("L1");
region0.AddLayer(layer1);
layer1.HtmModules.Add("encoder", encoder);
layer1.HtmModules.Add("sp", spatialPooler);
layer1.HtmModules.Add("tm", temporalMemory);
bool learn = true;
int maxCycles = 3500;
for (int i = 0; i < maxCycles; i++)
{
foreach (var input in inputs)
{
var lyrOut = layer1.Compute(input, learn) as ComputeCycle;
}
}
}
public void InputBitsExperiment(int W, int InputB, int loop)
{
string filename = "InputBits" + InputB + ".csv";
using (StreamWriter writer = new StreamWriter(filename))
{
Debug.WriteLine($"Learning Cycles: {460}");
Debug.WriteLine("Cycle;Similarity");
//Parent Loop
for (int j = 0; j < loop; j++)
{
int inputBits = InputB;
bool learn = true;
Parameters p = Parameters.getAllDefaultParameters();
p.Set(KEY.RANDOM, new ThreadSafeRandom(42));
p.Set(KEY.INPUT_DIMENSIONS, new int[] { inputBits });
p.Set(KEY.CELLS_PER_COLUMN, 5);
p.Set(KEY.COLUMN_DIMENSIONS, new int[] { 500 });
CortexNetwork net = new CortexNetwork("my cortex");
List <CortexRegion> regions = new List <CortexRegion>();
CortexRegion region0 = new CortexRegion("1st Region");
regions.Add(region0);
SpatialPoolerMT sp1 = new SpatialPoolerMT();
TemporalMemory tm1 = new TemporalMemory();
var mem = new Connections();
p.apply(mem);
sp1.Init(mem, UnitTestHelpers.GetMemory());
tm1.Init(mem);
Dictionary <string, object> settings = new Dictionary <string, object>()
{
{ "W", W },
{ "N", inputBits },
{ "Radius", -1.0 },
{ "MinVal", 0.0 },
// { "MaxVal", 20.0 },
{ "Periodic", false },
{ "Name", "scalar" },
{ "ClipInput", false },
};
double max = 10;
List <double> lst = new List <double>();
for (double i = max - 1; i >= 0; i--)
{
lst.Add(i);
}
settings["MaxVal"] = max;
EncoderBase encoder = new ScalarEncoder(settings);
CortexLayer <object, object> layer1 = new CortexLayer <object, object>("L1");
//
// NewBorn learning stage.
region0.AddLayer(layer1);
layer1.HtmModules.Add("encoder", encoder);
layer1.HtmModules.Add("sp", sp1);
HtmClassifier <double, ComputeCycle> cls = new HtmClassifier <double, ComputeCycle>();
double[] inputs = lst.ToArray();
//
// This trains SP.
foreach (var input in inputs)
{
Debug.WriteLine($" ** {input} **");
for (int i = 0; i < 3; i++)
{
var lyrOut = layer1.Compute(input, learn) as ComputeCycle;
}
}
// Here we add TM module to the layer.
layer1.HtmModules.Add("tm", tm1);
int cycle = 0;
int matches = 0;
double lastPredictedValue = 0;
//
// Now, training with SP+TM. SP is pretrained on pattern.
//Child loop
for (int i = 0; i < 460; i++)
{
matches = 0;
cycle++;
foreach (var input in inputs)
{
var lyrOut = layer1.Compute(input, learn) as ComputeCycle;
cls.Learn(input, lyrOut.ActiveCells.ToArray(), lyrOut.PredictiveCells.ToArray());
Debug.WriteLine($"-------------- {input} ---------------");
if (learn == false)
{
Debug.WriteLine($"Inference mode");
}
Debug.WriteLine($"W: {Helpers.StringifyVector(lyrOut.WinnerCells.Select(c => c.Index).ToArray())}");
Debug.WriteLine($"P: {Helpers.StringifyVector(lyrOut.PredictiveCells.Select(c => c.Index).ToArray())}");
var predictedValue = cls.GetPredictedInputValue(lyrOut.PredictiveCells.ToArray());
Debug.WriteLine($"Current Input: {input} \t| - Predicted value in previous cycle: {lastPredictedValue} \t| Predicted Input for the next cycle: {predictedValue}");
if (input == lastPredictedValue)
{
matches++;
Debug.WriteLine($"Match {input}");
}
else
{
Debug.WriteLine($"Missmatch Actual value: {input} - Predicted value: {lastPredictedValue}");
}
lastPredictedValue = predictedValue;
}
if (i == 500)
{
Debug.WriteLine("Stop Learning From Here. Entering inference mode.");
learn = false;
}
//tm1.reset(mem);
Debug.WriteLine($"Cycle: {cycle}\tMatches={matches} of {inputs.Length}\t {matches / (double)inputs.Length * 100.0}%");
if (matches / (double)inputs.Length == 1)
{
writer.WriteLine($"{cycle}");
break;
}
}
}
Debug.WriteLine("New Iteration");
}
//cls.TraceState();
Debug.WriteLine("------------------------------------------------------------------------\n----------------------------------------------------------------------------");
}
/// <summary>
///
/// </summary>
private void RunExperiment(int inputBits, Parameters p, EncoderBase encoder, List <double> inputValues)
{
Stopwatch sw = new Stopwatch();
sw.Start();
int maxMatchCnt = 0;
bool learn = true;
CortexNetwork net = new CortexNetwork("my cortex");
List <CortexRegion> regions = new List <CortexRegion>();
CortexRegion region0 = new CortexRegion("1st Region");
regions.Add(region0);
var mem = new Connections();
p.apply(mem);
//bool isInStableState = false;
//HtmClassifier<double, ComputeCycle> cls = new HtmClassifier<double, ComputeCycle>();
HtmClassifier <string, ComputeCycle> cls = new HtmClassifier <string, ComputeCycle>();
var numInputs = inputValues.Distinct().ToList().Count;
TemporalMemory tm1 = new TemporalMemory();
HomeostaticPlasticityController hpa = new HomeostaticPlasticityController(mem, numInputs * 55, (isStable, numPatterns, actColAvg, seenInputs) =>
{
if (isStable)
{
// Event should be fired when entering the stable state.
Debug.WriteLine($"STABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
else
{
// Ideal SP should never enter unstable state after stable state.
Debug.WriteLine($"INSTABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
Assert.IsTrue(numPatterns == numInputs);
//isInStableState = true;
cls.ClearState();
tm1.Reset(mem);
}, numOfCyclesToWaitOnChange: 25);
SpatialPoolerMT sp1 = new SpatialPoolerMT(hpa);
sp1.Init(mem, UnitTestHelpers.GetMemory());
tm1.Init(mem);
CortexLayer <object, object> layer1 = new CortexLayer <object, object>("L1");
region0.AddLayer(layer1);
layer1.HtmModules.Add("encoder", encoder);
layer1.HtmModules.Add("sp", sp1);
layer1.HtmModules.Add("tm", tm1);
double[] inputs = inputValues.ToArray();
int[] prevActiveCols = new int[0];
int cycle = 0;
int matches = 0;
string lastPredictedValue = "0";
Dictionary <double, List <List <int> > > activeColumnsLst = new Dictionary <double, List <List <int> > >();
foreach (var input in inputs)
{
if (activeColumnsLst.ContainsKey(input) == false)
{
activeColumnsLst.Add(input, new List <List <int> >());
}
}
int maxCycles = 3500;
int maxPrevInputs = inputValues.Count - 1;
List <string> previousInputs = new List <string>();
previousInputs.Add("-1.0");
//
// Now training with SP+TM. SP is pretrained on the given input pattern.
for (int i = 0; i < maxCycles; i++)
{
matches = 0;
cycle++;
Debug.WriteLine($"-------------- Cycle {cycle} ---------------");
foreach (var input in inputs)
{
Debug.WriteLine($"-------------- {input} ---------------");
var lyrOut = layer1.Compute(input, learn) as ComputeCycle;
var activeColumns = layer1.GetResult("sp") as int[];
activeColumnsLst[input].Add(activeColumns.ToList());
previousInputs.Add(input.ToString());
if (previousInputs.Count > maxPrevInputs + 1)
{
previousInputs.RemoveAt(0);
}
string key = GetKey(previousInputs, input);
List <Cell> actCells;
if (lyrOut.ActiveCells.Count == lyrOut.WinnerCells.Count)
{
actCells = lyrOut.ActiveCells;
}
else
{
actCells = lyrOut.WinnerCells;
}
cls.Learn(key, actCells.ToArray());
if (learn == false)
{
Debug.WriteLine($"Inference mode");
}
Debug.WriteLine($"Col SDR: {Helpers.StringifyVector(lyrOut.ActivColumnIndicies)}");
Debug.WriteLine($"Cell SDR: {Helpers.StringifyVector(actCells.Select(c => c.Index).ToArray())}");
if (key == lastPredictedValue)
{
matches++;
Debug.WriteLine($"Match. Actual value: {key} - Predicted value: {lastPredictedValue}");
}
else
{
Debug.WriteLine($"Missmatch! Actual value: {key} - Predicted value: {lastPredictedValue}");
}
if (lyrOut.PredictiveCells.Count > 0)
{
var predictedInputValue = cls.GetPredictedInputValue(lyrOut.PredictiveCells.ToArray());
Debug.WriteLine($"Current Input: {input} \t| Predicted Input: {predictedInputValue}");
lastPredictedValue = predictedInputValue;
}
else
{
Debug.WriteLine($"NO CELLS PREDICTED for next cycle.");
lastPredictedValue = string.Empty;
}
}
// The brain does not do that this way, so we don't use it.
// tm1.reset(mem);
double accuracy = matches / (double)inputs.Length * 100.0;
Debug.WriteLine($"Cycle: {cycle}\tMatches={matches} of {inputs.Length}\t {accuracy}%");
if (accuracy == 100.0)
{
maxMatchCnt++;
Debug.WriteLine($"100% accuracy reched {maxMatchCnt} times.");
if (maxMatchCnt >= 30)
{
sw.Stop();
Debug.WriteLine($"Exit experiment in the stable state after 30 repeats with 100% of accuracy. Elapsed time: {sw.ElapsedMilliseconds / 1000 / 60} min.");
learn = false;
//var testInputs = new double[] { 0.0, 2.0, 3.0, 4.0, 5.0, 6.0, 5.0, 4.0, 3.0, 7.0, 1.0, 9.0, 12.0, 11.0, 0.0, 1.0 };
// C-0, D-1, E-2, F-3, G-4, H-5
//var testInputs = new double[] { 0.0, 0.0, 4.0, 4.0, 5.0, 5.0, 4.0, 3.0, 3.0, 2.0, 2.0, 1.0, 1.0, 0.0 };
//// Traverse the sequence and check prediction.
//foreach (var input in inputValues)
//{
// var lyrOut = layer1.Compute(input, learn) as ComputeCycle;
// predictedInputValue = cls.GetPredictedInputValue(lyrOut.predictiveCells.ToArray());
// Debug.WriteLine($"I={input} - P={predictedInputValue}");
//}
/*
* //
* // Here we let the HTM predict sequence five times on its own.
* // We start with last predicted value.
* int cnt = 5 * inputValues.Count;
*
* Debug.WriteLine("---- Start Predicting the Sequence -----");
*
* //
* // This code snippet starts with some input value and tries to predict all next inputs
* // as they have been learned as a sequence.
* // We take a random value to start somwhere in the sequence.
* var predictedInputValue = inputValues[new Random().Next(0, inputValues.Count - 1)].ToString();
*
* List<string> predictedValues = new List<string>();
*
* while (--cnt > 0)
* {
* //var lyrOut = layer1.Compute(predictedInputValue, learn) as ComputeCycle;
* var lyrOut = layer1.Compute(double.Parse(predictedInputValue[predictedInputValue.Length - 1].ToString()), false) as ComputeCycle;
* predictedInputValue = cls.GetPredictedInputValue(lyrOut.PredictiveCells.ToArray());
* predictedValues.Add(predictedInputValue);
* };
*
* // Now we have a sequence of elements and watch in the trace if it matches to defined input set.
* foreach (var item in predictedValues)
* {
* Debug.Write(item);
* Debug.Write(" ,");
* }*/
break;
}
}
else if (maxMatchCnt > 0)
{
Debug.WriteLine($"At 100% accuracy after {maxMatchCnt} repeats we get a drop of accuracy with {accuracy}. This indicates instable state. Learning will be continued.");
maxMatchCnt = 0;
}
}
Debug.WriteLine("---- cell state trace ----");
cls.TraceState($"cellState_MinPctOverlDuty-{p[KEY.MIN_PCT_OVERLAP_DUTY_CYCLES]}_MaxBoost-{p[KEY.MAX_BOOST]}.csv");
Debug.WriteLine("---- Spatial Pooler column state ----");
foreach (var input in activeColumnsLst)
{
using (StreamWriter colSw = new StreamWriter($"ColumState_MinPctOverlDuty-{p[KEY.MIN_PCT_OVERLAP_DUTY_CYCLES]}_MaxBoost-{p[KEY.MAX_BOOST]}_input-{input.Key}.csv"))
{
Debug.WriteLine($"------------ {input.Key} ------------");
foreach (var actCols in input.Value)
{
Debug.WriteLine(Helpers.StringifyVector(actCols.ToArray()));
colSw.WriteLine(Helpers.StringifyVector(actCols.ToArray()));
}
}
}
Debug.WriteLine("------------ END ------------");
}
public void CategorySequenceExperiment()
{
bool learn = true;
Parameters p = Parameters.getAllDefaultParameters();
p.Set(KEY.RANDOM, new ThreadSafeRandom(42));
p.Set(KEY.INPUT_DIMENSIONS, new int[] { 100 });
p.Set(KEY.CELLS_PER_COLUMN, 30);
string[] categories = new string[] { "A", "B", "C", "D" };
//string[] categories = new string[] { "A", "B", "C", "D", "E", "F", "G", "H", "I", "K", "L" , "M", "O", "P", "Q", "R", "S", "T", "U", "V", "W", "X", "Y", "Z", "Ö" };
CortexNetwork net = new CortexNetwork("my cortex");
List <CortexRegion> regions = new List <CortexRegion>();
CortexRegion region0 = new CortexRegion("1st Region");
regions.Add(region0);
SpatialPoolerMT sp1 = new SpatialPoolerMT();
TemporalMemory tm1 = new TemporalMemory();
var mem = new Connections();
p.apply(mem);
sp1.Init(mem, UnitTestHelpers.GetMemory());
tm1.Init(mem);
Dictionary <string, object> settings = new Dictionary <string, object>();
//settings.Add("W", 25);
settings.Add("N", 100);
//settings.Add("Radius", 1);
EncoderBase encoder = new CategoryEncoder(categories, settings);
//encoder.Encode()
CortexLayer <object, object> layer1 = new CortexLayer <object, object>("L1");
region0.AddLayer(layer1);
layer1.HtmModules.Add("encoder", encoder);
layer1.HtmModules.Add("sp", sp1);
//layer1.HtmModules.Add(tm1);
//layer1.Compute();
//IClassifier<string, ComputeCycle> cls = new HtmClassifier<string, ComputeCycle>();
HtmClassifier <string, ComputeCycle> cls = new HtmClassifier <string, ComputeCycle>();
HtmUnionClassifier <string, ComputeCycle> cls1 = new HtmUnionClassifier <string, ComputeCycle>();
//string[] inputs = new string[] { "A", "B", "C", "D" };
string[] inputs = new string[] { "A", "B", "C", "D" };
//
// This trains SP.
foreach (var input in inputs)
{
Debug.WriteLine($" ** {input} **");
for (int i = 0; i < 3; i++)
{
var lyrOut = layer1.Compute((object)input, learn) as ComputeCycle;
}
}
// Here we add TM module to the layer.
layer1.HtmModules.Add("tm", tm1);
//
// Now, training with SP+TM. SP is pretrained on pattern.
for (int i = 0; i < 200; i++)
{
foreach (var input in inputs)
{
var lyrOut = layer1.Compute(input, learn) as ComputeCycle;
//cls1.Learn(input, lyrOut.activeCells.ToArray(), learn);
//Debug.WriteLine($"Current Input: {input}");
cls.Learn(input, lyrOut.ActiveCells.ToArray());
Debug.WriteLine($"Current Input: {input}");
if (learn == false)
{
Debug.WriteLine($"Predict Input When Not Learn: {cls.GetPredictedInputValue(lyrOut.PredictiveCells.ToArray())}");
}
else
{
Debug.WriteLine($"Predict Input: {cls.GetPredictedInputValue(lyrOut.PredictiveCells.ToArray())}");
}
Debug.WriteLine("-----------------------------------------------------------\n----------------------------------------------------------");
}
if (i == 10)
{
Debug.WriteLine("Stop Learning From Here----------------------------");
learn = false;
}
// tm1.reset(mem);
}
Debug.WriteLine("------------------------------------------------------------------------\n----------------------------------------------------------------------------");
/*
* learn = false;
* for (int i = 0; i < 19; i++)
* {
* foreach (var input in inputs)
* {
* layer1.Compute((object)input, learn);
* }
* }
*/
}
/// <summary>
///
/// </summary>
private async Task RunExperimentNeuroVisualizer(int inputBits, Parameters p, EncoderBase encoder, List <double> inputValues)
{
Stopwatch sw = new Stopwatch();
sw.Start();
int maxMatchCnt = 0;
bool learn = true;
INeuroVisualizer vis = new WSNeuroVisualizer();
GenerateNeuroModel model = new GenerateNeuroModel();
await vis.ConnectToWSServerAsync();
await vis.InitModelAsync(model.CreateNeuroModel(new int[] { 1 }, (long[, ])p[KEY.COLUMN_DIMENSIONS], (int)p[KEY.CELLS_PER_COLUMN]));
CortexNetwork net = new CortexNetwork("my cortex");
List <CortexRegion> regions = new List <CortexRegion>();
CortexRegion region0 = new CortexRegion("1st Region");
regions.Add(region0);
SpatialPoolerMT sp1 = new SpatialPoolerMT();
TemporalMemory tm1 = new TemporalMemory();
var mem = new Connections();
p.apply(mem);
sp1.Init(mem, UnitTestHelpers.GetMemory());
tm1.Init(mem);
CortexLayer <object, object> layer1 = new CortexLayer <object, object>("L1");
//
// NewBorn learning stage.
region0.AddLayer(layer1);
layer1.HtmModules.Add("encoder", encoder);
layer1.HtmModules.Add("sp", sp1);
//HtmClassifier<double, ComputeCycle> cls = new HtmClassifier<double, ComputeCycle>();
HtmClassifier <string, ComputeCycle> cls = new HtmClassifier <string, ComputeCycle>();
double[] inputs = inputValues.ToArray();
int[] prevActiveCols = new int[0];
int maxSPLearningCycles = 50;
//
// This trains SP on input pattern.
// It performs some kind of unsupervised new-born learning.
foreach (var input in inputs)
{
List <(int Cycle, double Similarity)> elementOscilationResult = new List <(int Cycle, double Similarity)>();
Debug.WriteLine($"Learning ** {input} **");
for (int i = 0; i < maxSPLearningCycles; i++)
{
var lyrOut = layer1.Compute((object)input, learn) as ComputeCycle;
var activeColumns = layer1.GetResult("sp") as int[];
var actCols = activeColumns.OrderBy(c => c).ToArray();
var similarity = MathHelpers.CalcArraySimilarity(prevActiveCols, actCols);
await vis.UpdateColumnAsync(GetColumns(actCols));
Debug.WriteLine($" {i.ToString("D4")} SP-OUT: [{actCols.Length}/{similarity.ToString("0.##")}] - {Helpers.StringifyVector(actCols)}");
prevActiveCols = activeColumns;
}
}
// Here we add TM module to the layer.
layer1.HtmModules.Add("tm", tm1);
int cycle = 0;
int matches = 0;
string lastPredictedValue = "0";
Dictionary <double, List <List <int> > > activeColumnsLst = new Dictionary <double, List <List <int> > >();
foreach (var input in inputs)
{
if (activeColumnsLst.ContainsKey(input) == false)
{
activeColumnsLst.Add(input, new List <List <int> >());
}
}
int maxCycles = 3500;
//
// Now training with SP+TM. SP is pretrained on the given input pattern.
for (int i = 0; i < maxCycles; i++)
{
matches = 0;
cycle++;
Debug.WriteLine($"-------------- Cycle {cycle} ---------------");
string prevInput = "-1.0";
//
// Activate the 'New - Born' effect.
//if (i == 300)
//{
// mem.setMaxBoost(0.0);
// mem.updateMinPctOverlapDutyCycles(0.0);
// cls.ClearState();
//}
foreach (var input in inputs)
{
Debug.WriteLine($"-------------- {input} ---------------");
var lyrOut = layer1.Compute(input, learn) as ComputeCycle;
var activeColumns = layer1.GetResult("sp") as int[];
activeColumnsLst[input].Add(activeColumns.ToList());
//cls.Learn(input, lyrOut.ActiveCells.ToArray());
cls.Learn(GetKey(prevInput, input), lyrOut.ActiveCells.ToArray());
List <Synapse> synapses = new List <Synapse>();
Cell cell = new Cell(0, 1, 6, 0, CellActivity.ActiveCell); // where to get all these values
Synapse synap = new Synapse(cell, 1, 1, 0.78); // here is just supposed to update the permanence, all other values remains same; where do we get all other values
synapses.Add(synap);
await vis.UpdateSynapsesAsync(synapses); //update Synapse or add new ones
await vis.UpdateCellsAsync(GetCells(lyrOut.ActiveCells));
if (learn == false)
{
Debug.WriteLine($"Inference mode");
}
if (GetKey(prevInput, input) == lastPredictedValue)
{
matches++;
Debug.WriteLine($"Match {input}");
}
else
{
Debug.WriteLine($"Missmatch Actual value: {GetKey(prevInput, input)} - Predicted value: {lastPredictedValue}");
}
if (lyrOut.PredictiveCells.Count > 0)
{
var predictedInputValue = cls.GetPredictedInputValue(lyrOut.PredictiveCells.ToArray());
Debug.WriteLine($"Current Input: {input} \t| Predicted Input: {predictedInputValue}");
lastPredictedValue = predictedInputValue;
}
else
{
Debug.WriteLine($"NO CELLS PREDICTED for next cycle.");
}
prevInput = input.ToString();
}
//tm1.reset(mem);
double accuracy = (double)matches / (double)inputs.Length * 100.0;
Debug.WriteLine($"Cycle: {cycle}\tMatches={matches} of {inputs.Length}\t {accuracy}%");
if (accuracy == 100.0)
{
maxMatchCnt++;
Debug.WriteLine($"100% accuracy reched {maxMatchCnt} times.");
if (maxMatchCnt >= 20)
{
sw.Stop();
Debug.WriteLine($"Exit experiment in the stable state after 10 repeats with 100% of accuracy. Elapsed time: {sw.ElapsedMilliseconds / 1000 / 60} min.");
learn = false;
//var testInputs = new double[] { 0.0, 2.0, 3.0, 4.0, 5.0, 6.0, 5.0, 4.0, 3.0, 7.0, 1.0, 9.0, 12.0, 11.0, 0.0, 1.0 };
// C-0, D-1, E-2, F-3, G-4, H-5
//var testInputs = new double[] { 0.0, 0.0, 4.0, 4.0, 5.0, 5.0, 4.0, 3.0, 3.0, 2.0, 2.0, 1.0, 1.0, 0.0 };
//// Traverse the sequence and check prediction.
//foreach (var input in inputValues)
//{
// var lyrOut = layer1.Compute(input, learn) as ComputeCycle;
// predictedInputValue = cls.GetPredictedInputValue(lyrOut.predictiveCells.ToArray());
// Debug.WriteLine($"I={input} - P={predictedInputValue}");
//}
//
// Here we let the HTM predict seuence five times on its own.
// We start with last predicted value.
int cnt = 5 * inputValues.Count;
Debug.WriteLine("---- Start Predicting the Sequence -----");
// We take a random value to start somwhere in the sequence.
var predictedInputValue = inputValues[new Random().Next(0, inputValues.Count - 1)].ToString();
List <string> predictedValues = new List <string>();
while (--cnt > 0)
{
//var lyrOut = layer1.Compute(predictedInputValue, learn) as ComputeCycle;
var lyrOut = layer1.Compute(double.Parse(predictedInputValue[predictedInputValue.Length - 1].ToString()), learn) as ComputeCycle;
predictedInputValue = cls.GetPredictedInputValue(lyrOut.PredictiveCells.ToArray());
predictedValues.Add(predictedInputValue);
}
;
foreach (var item in predictedValues)
{
Debug.Write(item);
Debug.Write(" ,");
}
break;
}
}
else if (maxMatchCnt > 0)
{
Debug.WriteLine($"At 100% accuracy after {maxMatchCnt} repeats we get a drop of accuracy with {accuracy}. This indicates instable state. Learning will be continued.");
maxMatchCnt = 0;
}
}
Debug.WriteLine("---- cell state trace ----");
cls.TraceState($"cellState_MinPctOverlDuty-{p[KEY.MIN_PCT_OVERLAP_DUTY_CYCLES]}_MaxBoost-{p[KEY.MAX_BOOST]}.csv");
Debug.WriteLine("---- column state trace ----");
foreach (var input in activeColumnsLst)
{
using (StreamWriter colSw = new StreamWriter($"ColumState_MinPctOverlDuty-{p[KEY.MIN_PCT_OVERLAP_DUTY_CYCLES]}_MaxBoost-{p[KEY.MAX_BOOST]}_input-{input.Key}.csv"))
{
Debug.WriteLine($"------------ {input} ------------");
foreach (var actCols in input.Value)
{
Debug.WriteLine(Helpers.StringifyVector(actCols.ToArray()));
colSw.WriteLine(Helpers.StringifyVector(actCols.ToArray()));
}
}
}
Debug.WriteLine("------------ END ------------");
}
public void SimilarityExperimentWithEncoder()
{
int stableStateCnt = 100;
double minOctOverlapCycles = 1.0;
double maxBoost = 10.0;
int inputBits = 100;
var colDims = new int[] { 64 * 64 };
int numOfActCols = colDims[0];
int numColumns = colDims[0];
string TestOutputFolder = $"Output-{nameof(ImageSimilarityExperiment)}";
Directory.CreateDirectory($"{nameof(ImageSimilarityExperiment)}");
//int counter = 0;
//var parameters = GetDefaultParams();
////parameters.Set(KEY.DUTY_CYCLE_PERIOD, 20);
////parameters.Set(KEY.MAX_BOOST, 1);
////parameters.setInputDimensions(new int[] { imageSize[imSizeIndx], imageSize[imSizeIndx] });
////parameters.setColumnDimensions(new int[] { topologies[topologyIndx], topologies[topologyIndx] });
////parameters.setNumActiveColumnsPerInhArea(0.02 * numOfActCols);
//parameters.Set(KEY.NUM_ACTIVE_COLUMNS_PER_INH_AREA, 0.06 * 4096); // TODO. Experiment with different sizes
//parameters.Set(KEY.POTENTIAL_RADIUS, inputBits);
//parameters.Set(KEY.POTENTIAL_PCT, 1.0);
//parameters.Set(KEY.GLOBAL_INHIBITION, true); // TODO: Experiment with local inhibition too. Note also the execution time of the experiment.
//// Num of active synapces in order to activate the column.
//parameters.Set(KEY.STIMULUS_THRESHOLD, 50.0);
//parameters.Set(KEY.SYN_PERM_INACTIVE_DEC, 0.008);
//parameters.Set(KEY.SYN_PERM_ACTIVE_INC, 0.05);
//parameters.Set(KEY.INHIBITION_RADIUS, (int)0.15 * inputBits); // TODO. check if this has influence in a case of the global inhibition. ALso check how this parameter influences the similarity of SDR.
//parameters.Set(KEY.SYN_PERM_CONNECTED, 0.2);
//parameters.Set(KEY.MIN_PCT_OVERLAP_DUTY_CYCLES, 1.0);
//parameters.Set(KEY.MIN_PCT_ACTIVE_DUTY_CYCLES, 0.001);
//parameters.Set(KEY.DUTY_CYCLE_PERIOD, 100);
//parameters.Set(KEY.MAX_BOOST, 10);
//parameters.Set(KEY.WRAP_AROUND, true);
//parameters.Set(KEY.SEED, 1969);
//parameters.setInputDimensions(new int[] {inputBits });
//parameters.setColumnDimensions(colDims);
Parameters p = Parameters.getAllDefaultParameters();
p.Set(KEY.RANDOM, new ThreadSafeRandom(42));
p.Set(KEY.INPUT_DIMENSIONS, new int[] { inputBits });
p.Set(KEY.COLUMN_DIMENSIONS, colDims);
p.Set(KEY.CELLS_PER_COLUMN, 10);
p.Set(KEY.MAX_BOOST, maxBoost);
p.Set(KEY.DUTY_CYCLE_PERIOD, 50);
p.Set(KEY.MIN_PCT_OVERLAP_DUTY_CYCLES, minOctOverlapCycles);
// Global inhibition
// N of 40 (40= 0.02*2048 columns) active cells required to activate the segment.
p.Set(KEY.GLOBAL_INHIBITION, true);
p.setNumActiveColumnsPerInhArea(0.02 * numColumns);
p.Set(KEY.POTENTIAL_RADIUS, (int)(.7 * inputBits));
p.Set(KEY.LOCAL_AREA_DENSITY, -1); // In a case of global inhibition.
//p.setInhibitionRadius( Automatically set on the columns pace in a case of global inhibition.);
// Activation threshold is 10 active cells of 40 cells in inhibition area.
p.setActivationThreshold(10);
// Max number of synapses on the segment.
p.setMaxNewSynapsesPerSegmentCount((int)(0.02 * numColumns));
double max = 20;
Dictionary <string, object> settings = new Dictionary <string, object>()
{
{ "W", 15 },
{ "N", inputBits },
{ "Radius", -1.0 },
{ "MinVal", 0.0 },
{ "Periodic", false },
{ "Name", "scalar" },
{ "ClipInput", false },
{ "MaxVal", max }
};
var encoder = new ScalarEncoder(settings);
bool isInStableState = false;
var mem = new Connections();
p.apply(mem);
var inputs = new int[] { 0, 1, 2, 3, 4, 5 };
HomeostaticPlasticityController hpa = new HomeostaticPlasticityController(mem, inputs.Length * 150, (isStable, numPatterns, actColAvg, seenInputs) =>
{
// Event should only be fired when entering the stable state.
// Ideal SP should never enter unstable state after stable state.
Assert.IsTrue(isStable);
//Assert.IsTrue(numPatterns == inputs.Length);
isInStableState = true;
Debug.WriteLine($"Entered STABLE state: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
});
SpatialPooler sp = new SpatialPoolerMT(hpa);
sp.Init(mem, UnitTestHelpers.GetMemory());
string outFolder = $"{TestOutputFolder}";
Directory.CreateDirectory(outFolder);
string outputHamDistFile = $"{outFolder}\\hamming.txt";
string outputActColFile = $"{outFolder}\\activeCol.txt";
using (StreamWriter swHam = new StreamWriter(outputHamDistFile))
{
using (StreamWriter swActCol = new StreamWriter(outputActColFile))
{
int cycle = 0;
Dictionary <string, int[]> sdrs = new Dictionary <string, int[]>();
while (!isInStableState)
{
foreach (var digit in inputs)
{
int[] activeArray = new int[numOfActCols];
int[] oldArray = new int[activeArray.Length];
List <double[, ]> overlapArrays = new List <double[, ]>();
List <double[, ]> bostArrays = new List <double[, ]>();
var inputVector = encoder.Encode(digit);
sp.compute(inputVector, activeArray, true);
string actColFileName = Path.Combine(outFolder, $"{digit}.actcols.txt");
if (cycle == 0 && File.Exists(actColFileName))
{
File.Delete(actColFileName);
}
var activeCols = ArrayUtils.IndexWhere(activeArray, (el) => el == 1);
using (StreamWriter swCols = new StreamWriter(actColFileName, true))
{
swCols.WriteLine(Helpers.StringifyVector(activeCols));
}
Debug.WriteLine($"'Cycle: {cycle} - {digit}'");
Debug.WriteLine($"IN :{Helpers.StringifyVector(inputVector)}");
Debug.WriteLine($"OUT:{Helpers.StringifyVector(activeCols)}\n");
if (isInStableState)
{
if (--stableStateCnt <= 0)
{
return;
}
}
/*
* if (isInStableState)
* {
* swActCol.WriteLine($"\nDigit {digit}");
*
* sdrs.Add(digit.ToString(), activeCols);
*
* //
* // To be sure that same input produces the same output after entered the stable state.
* for (int i = 0; i < 100; i++)
* {
* activeArray = new int[numOfActCols];
*
* sp.compute(inputVector, activeArray, true);
*
* var distance = MathHelpers.GetHammingDistance(oldArray, activeArray, true);
*
* var actColsIndxes = ArrayUtils.IndexWhere(activeArray, i => i == 1);
* var oldActColsIndxes = ArrayUtils.IndexWhere(oldArray, i => i == 1);
*
* var similarity = MathHelpers.CalcArraySimilarity(actColsIndxes, oldActColsIndxes);
*
* swHam.Write($"Digit {digit}: Dist/Similarity: {distance} | {similarity}\t");
* Debug.Write($"Digit {digit}: Dist/Similarity: {distance} | {similarity}\t");
* Debug.WriteLine($"{Helpers.StringifyVector(actColsIndxes)}");
*
* if (i > 5 && similarity < 100)
* {
*
* }
*
* oldArray = new int[numOfActCols];
* activeArray.CopyTo(oldArray, 0);
* }
* }
*
* Debug.WriteLine($"Cycle {cycle++}");*/
}
cycle++;
}
CalculateSimilarity(sdrs, null);//todo
}
}
}
public void LearnCSI(int iteration, int colDimSize, int inputDimSize, int Hammingdim)
{
FolderInit();
Debug.WriteLine("** running learning SP with Continuous Same Inputs feeding");
/**************************************************************************************
* |INITIATION| |SPATIALPOOLER|
* Setting the parameters variable
* changing the KEY's values
*/
var parameters = GetDefaultParams();
parameters.Set(KEY.POTENTIAL_RADIUS, inputDimSize * inputDimSize);
parameters.Set(KEY.POTENTIAL_PCT, 0.75);
parameters.Set(KEY.GLOBAL_INHIBITION, true);
parameters.Set(KEY.STIMULUS_THRESHOLD, 5);//5
parameters.Set(KEY.INHIBITION_RADIUS, (int)(0.01 * colDimSize * colDimSize));
parameters.Set(KEY.LOCAL_AREA_DENSITY, -1);
parameters.Set(KEY.NUM_ACTIVE_COLUMNS_PER_INH_AREA, (int)(0.02 * colDimSize * colDimSize));
parameters.Set(KEY.DUTY_CYCLE_PERIOD, 1000); //1000
parameters.Set(KEY.MAX_BOOST, 10.0); //0
parameters.Set(KEY.SYN_PERM_INACTIVE_DEC, 0.008);
parameters.Set(KEY.SYN_PERM_ACTIVE_INC, 0.05);
parameters.Set(KEY.MIN_PCT_OVERLAP_DUTY_CYCLES, 0.001);
parameters.Set(KEY.SEED, 42);
parameters.setInputDimensions(new int[] { inputDimSize, inputDimSize });
parameters.setColumnDimensions(new int[] { colDimSize, colDimSize });
var sp = new SpatialPoolerMT();
var mem = new Connections();
//var rnd = new Random();
parameters.apply(mem);
sp.Init(mem);
/***************************************************************************************
* |LOGICTEST| |CONTINOUS_PATTERN_LEARNING|
* Using input from scalar encoder
*/
List <int[]> inputVectors;
double[] inputDoubleArray = GetInputDoubleArray(-100, 100, Hammingdim * Hammingdim);
inputVectors = ScalarEncoderDoubleArray(inputDoubleArray, inputDimSize, CSI);
int vectorIndex = 0;
Debug.WriteLine($"InputVectors.Count = {inputVectors.Count}");
int[][] activeArray = new int[inputVectors.Count][];
foreach (var inputVector in inputVectors)
{
activeArray[vectorIndex] = new int[colDimSize * colDimSize];
for (int i = 0; i < iteration; i++)
{
Debug.WriteLine(
$"compute VECTOR {vectorIndex} " +
$"iter: {i} " +
$"val: {NeoCortexApi.Helpers.StringifyVector(inputVector)} ");
sp.compute(inputVector, activeArray[vectorIndex], true);
}
sp.compute(inputVector, activeArray[vectorIndex], false);
vectorIndex++;
}
vectorIndex = 0;
foreach (var inputVector in inputVectors)
{
int[,] arrayDraw = ArrayUtils.Transpose(ArrayUtils.Make2DArray <int>(activeArray[vectorIndex], colDimSize, colDimSize));
NeoCortexUtils.DrawBitmap(arrayDraw,
OutImgSize, OutImgSize,
$"{outputFolder}\\{CSI}\\{SDROuputFolder_CSI}\\Vector_{vectorIndex}.png",
Color.FromArgb(44, 44, 44),
Color.FromArgb(250, 100, 100),
$"No.{vectorIndex} val {inputDoubleArray[vectorIndex]}");
vectorIndex++;
}
vectorIndex = 0;
//************************drawing similarities*******************************************
int[] dummyArray = new int[colDimSize * colDimSize];
int[][] diffArray = new int[inputVectors.Count][];
for (int i = 0; i < inputVectors.Count; i++)
{
sp.compute(inputVectors[i], dummyArray, false);
int[,] drawArray = ArrayUtils.Transpose(ArrayUtils.Make2DArray <int>(dummyArray, colDimSize, colDimSize));
NeoCortexUtils.DrawBitmap(drawArray,
OutImgSize, OutImgSize,
$"{outputFolder}\\{CSI}\\{FreshComputed_CSI}\\Vector_{i}.png",
Color.FromArgb(44, 44, 44),
Color.FromArgb(100, 250, 100),
$"No.{i} newCom, Val_{inputDoubleArray[i]}");
diffArray[i] = AddArray(activeArray[i], dummyArray);
}
foreach (var inputVector in inputVectors)
{
int[,] arrayDraw = ArrayUtils.Transpose(ArrayUtils.Make2DArray <int>(diffArray[vectorIndex], colDimSize, colDimSize));
Debug.WriteLine($"Drawing SDRdiff index : {vectorIndex} .........");
DrawdiffArray(arrayDraw,
OutImgSize / colDimSize,
$"{outputFolder}\\{CSI}\\{SDROuputDiff_CSI}\\Vector_{vectorIndex}.png",
Color.FromArgb(44, 44, 44),
Color.FromArgb(100, 100, 250),
Color.FromArgb(250, 100, 100),
Color.FromArgb(100, 250, 100),
$"No.{vectorIndex} val {inputDoubleArray[vectorIndex]}");
vectorIndex++;
}
//*********** Checking phase with getHammingDistance ************************************
double[] hammingArray = GetHammingArray(inputVectors, activeArray, sp);
LogToCSV(hammingArray, inputVectors, $"{CSI}\\{logOutputFolder}\\HammingOutput.csv");
DrawHammingBitmap(
ArrayUtils.Make2DArray <double>(hammingArray, Hammingdim, Hammingdim),
50,
$"{outputFolder}\\{CSI}\\{logOutputFolder}\\HammingBitmap.png");
}
/// <summary>
/// Implements the experiment.
/// </summary>
/// <param name="cfg"></param>
/// <param name="encoder"></param>
/// <param name="inputValues"></param>
private static void RunExperiment(HtmConfig cfg, EncoderBase encoder, List <int[]> inputValues)
{
// Creates the htm memory.
var mem = new Connections(cfg);
bool isInStableState = false;
//
// HPC extends the default Spatial Pooler algorithm.
// The purpose of HPC is to set the SP in the new-born stage at the begining of the learning process.
// In this stage the boosting is very active, but the SP behaves instable. After this stage is over
// (defined by the second argument) the HPC is controlling the learning process of the SP.
// Once the SDR generated for every input gets stable, the HPC will fire event that notifies your code
// that SP is stable now.
HomeostaticPlasticityController hpa = new HomeostaticPlasticityController(mem, inputValues.Count * 40,
(isStable, numPatterns, actColAvg, seenInputs) =>
{
// Event should only be fired when entering the stable state.
// Ideal SP should never enter unstable state after stable state.
if (isStable == false)
{
Debug.WriteLine($"INSTABLE STATE");
// This should usually not happen.
isInStableState = false;
}
else
{
Debug.WriteLine($"STABLE STATE");
// Here you can perform any action if required.
isInStableState = true;
}
}, requiredSimilarityThreshold: 0.975);
// It creates the instance of Spatial Pooler Multithreaded version.
SpatialPooler sp = new SpatialPoolerMT(hpa);
// Initializes the
sp.Init(mem);
// Holds the indicies of active columns of the SDR.
Dictionary <string, int[]> prevActiveColIndicies = new Dictionary <string, int[]>();
// Holds the active column SDRs.
Dictionary <string, int[]> prevActiveCols = new Dictionary <string, int[]>();
// Will hold the similarity of SDKk and SDRk-1 fro every input.
Dictionary <string, double> prevSimilarity = new Dictionary <string, double>();
//
// Initiaize start similarity to zero.
for (int i = 0; i < inputValues.Count; i++)
{
string inputKey = GetInputGekFromIndex(i);
prevSimilarity.Add(inputKey, 0.0);
prevActiveColIndicies.Add(inputKey, new int[0]);
}
// Learning process will take 1000 iterations (cycles)
int maxSPLearningCycles = 1000;
for (int cycle = 0; cycle < maxSPLearningCycles; cycle++)
{
//Debug.WriteLine($"Cycle ** {cycle} ** Stability: {isInStableState}");
//
// This trains the layer on input pattern.
for (int inputIndx = 0; inputIndx < inputValues.Count; inputIndx++)
{
string inputKey = GetInputGekFromIndex(inputIndx);
int[] input = inputValues[inputIndx];
double similarity;
int[] activeColumns = new int[(int)cfg.NumColumns];
// Learn the input pattern.
// Output lyrOut is the output of the last module in the layer.
sp.compute(input, activeColumns, true);
// DrawImages(cfg, inputKey, input, activeColumns);
var actColsIndicies = ArrayUtils.IndexWhere(activeColumns, c => c == 1);
similarity = MathHelpers.CalcArraySimilarity(actColsIndicies, prevActiveColIndicies[inputKey]);
Debug.WriteLine($"[i={inputKey}, cols=:{actColsIndicies.Length} s={similarity}] SDR: {Helpers.StringifyVector(actColsIndicies)}");
prevActiveCols[inputKey] = activeColumns;
prevActiveColIndicies[inputKey] = actColsIndicies;
prevSimilarity[inputKey] = similarity;
if (isInStableState)
{
GenerateResult(cfg, inputValues, prevActiveColIndicies, prevActiveCols);
return;
}
}
}
}
public void NoiseExperimentTest()
{
const int colDimSize = 64;
const int noiseStepPercent = 5;
var parameters = GetDefaultParams();
parameters.Set(KEY.POTENTIAL_RADIUS, 32 * 32);
parameters.Set(KEY.POTENTIAL_PCT, 1.0);
parameters.Set(KEY.GLOBAL_INHIBITION, true);
parameters.Set(KEY.STIMULUS_THRESHOLD, 0.5);
parameters.Set(KEY.INHIBITION_RADIUS, (int)0.01 * colDimSize * colDimSize);
parameters.Set(KEY.LOCAL_AREA_DENSITY, -1);
parameters.Set(KEY.NUM_ACTIVE_COLUMNS_PER_INH_AREA, 0.02 * colDimSize * colDimSize);
parameters.Set(KEY.DUTY_CYCLE_PERIOD, 1000);
parameters.Set(KEY.MAX_BOOST, 0.0);
parameters.Set(KEY.SYN_PERM_INACTIVE_DEC, 0.008);
parameters.Set(KEY.SYN_PERM_ACTIVE_INC, 0.01);
parameters.Set(KEY.MIN_PCT_OVERLAP_DUTY_CYCLES, 0.0);
parameters.Set(KEY.SEED, 42);
parameters.setInputDimensions(new int[] { 32, 32 });
parameters.setColumnDimensions(new int[] { colDimSize, colDimSize });
var sp = new SpatialPoolerMT();
var mem = new Connections();
parameters.apply(mem);
sp.Init(mem);
List <int[]> inputVectors = new List <int[]>();
inputVectors.Add(getInputVector1());
inputVectors.Add(getInputVector2());
int vectorIndex = 0;
int[][] activeColumnsWithZeroNoise = new int[inputVectors.Count][];
foreach (var inputVector in inputVectors)
{
var x = getNumBits(inputVector);
Debug.WriteLine("");
Debug.WriteLine($"----- VECTOR {vectorIndex} ----------");
// Array of active columns with zero noise. The reference (ideal) output.
activeColumnsWithZeroNoise[vectorIndex] = new int[colDimSize * colDimSize];
int[] activeArray = null;
for (int j = 0; j < 25; j += noiseStepPercent)
{
Debug.WriteLine($"--- Vector {0} - Noise Iteration {j} ----------");
int[] noisedInput;
if (j > 0)
{
noisedInput = ArrayUtils.FlipBit(inputVector, (double)((double)j / 100.00));
}
else
{
noisedInput = inputVector;
}
// TODO: Try CalcArraySimilarity
var d = MathHelpers.GetHammingDistance(inputVector, noisedInput, true);
Debug.WriteLine($"Input with noise {j} - HamDist: {d}");
Debug.WriteLine($"Original: {Helpers.StringifyVector(inputVector)}");
Debug.WriteLine($"Noised: {Helpers.StringifyVector(noisedInput)}");
for (int i = 0; i < 10; i++)
{
activeArray = sp.Compute(noisedInput, true, returnActiveColIndiciesOnly: false) as int[];
if (j > 0)
{
Debug.WriteLine($"{ MathHelpers.GetHammingDistance(activeColumnsWithZeroNoise[vectorIndex], activeArray, true)} -> {Helpers.StringifyVector(ArrayUtils.IndexWhere(activeArray, (el) => el == 1))}");
}
}
if (j == 0)
{
Array.Copy(activeArray, activeColumnsWithZeroNoise[vectorIndex], activeColumnsWithZeroNoise[vectorIndex].Length);
}
var activeCols = ArrayUtils.IndexWhere(activeArray, (el) => el == 1);
var d2 = MathHelpers.GetHammingDistance(activeColumnsWithZeroNoise[vectorIndex], activeArray, true);
Debug.WriteLine($"Output with noise {j} - Ham Dist: {d2}");
Debug.WriteLine($"Original: {Helpers.StringifyVector(ArrayUtils.IndexWhere(activeColumnsWithZeroNoise[vectorIndex], (el) => el == 1))}");
Debug.WriteLine($"Noised: {Helpers.StringifyVector(ArrayUtils.IndexWhere(activeArray, (el) => el == 1))}");
List <int[, ]> arrays = new List <int[, ]>();
int[,] twoDimenArray = ArrayUtils.Make2DArray <int>(activeArray, 64, 64);
twoDimenArray = ArrayUtils.Transpose(twoDimenArray);
arrays.Add(ArrayUtils.Transpose(ArrayUtils.Make2DArray <int>(noisedInput, 32, 32)));
arrays.Add(ArrayUtils.Transpose(ArrayUtils.Make2DArray <int>(activeArray, 64, 64)));
// NeoCortexUtils.DrawHeatmaps(bostArrays, $"{outputImage}_boost.png", 1024, 1024, 150, 50, 5);
NeoCortexUtils.DrawBitmaps(arrays, $"Vector_{vectorIndex}_Noise_{j * 10}.png", Color.Yellow, Color.Gray, OutImgSize, OutImgSize);
}
vectorIndex++;
}
vectorIndex = OutputPredictionResult(sp, inputVectors, activeColumnsWithZeroNoise);
}
/// <summary>
///
/// </summary>
private void RunExperiment(int inputBits, HtmConfig cfg, EncoderBase encoder, List <double> inputValues)
{
Stopwatch sw = new Stopwatch();
sw.Start();
int maxMatchCnt = 0;
bool learn = true;
var mem = new Connections(cfg);
bool isInStableState = false;
HtmClassifier <string, ComputeCycle> cls = new HtmClassifier <string, ComputeCycle>();
var numInputs = inputValues.Distinct <double>().ToList().Count;
CortexLayer <object, object> layer1 = new CortexLayer <object, object>("L1");
TemporalMemory tm = new TemporalMemory();
HomeostaticPlasticityController hpa = new HomeostaticPlasticityController(mem, numInputs * 150, (isStable, numPatterns, actColAvg, seenInputs) =>
{
if (isStable)
{
// Event should be fired when entering the stable state.
Debug.WriteLine($"STABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
else
{
// Ideal SP should never enter unstable state after stable state.
Debug.WriteLine($"INSTABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
// We are not learning in instable state.
learn = isInStableState = isStable;
//if (isStable && layer1.HtmModules.ContainsKey("tm") == false)
// layer1.HtmModules.Add("tm", tm);
// Clear all learned patterns in the classifier.
cls.ClearState();
// Clear active and predictive cells.
//tm.Reset(mem);
}, numOfCyclesToWaitOnChange: 50);
SpatialPoolerMT sp = new SpatialPoolerMT(hpa);
sp.Init(mem);
tm.Init(mem);
layer1.HtmModules.Add("encoder", encoder);
layer1.HtmModules.Add("sp", sp);
double[] inputs = inputValues.ToArray();
int[] prevActiveCols = new int[0];
int cycle = 0;
int matches = 0;
string lastPredictedValue = "0";
//Dictionary<double, List<List<int>>> activeColumnsLst = new Dictionary<double, List<List<int>>>();
//foreach (var input in inputs)
//{
// if (activeColumnsLst.ContainsKey(input) == false)
// activeColumnsLst.Add(input, new List<List<int>>());
//}
int maxCycles = 3500;
int maxPrevInputs = inputValues.Count - 1;
List <string> previousInputs = new List <string>();
previousInputs.Add("-1.0");
//
// Training SP to get stable. New-born stage.
//
for (int i = 0; i < maxCycles; i++)
{
matches = 0;
cycle++;
Debug.WriteLine($"-------------- Newborn Cycle {cycle} ---------------");
foreach (var input in inputs)
{
Debug.WriteLine($" -- {input} --");
var lyrOut = layer1.Compute(input, learn);
if (isInStableState)
{
break;
}
}
if (isInStableState)
{
break;
}
}
layer1.HtmModules.Add("tm", tm);
//
// Now training with SP+TM. SP is pretrained on the given input pattern set.
for (int i = 0; i < maxCycles; i++)
{
matches = 0;
cycle++;
Debug.WriteLine($"-------------- Cycle {cycle} ---------------");
foreach (var input in inputs)
{
Debug.WriteLine($"-------------- {input} ---------------");
var lyrOut = layer1.Compute(input, learn) as ComputeCycle;
// lyrOut is null when the TM is added to the layer inside of HPC callback by entering of the stable state.
//if (isInStableState && lyrOut != null)
{
var activeColumns = layer1.GetResult("sp") as int[];
//layer2.Compute(lyrOut.WinnerCells, true);
//activeColumnsLst[input].Add(activeColumns.ToList());
previousInputs.Add(input.ToString());
if (previousInputs.Count > (maxPrevInputs + 1))
{
previousInputs.RemoveAt(0);
}
// In the pretrained SP with HPC, the TM will quickly learn cells for patterns
// In that case the starting sequence 4-5-6 might have the sam SDR as 1-2-3-4-5-6,
// Which will result in returning of 4-5-6 instead of 1-2-3-4-5-6.
// HtmClassifier allways return the first matching sequence. Because 4-5-6 will be as first
// memorized, it will match as the first one.
if (previousInputs.Count < maxPrevInputs)
{
continue;
}
string key = GetKey(previousInputs, input);
List <Cell> actCells;
if (lyrOut.ActiveCells.Count == lyrOut.WinnerCells.Count)
{
actCells = lyrOut.ActiveCells;
}
else
{
actCells = lyrOut.WinnerCells;
}
cls.Learn(key, actCells.ToArray());
if (learn == false)
{
Debug.WriteLine($"Inference mode");
}
Debug.WriteLine($"Col SDR: {Helpers.StringifyVector(lyrOut.ActivColumnIndicies)}");
Debug.WriteLine($"Cell SDR: {Helpers.StringifyVector(actCells.Select(c => c.Index).ToArray())}");
if (key == lastPredictedValue)
{
matches++;
Debug.WriteLine($"Match. Actual value: {key} - Predicted value: {lastPredictedValue}");
}
else
{
Debug.WriteLine($"Missmatch! Actual value: {key} - Predicted value: {lastPredictedValue}");
}
if (lyrOut.PredictiveCells.Count > 0)
{
var predictedInputValue = cls.GetPredictedInputValue(lyrOut.PredictiveCells.ToArray());
Debug.WriteLine($"Current Input: {input} \t| Predicted Input: {predictedInputValue}");
lastPredictedValue = predictedInputValue;
}
else
{
Debug.WriteLine($"NO CELLS PREDICTED for next cycle.");
lastPredictedValue = String.Empty;
}
}
}
// The brain does not do that this way, so we don't use it.
// tm1.reset(mem);
double accuracy = (double)matches / (double)inputs.Length * 100.0;
Debug.WriteLine($"Cycle: {cycle}\tMatches={matches} of {inputs.Length}\t {accuracy}%");
if (accuracy == 100.0)
{
maxMatchCnt++;
Debug.WriteLine($"100% accuracy reched {maxMatchCnt} times.");
//
// Experiment is completed if we are 30 cycles long at the 100% accuracy.
if (maxMatchCnt >= 30)
{
sw.Stop();
Debug.WriteLine($"Exit experiment in the stable state after 30 repeats with 100% of accuracy. Elapsed time: {sw.ElapsedMilliseconds / 1000 / 60} min.");
learn = false;
break;
}
}
else if (maxMatchCnt > 0)
{
Debug.WriteLine($"At 100% accuracy after {maxMatchCnt} repeats we get a drop of accuracy with {accuracy}. This indicates instable state. Learning will be continued.");
maxMatchCnt = 0;
}
}
Debug.WriteLine("------------ END ------------");
}
/// <summary>
///
/// </summary>
private static void RunExperiment(int inputBits, HtmConfig cfg, EncoderBase encoder, List <double> inputValues)
{
Stopwatch sw = new Stopwatch();
sw.Start();
int maxMatchCnt = 0;
bool learn = true;
CortexNetwork net = new CortexNetwork("my cortex");
List <CortexRegion> regions = new List <CortexRegion>();
CortexRegion region0 = new CortexRegion("1st Region");
regions.Add(region0);
var mem = new Connections(cfg);
bool isInStableState;
HtmClassifier <string, ComputeCycle> cls = new HtmClassifier <string, ComputeCycle>();
var numInputs = inputValues.Distinct <double>().ToList().Count;
TemporalMemory tm1 = new TemporalMemory();
HomeostaticPlasticityController hpa = new HomeostaticPlasticityController(mem, numInputs * 55, (isStable, numPatterns, actColAvg, seenInputs) =>
{
if (isStable)
{
// Event should be fired when entering the stable state.
Debug.WriteLine($"STABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
else
{
// Ideal SP should never enter unstable state after stable state.
Debug.WriteLine($"INSTABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
if (numPatterns != numInputs)
{
throw new InvalidOperationException("Stable state must observe all input patterns");
}
isInStableState = true;
cls.ClearState();
tm1.Reset(mem);
}, numOfCyclesToWaitOnChange: 25);
SpatialPoolerMT sp1 = new SpatialPoolerMT(hpa);
sp1.Init(mem, new DistributedMemory()
{
ColumnDictionary = new InMemoryDistributedDictionary <int, NeoCortexApi.Entities.Column>(1),
});
tm1.Init(mem);
CortexLayer <object, object> layer1 = new CortexLayer <object, object>("L1");
region0.AddLayer(layer1);
layer1.HtmModules.Add("encoder", encoder);
layer1.HtmModules.Add("sp", sp1);
layer1.HtmModules.Add("tm", tm1);
double[] inputs = inputValues.ToArray();
int[] prevActiveCols = new int[0];
int cycle = 0;
int matches = 0;
string lastPredictedValue = "0";
String prediction = null;
Dictionary <double, List <List <int> > > activeColumnsLst = new Dictionary <double, List <List <int> > >();
foreach (var input in inputs)
{
if (activeColumnsLst.ContainsKey(input) == false)
{
activeColumnsLst.Add(input, new List <List <int> >());
}
}
int maxCycles = 3500;
int maxPrevInputs = inputValues.Count - 1;
List <string> previousInputs = new List <string>();
previousInputs.Add("-1.0");
//
// Now training with SP+TM. SP is pretrained on the given input pattern.
for (int i = 0; i < maxCycles; i++)
{
matches = 0;
cycle++;
Debug.WriteLine($"-------------- Cycle {cycle} ---------------");
foreach (var input in inputs)
{
Debug.WriteLine($"-------------- {input} ---------------");
var lyrOut = layer1.Compute(input, learn) as ComputeCycle;
var activeColumns = layer1.GetResult("sp") as int[];
activeColumnsLst[input].Add(activeColumns.ToList());
previousInputs.Add(input.ToString());
if (previousInputs.Count > (maxPrevInputs + 1))
{
previousInputs.RemoveAt(0);
}
string key = GetKey(previousInputs, input);
cls.Learn(key, lyrOut.ActiveCells.ToArray());
if (learn == false)
{
Debug.WriteLine($"Inference mode");
}
Debug.WriteLine($"Col SDR: {Helpers.StringifyVector(lyrOut.ActivColumnIndicies)}");
Debug.WriteLine($"Cell SDR: {Helpers.StringifyVector(lyrOut.ActiveCells.Select(c => c.Index).ToArray())}");
if (key == lastPredictedValue)
{
matches++;
Debug.WriteLine($"Match. Actual value: {key} - Predicted value: {lastPredictedValue}");
}
else
{
Debug.WriteLine($"Missmatch! Actual value: {key} - Predicted value: {lastPredictedValue}");
}
if (lyrOut.PredictiveCells.Count > 0)
{
var predictedInputValue = cls.GetPredictedInputValues(lyrOut.PredictiveCells.ToArray(), 3);
Debug.WriteLine($"Current Input: {input}");
Debug.WriteLine("The predictions with similarity greater than 50% are");
foreach (var t in predictedInputValue)
{
if (t.Similarity >= (double)50.00)
{
Debug.WriteLine($"Predicted Input: {string.Join(", ", t.PredictedInput)},\tSimilarity Percentage: {string.Join(", ", t.Similarity)}, \tNumber of Same Bits: {string.Join(", ", t.NumOfSameBits)}");
}
}
lastPredictedValue = predictedInputValue.First().PredictedInput;
}
else
{
Debug.WriteLine($"NO CELLS PREDICTED for next cycle.");
lastPredictedValue = String.Empty;
}
}
double accuracy = (double)matches / (double)inputs.Length * 100.0;
Debug.WriteLine($"Cycle: {cycle}\tMatches={matches} of {inputs.Length}\t {accuracy}%");
if (accuracy == 100.0)
{
maxMatchCnt++;
Debug.WriteLine($"100% accuracy reched {maxMatchCnt} times.");
if (maxMatchCnt >= 30)
{
sw.Stop();
Debug.WriteLine($"Exit experiment in the stable state after 30 repeats with 100% of accuracy. Elapsed time: {sw.ElapsedMilliseconds / 1000 / 60} min.");
learn = false;
break;
}
}
else if (maxMatchCnt > 0)
{
Debug.WriteLine($"At 100% accuracy after {maxMatchCnt} repeats we get a drop of accuracy with {accuracy}. This indicates instable state. Learning will be continued.");
maxMatchCnt = 0;
}
}
Debug.WriteLine("---- cell state trace ----");
cls.TraceState($"cellState_MinPctOverlDuty-{cfg.MinPctOverlapDutyCycles}_MaxBoost-{cfg.MaxBoost}.csv");
Debug.WriteLine("---- Spatial Pooler column state ----");
foreach (var input in activeColumnsLst)
{
using (StreamWriter colSw = new StreamWriter($"ColumState_MinPctOverlDuty-{cfg.MinPctOverlapDutyCycles}_MaxBoost-{cfg.MaxBoost}_input-{input.Key}.csv"))
{
Debug.WriteLine($"------------ {input.Key} ------------");
foreach (var actCols in input.Value)
{
Debug.WriteLine(Helpers.StringifyVector(actCols.ToArray()));
colSw.WriteLine(Helpers.StringifyVector(actCols.ToArray()));
}
}
}
Debug.WriteLine("------------ END ------------");
Console.WriteLine("\n Please enter a number that has been learnt");
int inputNumber = Convert.ToInt16(Console.ReadLine());
Inference(inputNumber, false, layer1, cls);
}
/// <summary>
/// Implements the experiment.
/// </summary>
/// <param name="cfg"></param>
/// <param name="encoder"></param>
/// <param name="inputValues"></param>
private static void RunExperiment(HtmConfig cfg, EncoderBase encoder, List <double> inputValues)
{
// Creates the htm memory.
var mem = new Connections(cfg);
bool isInStableState = false;
//
// HPC extends the default Spatial Pooler algorithm.
// The purpose of HPC is to set the SP in the new-born stage at the begining of the learning process.
// In this stage the boosting is very active, but the SP behaves instable. After this stage is over
// (defined by the second argument) the HPC is controlling the learning process of the SP.
// Once the SDR generated for every input gets stable, the HPC will fire event that notifies your code
// that SP is stable now.
HomeostaticPlasticityController hpa = new HomeostaticPlasticityController(mem, inputValues.Count * 15,
(isStable, numPatterns, actColAvg, seenInputs) =>
{
// Event should only be fired when entering the stable state.
// Ideal SP should never enter unstable state after stable state.
if (isStable == false)
{
Debug.WriteLine($"INSTABLE");
// This should usually not happen.
isInStableState = false;
}
else
{
Debug.WriteLine($"STABILITY");
// Here you can perform any action if required.
isInStableState = true;
}
});
// It creates the instance of Spatial Pooler Multithreaded version.
SpatialPooler sp = new SpatialPoolerMT(hpa);
// Initializes the
sp.Init(mem, new DistributedMemory()
{
ColumnDictionary = new InMemoryDistributedDictionary <int, NeoCortexApi.Entities.Column>(1)
});
// It creates the instance of the neo-cortex layer.
// Algorithm will be performed inside of that layer.
CortexLayer <object, object> cortexLayer = new CortexLayer <object, object>("L1");
// Add encoder as the very first module. This model is connected to the sensory input cells
// that receive the input. Encoder will receive the input and forward the encoded signal
// to the next module.
cortexLayer.HtmModules.Add("encoder", encoder);
// The next module in the layer is Spatial Pooler. This module will receive the output of the
// encoder.
cortexLayer.HtmModules.Add("sp", sp);
double[] inputs = inputValues.ToArray();
// Will hold the SDR of every inputs.
Dictionary <double, int[]> prevActiveCols = new Dictionary <double, int[]>();
// Will hold the similarity of SDKk and SDRk-1 fro every input.
Dictionary <double, double> prevSimilarity = new Dictionary <double, double>();
//
// Initiaize start similarity to zero.
foreach (var input in inputs)
{
prevSimilarity.Add(input, 0.0);
prevActiveCols.Add(input, new int[0]);
}
// Learning process will take 1000 iterations (cycles)
int maxSPLearningCycles = 1000;
for (int cycle = 0; cycle < maxSPLearningCycles; cycle++)
{
Debug.WriteLine($"Cycle ** {cycle} ** Stability: {isInStableState}");
//
// This trains the layer on input pattern.
foreach (var input in inputs)
{
double similarity;
// Learn the input pattern.
// Output lyrOut is the output of the last module in the layer.
//
var lyrOut = cortexLayer.Compute((object)input, true) as int[];
// This is a general way to get the SpatialPooler result from the layer.
var activeColumns = cortexLayer.GetResult("sp") as int[];
var actCols = activeColumns.OrderBy(c => c).ToArray();
similarity = MathHelpers.CalcArraySimilarity(activeColumns, prevActiveCols[input]);
Debug.WriteLine($"[cycle={cycle.ToString("D4")}, i={input}, cols=:{actCols.Length} s={similarity}] SDR: {Helpers.StringifyVector(actCols)}");
prevActiveCols[input] = activeColumns;
prevSimilarity[input] = similarity;
}
}
}
private void RunExperiment(int inputBits, HtmConfig cfgL4, EncoderBase encoder, List <double> inputValues, HtmConfig cfgL2)
{
Stopwatch swL2 = new Stopwatch();
//int maxMatchCnt = 0;
bool learn = true;
bool isSP1Stable = false;
bool isSP2STable = false;
var memL4 = new Connections(cfgL4);
var memL2 = new Connections(cfgL2);
var numInputs = inputValues.Distinct <double>().ToList().Count;
HtmClassifier <string, ComputeCycle> cls = new HtmClassifier <string, ComputeCycle>();
layerL4 = new CortexLayer <object, object>("L4");
layerL2 = new CortexLayer <object, object>("L2");
tm4 = new TemporalMemoryMT();
tm2 = new TemporalMemoryMT();
//
// HPC for Layer 4 SP
HomeostaticPlasticityController hpa_sp_L4 = new HomeostaticPlasticityController(memL4, numInputs * 50, (isStable, numPatterns, actColAvg, seenInputs) =>
{
if (isStable)
{
Debug.WriteLine($"SP L4 STABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
else
{
Debug.WriteLine($"SP L4 INSTABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
learn = isSP1Stable = isStable;
//cls.ClearState();
}, numOfCyclesToWaitOnChange: 50);
//
// HPC for Layer 2 SP
HomeostaticPlasticityController hpa_sp_L2 = new HomeostaticPlasticityController(memL2, numInputs * 50, (isStable, numPatterns, actColAvg, seenInputs) =>
{
if (isStable)
{
Debug.WriteLine($"SP L2 STABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
else
{
Debug.WriteLine($"SP L2 INSTABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
learn = isSP2STable = isStable;
//cls.ClearState();
}, numOfCyclesToWaitOnChange: 50);
SpatialPooler sp4 = new SpatialPoolerMT(hpa_sp_L4);
SpatialPooler sp2 = new SpatialPoolerMT(hpa_sp_L2);
sp4.Init(memL4);
sp2.Init(memL2);
// memL2.TraceInputPotential();
tm4.Init(memL4);
tm2.Init(memL2);
layerL4.HtmModules.Add("encoder", encoder);
layerL4.HtmModules.Add("sp", sp4);
layerL4.HtmModules.Add("tm", tm4);
layerL2.HtmModules.Add("sp", sp2);
layerL2.HtmModules.Add("tm", tm2);
int[] inpCellsL4ToL2 = new int[cfgL4.CellsPerColumn * cfgL4.NumColumns];
double[] inputs = inputValues.ToArray();
int[] prevActiveCols = new int[0];
int cycle = 0;
int matches = 0;
// string lastPredictedValue = "0";
int maxCycles = 3500;
int maxPrevInputs = inputValues.Count - 1;
List <string> previousInputs = new List <string>();
//
// Training SP at Layer 4 to get stable. New-born stage.
//
using (StreamWriter swL4Sdrs = new StreamWriter($"L4-SDRs-in_{cfgL2.NumInputs}-col_{cfgL2.NumColumns}-r_{cfgL2.PotentialRadius}.txt"))
{
using (StreamWriter sw = new StreamWriter($"in_{cfgL2.NumInputs}-col_{cfgL2.NumColumns}-r_{cfgL2.PotentialRadius}.txt"))
{
for (int i = 0; i < maxCycles; i++)
{
matches = 0;
cycle++;
Debug.WriteLine($"-------------- Newborn Cycle {cycle} at L4 SP region ---------------");
foreach (var input in inputs)
{
Debug.WriteLine($" INPUT: '{input}'\tCycle:{cycle}");
Debug.Write("L4: ");
var lyrOut = layerL4.Compute(input, learn);
InitArray(inpCellsL4ToL2, 0);
if (isSP1Stable)
{
var cellSdrL4Indexes = memL4.ActiveCells.Select(c => c.Index).ToArray();
// Write SDR as output of L4 and input of L2
swL4Sdrs.WriteLine($"{input} - {Helpers.StringifyVector(cellSdrL4Indexes)}");
// Set the output active cell array
ArrayUtils.SetIndexesTo(inpCellsL4ToL2, cellSdrL4Indexes, 1);
Debug.Write("L2: ");
swL2.Restart();
layerL2.Compute(inpCellsL4ToL2, true);
swL2.Stop();
Debug.WriteLine($"{swL2.ElapsedMilliseconds / 1000}");
sw.WriteLine($"{swL2.ElapsedMilliseconds / 1000}");
sw.Flush();
Debug.WriteLine($"L4 out sdr: {Helpers.StringifyVector(cellSdrL4Indexes)}");
var overlaps = ArrayUtils.IndexWhere(memL2.Overlaps, o => o > 0);
var strOverlaps = Helpers.StringifyVector(overlaps);
Debug.WriteLine($"Potential columns: {overlaps.Length}, overlaps: {strOverlaps}");
}
if (isSP1Stable && isSP2STable)
{
break;
}
}
}
}
}
Debug.WriteLine($"-------------- L4 SP region is {isSP1Stable} ---------------");
//layerL4.HtmModules.Add("tm", tm4);
// SP+TM at L4
for (int i = 0; i < maxCycles; i++)
{
matches -= 0;
cycle++;
Debug.WriteLine($"-------------- L4 TM Train region Cycle {cycle} ---------------");
foreach (var input in inputs)
{
Debug.WriteLine($"-------------- {input} ---------------");
var lyrOut = layerL4.Compute(input, learn) as ComputeCycle;
previousInputs.Add(input.ToString());
if (previousInputs.Count > (maxPrevInputs + 1))
{
previousInputs.RemoveAt(0);
}
if (previousInputs.Count < maxPrevInputs)
{
continue;
}
if (lyrOut.ActiveCells.Count == lyrOut.WinnerCells.Count)
{
/*var activeColumns = L4.GetResult("sp") as int[];
* foreach (var item in activeColumns)
* {
* L2.Compute(item, true);
* }*/
// Reset tha array
InitArray(inpCellsL4ToL2, 0);
// Set the output active cell array
ArrayUtils.SetIndexesTo(inpCellsL4ToL2, memL4.ActiveCells.Select(c => c.Index).ToArray(), 1);
// 4102,25072, 25363, 25539, 25738, 25961, 26009, 26269, 26491, 26585, 26668, 26920, 26934, 27040, 27107, 27262, 27392, 27826, 27948, 28174, 28243, 28270, 28294, 28308, 28429, 28577, 28671, 29139, 29618, 29637, 29809, 29857, 29897, 29900, 29969, 30057, 30727, 31111, 49805, 49972,
layerL2.Compute(inpCellsL4ToL2, true);
/*foreach (var item in lyrOut.ActiveCells)
* {
* L2.Compute(item, true);
* }*/
//var activeCell = Helpers.StringifyVector(lyrOut.ActiveCells.Select(c => c.Index).ToArray());
//L2.Compute(activeCell, true);
}
}
}
}
/// <summary>
///
/// </summary>
private HtmPredictionEngine RunExperiment(int inputBits, HtmConfig cfg, EncoderBase encoder, Dictionary <string, List <double> > sequences)
{
Stopwatch sw = new Stopwatch();
sw.Start();
int maxMatchCnt = 0;
var mem = new Connections(cfg);
bool isInStableState = false;
HtmClassifier <string, ComputeCycle> cls = new HtmClassifier <string, ComputeCycle>();
var numUniqueInputs = GetNumberOfInputs(sequences);
CortexLayer <object, object> layer1 = new CortexLayer <object, object>("L1");
TemporalMemory tm = new TemporalMemory();
// For more information see following paper: https://www.scitepress.org/Papers/2021/103142/103142.pdf
HomeostaticPlasticityController hpc = new HomeostaticPlasticityController(mem, numUniqueInputs * 150, (isStable, numPatterns, actColAvg, seenInputs) =>
{
if (isStable)
{
// Event should be fired when entering the stable state.
Debug.WriteLine($"STABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
else
{
// Ideal SP should never enter unstable state after stable state.
Debug.WriteLine($"INSTABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
// We are not learning in instable state.
isInStableState = isStable;
// Clear active and predictive cells.
//tm.Reset(mem);
}, numOfCyclesToWaitOnChange: 50);
SpatialPoolerMT sp = new SpatialPoolerMT(hpc);
sp.Init(mem);
tm.Init(mem);
// Please note that we do not add here TM in the layer.
// This is omitted for practical reasons, because we first eneter the newborn-stage of the algorithm
// In this stage we want that SP get boosted and see all elements before we start learning with TM.
// All would also work fine with TM in layer, but it would work much slower.
// So, to improve the speed of experiment, we first ommit the TM and then after the newborn-stage we add it to the layer.
layer1.HtmModules.Add("encoder", encoder);
layer1.HtmModules.Add("sp", sp);
//double[] inputs = inputValues.ToArray();
int[] prevActiveCols = new int[0];
int cycle = 0;
int matches = 0;
var lastPredictedValues = new List <string>(new string[] { "0" });
int maxCycles = 3500;
//
// Training SP to get stable. New-born stage.
//
for (int i = 0; i < maxCycles && isInStableState == false; i++)
{
matches = 0;
cycle++;
Debug.WriteLine($"-------------- Newborn Cycle {cycle} ---------------");
foreach (var inputs in sequences)
{
foreach (var input in inputs.Value)
{
Debug.WriteLine($" -- {inputs.Key} - {input} --");
var lyrOut = layer1.Compute(input, true);
if (isInStableState)
{
break;
}
}
if (isInStableState)
{
break;
}
}
}
// Clear all learned patterns in the classifier.
cls.ClearState();
// We activate here the Temporal Memory algorithm.
layer1.HtmModules.Add("tm", tm);
//
// Loop over all sequences.
foreach (var sequenceKeyPair in sequences)
{
Debug.WriteLine($"-------------- Sequences {sequenceKeyPair.Key} ---------------");
int maxPrevInputs = sequenceKeyPair.Value.Count - 1;
List <string> previousInputs = new List <string>();
previousInputs.Add("-1.0");
//
// Now training with SP+TM. SP is pretrained on the given input pattern set.
for (int i = 0; i < maxCycles; i++)
{
matches = 0;
cycle++;
Debug.WriteLine("");
Debug.WriteLine($"-------------- Cycle {cycle} ---------------");
Debug.WriteLine("");
foreach (var input in sequenceKeyPair.Value)
{
Debug.WriteLine($"-------------- {input} ---------------");
var lyrOut = layer1.Compute(input, true) as ComputeCycle;
var activeColumns = layer1.GetResult("sp") as int[];
previousInputs.Add(input.ToString());
if (previousInputs.Count > (maxPrevInputs + 1))
{
previousInputs.RemoveAt(0);
}
// In the pretrained SP with HPC, the TM will quickly learn cells for patterns
// In that case the starting sequence 4-5-6 might have the sam SDR as 1-2-3-4-5-6,
// Which will result in returning of 4-5-6 instead of 1-2-3-4-5-6.
// HtmClassifier allways return the first matching sequence. Because 4-5-6 will be as first
// memorized, it will match as the first one.
if (previousInputs.Count < maxPrevInputs)
{
continue;
}
string key = GetKey(previousInputs, input, sequenceKeyPair.Key);
List <Cell> actCells;
if (lyrOut.ActiveCells.Count == lyrOut.WinnerCells.Count)
{
actCells = lyrOut.ActiveCells;
}
else
{
actCells = lyrOut.WinnerCells;
}
cls.Learn(key, actCells.ToArray());
Debug.WriteLine($"Col SDR: {Helpers.StringifyVector(lyrOut.ActivColumnIndicies)}");
Debug.WriteLine($"Cell SDR: {Helpers.StringifyVector(actCells.Select(c => c.Index).ToArray())}");
//
// If the list of predicted values from the previous step contains the currently presenting value,
// we have a match.
if (lastPredictedValues.Contains(key))
{
matches++;
Debug.WriteLine($"Match. Actual value: {key} - Predicted value: {lastPredictedValues.FirstOrDefault(key)}.");
}
else
{
Debug.WriteLine($"Missmatch! Actual value: {key} - Predicted values: {String.Join(',', lastPredictedValues)}");
}
if (lyrOut.PredictiveCells.Count > 0)
{
//var predictedInputValue = cls.GetPredictedInputValue(lyrOut.PredictiveCells.ToArray());
var predictedInputValues = cls.GetPredictedInputValues(lyrOut.PredictiveCells.ToArray(), 3);
foreach (var item in predictedInputValues)
{
Debug.WriteLine($"Current Input: {input} \t| Predicted Input: {item.PredictedInput} - {item.Similarity}");
}
lastPredictedValues = predictedInputValues.Select(v => v.PredictedInput).ToList();
}
else
{
Debug.WriteLine($"NO CELLS PREDICTED for next cycle.");
lastPredictedValues = new List <string> ();
}
}
// The first element (a single element) in the sequence cannot be predicted
double maxPossibleAccuraccy = (double)((double)sequenceKeyPair.Value.Count - 1) / (double)sequenceKeyPair.Value.Count * 100.0;
double accuracy = (double)matches / (double)sequenceKeyPair.Value.Count * 100.0;
Debug.WriteLine($"Cycle: {cycle}\tMatches={matches} of {sequenceKeyPair.Value.Count}\t {accuracy}%");
if (accuracy >= maxPossibleAccuraccy)
{
maxMatchCnt++;
Debug.WriteLine($"100% accuracy reched {maxMatchCnt} times.");
//
// Experiment is completed if we are 30 cycles long at the 100% accuracy.
if (maxMatchCnt >= 30)
{
sw.Stop();
Debug.WriteLine($"Sequence learned. The algorithm is in the stable state after 30 repeats with with accuracy {accuracy} of maximum possible {maxMatchCnt}. Elapsed sequence {sequenceKeyPair.Key} learning time: {sw.Elapsed}.");
break;
}
}
else if (maxMatchCnt > 0)
{
Debug.WriteLine($"At 100% accuracy after {maxMatchCnt} repeats we get a drop of accuracy with accuracy {accuracy}. This indicates instable state. Learning will be continued.");
maxMatchCnt = 0;
}
// This resets the learned state, so the first element starts allways from the beginning.
tm.Reset(mem);
}
}
Debug.WriteLine("------------ END ------------");
return(new HtmPredictionEngine {
Layer = layer1, Classifier = cls, Connections = mem
});
}
public void RunPowerPredictionExperiment()
{
const int inputBits = 300; /* without datetime component */ // 13420; /* with 4096 scalar bits */ // 10404 /* with 1024 bits */;
Parameters p = Parameters.getAllDefaultParameters();
p.Set(KEY.RANDOM, new ThreadSafeRandom(42));
p.Set(KEY.COLUMN_DIMENSIONS, new int[] { 2048 });
p.Set(KEY.INPUT_DIMENSIONS, new int[] { inputBits });
p.Set(KEY.CELLS_PER_COLUMN, 10 /* 50 */);
p.Set(KEY.GLOBAL_INHIBITION, true);
p.Set(KEY.CONNECTED_PERMANENCE, 0.1);
// N of 40 (40= 0.02*2048 columns) active cells required to activate the segment.
p.setNumActiveColumnsPerInhArea(0.02 * 2048);
// Activation threshold is 10 active cells of 40 cells in inhibition area.
p.setActivationThreshold(10 /*15*/);
p.setInhibitionRadius(15);
p.Set(KEY.MAX_BOOST, 0.0);
p.Set(KEY.DUTY_CYCLE_PERIOD, 100000);
p.setActivationThreshold(10);
p.setMaxNewSynapsesPerSegmentCount((int)(0.02 * 2048));
p.setPermanenceIncrement(0.17);
//p.Set(KEY.MAX_SYNAPSES_PER_SEGMENT, 32);
//p.Set(KEY.MAX_SEGMENTS_PER_CELL, 128);
//p.Set(KEY.MAX_NEW_SYNAPSE_COUNT, 200);
//p.Set(KEY.POTENTIAL_RADIUS, 700);
//p.Set(KEY.POTENTIAL_PCT, 0.5);
//p.Set(KEY.NUM_ACTIVE_COLUMNS_PER_INH_AREA, 42);
//p.Set(KEY.LOCAL_AREA_DENSITY, -1);
CortexRegion region0 = new CortexRegion("1st Region");
SpatialPoolerMT sp1 = new SpatialPoolerMT();
TemporalMemory tm1 = new TemporalMemory();
var mem = new Connections();
p.apply(mem);
sp1.Init(mem, UnitTestHelpers.GetMemory());
tm1.Init(mem);
Dictionary <string, object> settings = new Dictionary <string, object>();
CortexLayer <object, object> layer1 = new CortexLayer <object, object>("L1");
region0.AddLayer(layer1);
layer1.HtmModules.Add("sp", sp1);
HtmClassifier <double, ComputeCycle> cls = new HtmClassifier <double, ComputeCycle>();
Stopwatch sw = new Stopwatch();
sw.Start();
Train(inputBits, layer1, cls, true); // New born mode.
sw.Stop();
Debug.WriteLine($"NewBorn stage duration: {sw.ElapsedMilliseconds / 1000} s");
layer1.AddModule("tm", tm1);
sw.Start();
int hunderdAccCnt = 0;
for (int i = 0; i < 1000; i++)
{
float acc = Train(inputBits, layer1, cls, false);
Debug.WriteLine($"Accuracy = {acc}, Cycle = {i}");
if (acc == 100.0)
{
hunderdAccCnt++;
}
if (hunderdAccCnt >= 10)
{
break;
}
//tm1.reset(mem);
}
if (hunderdAccCnt >= 10)
{
Debug.WriteLine($"EXPERIMENT SUCCESS. Accurracy 100% reached.");
}
else
{
Debug.WriteLine($"Experiment FAILED!. Accurracy 100% was not reached.");
}
cls.TraceState();
sw.Stop();
Debug.WriteLine($"Training duration: {sw.ElapsedMilliseconds / 1000} s");
}
public void MyTestMethod()
{
//DIRECTORIES TO STORE INPUT AND OUTPUT FILES OF THE EXPERIMENT
//Encoder
string E_inFolder = "NoiseExperiments/Input"; //Encoder's raw input file directory
//Directory.CreateDirectory(E_inFolder);
//----------------INPUT FILE PATH-----------------
string E_inFile_train = $"{E_inFolder}\\sinusoidal.csv"; //Encoder's input file in "Training mode"
//<Robustness>
string E_inFile_robustness = $"{E_inFolder}/Noisy_N-0-2_sinusoidal.csv"; // Encoder's input file in "Testing mode - Robustness" - All the files with name of the form "Noisy_*.csv"
//</Robustness>
//< Specificity >
string E_inFile_specificity = $"{E_inFolder}/sinusoidal-specificity.csv"; // Encoder's input file in "Testing mode - Specificity"
//</ Specificity >
//------------------------------------------------
string E_outFolder = "NoiseExperiments/MyEncoderOutput"; //Encoder's graphical output (PNG format) during "Testing mode" will be created here
Directory.CreateDirectory(E_outFolder);
string E_outFolder_train = $"{E_outFolder}/train"; // Encoder's graphical output (PNG format) during "Training mode" will be created here
Directory.CreateDirectory(E_outFolder_train);
//Spatial Pooler
string SP_inFolder = "NoiseExperiments/MySPInput"; //Spatial Pooler's input file (Encoder's output (CSV format)) directory
Directory.CreateDirectory(SP_inFolder);
string SP_inFile_train = $"{SP_inFolder}/MyEncoderOut_train.csv"; //Spatial Pooler's input file during "Training mode"
string SP_inFile_robustness = $"{SP_inFolder}/MyEncoderOut_robustness.csv"; //Spatial Pooler's input file during "Testing mode - robustness"
string SP_inFile_specificity = $"{SP_inFolder}/MyEncoderOut_specificity.csv"; //Spatial Pooler's input file during "Testing mode - specificity"
string SP_outFolder = "MySPOutput"; //Spatial Pooler's graphical output (PNG format) will be stored here
Directory.CreateDirectory(SP_outFolder);
string SP_outFolder_compare = $"{SP_outFolder}/compare_445"; //This folder containing CSV files, which show Hamming distance between Spatial Pooler's output in "Training mode" and "Testing Mode"
Directory.CreateDirectory(SP_outFolder_compare);
//--------------------OUTPUT FILE PATH-------------------------
//<Robustness>
string SP_outFile_robustness = $"{SP_outFolder_compare}/compare_N-0-2.csv"; //The final output file in "Robustness test"
//</Robustness>
//<Specificity>
string SP_outFile_specificity = $"{SP_outFolder_compare}/compare_specificity.csv"; //The final output file in "Specificity test"
//</Specificity>
//-------------------------------------------------------------
//-------------------------------------------------------
//| HTM PARAMETERS |
//-------------------------------------------------------
const int E_outBits = 445; //Number of Scalar Encoder's output bits
const int columnsNumber = 2048; //Number of Spatial Pooler's output columns
Parameters p = Parameters.getAllDefaultParameters();
p.Set(KEY.RANDOM, new ThreadSafeRandom(42));
//------------------SPATIAL POOLER PARAMETERS-----------------
p.Set(KEY.INPUT_DIMENSIONS, new int[] { E_outBits });
p.Set(KEY.POTENTIAL_RADIUS, -1);
p.Set(KEY.POTENTIAL_PCT, 1);
p.Set(KEY.GLOBAL_INHIBITION, true);
p.Set(KEY.INHIBITION_RADIUS, 15);
//Leave it
p.Set(KEY.LOCAL_AREA_DENSITY, -1.0);
p.Set(KEY.NUM_ACTIVE_COLUMNS_PER_INH_AREA, 0.02 * columnsNumber);
p.Set(KEY.STIMULUS_THRESHOLD, 0.5);
p.Set(KEY.SYN_PERM_INACTIVE_DEC, 0.008);
p.Set(KEY.SYN_PERM_ACTIVE_INC, 0.01);
p.Set(KEY.SYN_PERM_CONNECTED, 0.10);
//Leave it
p.Set(KEY.SYN_PERM_BELOW_STIMULUS_INC, 0.01);
p.Set(KEY.SYN_PERM_TRIM_THRESHOLD, 0.05);
p.Set(KEY.MIN_PCT_OVERLAP_DUTY_CYCLES, 0.001);
p.Set(KEY.MIN_PCT_ACTIVE_DUTY_CYCLES, 0.001);
p.Set(KEY.DUTY_CYCLE_PERIOD, 100);
p.Set(KEY.MAX_BOOST, 10 /*10.0*/);
p.Set(KEY.WRAP_AROUND, true);
//p.Set(KEY.LEARN, true);
//-------------------TEMPORAL MEMORY PARAMETERS----------------
p.Set(KEY.COLUMN_DIMENSIONS, new int[] { columnsNumber });
p.Set(KEY.CELLS_PER_COLUMN, 32);
p.Set(KEY.ACTIVATION_THRESHOLD, 10);
p.Set(KEY.LEARNING_RADIUS, 10);
p.Set(KEY.MIN_THRESHOLD, 9);
p.Set(KEY.MAX_NEW_SYNAPSE_COUNT, 20);
p.Set(KEY.MAX_SYNAPSES_PER_SEGMENT, 225);
p.Set(KEY.MAX_SEGMENTS_PER_CELL, 225);
p.Set(KEY.INITIAL_PERMANENCE, 0.21);
p.Set(KEY.CONNECTED_PERMANENCE, 0.5);
p.Set(KEY.PERMANENCE_INCREMENT, 0.10);
p.Set(KEY.PERMANENCE_DECREMENT, 0.10);
p.Set(KEY.PREDICTED_SEGMENT_DECREMENT, 0.1);
//p.Set(KEY.LEARN, true);
//---------------------------------------------------
//| UNIT TEST |
//---------------------------------------------------
//Initiating HTM modules
SpatialPoolerMT sp1 = new SpatialPoolerMT();
TemporalMemory tm1 = new TemporalMemory();
var mem = new Connections();
p.apply(mem);
sp1.Init(mem, UnitTestHelpers.GetMemory());
tm1.Init(mem);
HtmClassifier <double, ComputeCycle> cls = new HtmClassifier <double, ComputeCycle>();
//-------------------ENCODING INPUTS----------------------
Encoding(E_inFile_train, SP_inFile_train, E_outFolder_train, E_outBits);
//--------------------TRAINING MODE---------------------
//SP training
for (int j = 0; j < 5; j++)
{
using (StreamReader sr = new StreamReader(SP_inFile_train))
{
string line;
while ((line = sr.ReadLine()) != null)
{
string[] tokens = line.Split(",");
int[] SP_input = new int[E_outBits];
for (int i = 0; i < E_outBits; i++)
{
if (tokens[i + 1] == "0")
{
SP_input[i] = 0;
}
else
{
SP_input[i] = 1;
}
}
for (int i = 0; i < 3; i++)
{
sp1.Compute(SP_input, true);
}
}
}
}
Debug.WriteLine("-----------------------------------------------------");
Debug.WriteLine("|-----------------FINISHED TRAINING-----------------|");
Debug.WriteLine("-----------------------------------------------------");
//TM + SP training
double lastPredictedValue = 0.0;
for (int j = 0; j < 20; j++)
{
using (StreamReader sr = new StreamReader(SP_inFile_train))
{
string line;
while ((line = sr.ReadLine()) != null)
{
string[] tokens = line.Split(",");
int[] SP_input = new int[E_outBits];
for (int i = 0; i < E_outBits; i++)
{
if (tokens[i + 1] == "0")
{
SP_input[i] = 0;
}
else
{
SP_input[i] = 1;
}
}
var SP_res = sp1.Compute(SP_input, true);
var TM_res = tm1.Compute(SP_res, true) as ComputeCycle;
double input = Convert.ToDouble(tokens[0], CultureInfo.InvariantCulture);
Debug.WriteLine($"Input: {input} - Predicted: {lastPredictedValue}");
//cls.Learn(input, TM_res.ActiveCells.ToArray(), TM_res.PredictiveCells.ToArray());
lastPredictedValue = cls.GetPredictedInputValue(TM_res.PredictiveCells.ToArray());
}
}
}
}
/// <summary>
/// Processes the test cases for Noise Test taking the necessary parameters.
/// It's the parent method which calls other utility methods to create the input vectors and the outputs.
/// </summary>
/// <param name="inputSequence">An array of double, consisting the starting indexes for each input vector</param>
/// <param name="inputs">
/// A parameter of the class "InputParameters", which contains all the input parameters needed as properties which can be set in a test case
/// </param>
/// <returns>Returns nothing</returns>
public void ProcessTestCase(List <double[]> inputSequences, InputParameters inputs)
{
string path = Directory.GetCurrentDirectory();
string parentDir = path.Substring(0, path.IndexOf("bin") - 1);
string resultDir = parentDir.Substring(0, parentDir.LastIndexOf(@"\"));
string timeStamp = DateTime.Now.ToString("yyyyMMddHHmmss");
string outFolder = resultDir + @"\SpatialPooler_Results\" + timeStamp + @"\Output\";
string inFolder = resultDir + @"\SpatialPooler_Results\" + timeStamp + @"\InputVectors";
if (!Directory.Exists(outFolder))
{
Directory.CreateDirectory(outFolder);
}
if (!Directory.Exists(inFolder + @"\"))
{
Directory.CreateDirectory(inFolder);
}
//int radius = 0;
var parameters = GetDefaultParams();
parameters.Set(KEY.POTENTIAL_RADIUS, 64 * 64);
parameters.Set(KEY.POTENTIAL_PCT, 1.0);
parameters.Set(KEY.GLOBAL_INHIBITION, false);
parameters.Set(KEY.STIMULUS_THRESHOLD, 0.5);
parameters.Set(KEY.INHIBITION_RADIUS, (int)0.25 * 64 * 64);
parameters.Set(KEY.LOCAL_AREA_DENSITY, -1);
parameters.Set(KEY.NUM_ACTIVE_COLUMNS_PER_INH_AREA, 0.1 * 64 * 64);
parameters.Set(KEY.DUTY_CYCLE_PERIOD, 1000000);
parameters.Set(KEY.MAX_BOOST, 5);
Console.WriteLine("Fetched all default parameters\n");
parameters.setInputDimensions(new int[] { 32, 32 });
parameters.setColumnDimensions(new int[] { 64, 64 });
parameters.setNumActiveColumnsPerInhArea(0.02 * 64 * 64);
var sp = new SpatialPoolerMT();
var mem = new Connections();
parameters.apply(mem);
Console.WriteLine("\nConfiguring the Inputs...\n");
sp.Init(mem, GetInMemoryDictionary());
int outFolderCount = 0;
int compareIndex = Convert.ToInt32(inputs.getCompareNumber());
double[][] recordOutput = null;
double[] hammingDistance = null;
foreach (double[] inputSequence in inputSequences)
{
outFolderCount++;
double minVal = 0.0;
for (int i = 0; i < inputSequence.Length; i++)
{
if (i == 0)
{
minVal = inputSequence[i];
}
else if (inputSequence[i] < minVal)
{
minVal = inputSequence[i];
}
}
minVal -= 1.0;
Console.WriteLine("\nGetting the Input Vectors...\n");
var inputVectors = GetEncodedSequence(inputSequence, minVal, inputs.getMaxIndex(), inputs, inFolder);
int count = 1;
//string output = String.Empty;
int max = 0;
for (int i = 0; i < inputVectors.Count; i++)
{
hammingDistance = null;
//output = String.Empty;
Console.WriteLine("Computing the Output for the vector no: " + count.ToString() + "...\n");
var activeArray = sp.Compute(inputVectors[i], true) as int[];
for (int j = 0; j < activeArray.Length; j++)
{
if (activeArray[j] > max)
{
max = activeArray[j];
}
}
//var str = Helpers.StringifyVector(activeArray);
int rows = Convert.ToInt32(Math.Ceiling(Math.Sqrt(Convert.ToDouble(max))));
int counter = 0;
int index = 0;
int[,] outTwoDArray = new int[rows, rows];
for (int j = 0; j < rows; j++)
{
for (int k = 0; k < rows; k++)
{
outTwoDArray[j, k] = 0;
}
}
for (int j = 0; j < rows; j++)
{
for (int k = 0; k < rows; k++)
{
counter++;
if (index < activeArray.Length && activeArray[index] == counter)
{
index++;
outTwoDArray[j, k] = 1;
}
}
}
double[][] comparingArray = new double[rows][];
for (int j = 0; j < rows; j++)
{
comparingArray[j] = new double[rows];
for (int k = 0; k < rows; k++)
{
comparingArray[j][k] = Convert.ToDouble(outTwoDArray[j, k]);
}
}
int[,] record2Darray = null;
if (inputSequence[i] == compareIndex)
{
if (recordOutput != null)
{
hammingDistance = MathHelpers.GetHammingDistance(recordOutput, comparingArray, false);
record2Darray = new int[recordOutput.Length, recordOutput.Length];
for (int j = 0; j < recordOutput.Length; j++)
{
for (int k = 0; k < recordOutput.Length; k++)
{
record2Darray[j, k] = Convert.ToInt32(recordOutput[j][k]);
}
}
}
recordOutput = new double[rows][];
for (int j = 0; j < rows; j++)
{
recordOutput[j] = new double[rows];
for (int k = 0; k < rows; k++)
{
recordOutput[j][k] = comparingArray[j][k];
}
}
}
if (hammingDistance != null)
{
int rowHam = Convert.ToInt32(Math.Ceiling(Math.Sqrt(hammingDistance.Length)));
int[,] hammingArray = new int[rowHam, rowHam];
int limit = 0;
for (int j = 0; j < rowHam; j++)
{
for (int k = 0; k < rowHam; k++)
{
if (limit < hammingDistance.Length)
{
//hj
hammingArray[j, k] = Convert.ToInt32(hammingDistance[limit]);
limit++;
}
}
}
int compare_no = 1;
if (!File.Exists($"{outFolder}\\Compare_{compareIndex}.png"))
{
DrawBitmapHamming(hammingArray, 1024, 1024, $"{outFolder}\\Compare_{compareIndex}.png", $"Compare_{compareIndex}");
DrawBitmapOverlap(record2Darray, outTwoDArray, 1024, 1024, $"{outFolder}\\Overlap_{compareIndex}.png", $"Overlap_{compareIndex}");
}
else
{
while (File.Exists($"{outFolder}\\Compare_{compareIndex}_{compare_no}.png"))
{
compare_no++;
}
DrawBitmapHamming(hammingArray, 1024, 1024, $"{outFolder}\\Compare_{compareIndex}_{compare_no}.png", $"Compare_{compareIndex}");
compare_no = 1;
while (File.Exists($"{outFolder}\\Overlap_{compareIndex}_{compare_no}.png"))
{
compare_no++;
}
DrawBitmapOverlap(record2Darray, outTwoDArray, 1024, 1024, $"{outFolder}\\Overlap_{compareIndex}_{compare_no}.png", $"Overlap_{compareIndex}");
}
}
if (!Directory.Exists(outFolder + @"\\" + outFolderCount.ToString()))
{
Directory.CreateDirectory(outFolder + @"\\" + outFolderCount.ToString());
}
int[,] out2dimArray = ArrayUtils.Transpose(outTwoDArray);
NeoCortexUtils.DrawBitmap(out2dimArray, 1024, 1024, $"{outFolder}\\{outFolderCount}\\{count}.png", Color.Black, Color.Green, text: inputSequence[i].ToString());
//File.WriteAllLines(outFolder + count.ToString() + ".txt", new string[] { str });
Console.WriteLine("Output is recorded in the path: " + outFolder + count.ToString() + ".txt\n");
count++;
}
}
}
public void CreateSdrsTest()
{
var colDims = new int[] { 64, 64 };
int numOfCols = 64 * 64;
string trainingFolder = @"..\..\..\TestFiles\Sdr";
int imgSize = 28;
var trainingImages = Directory.GetFiles(trainingFolder, "*.jpeg");
Directory.CreateDirectory($"{nameof(CreateSdrsTest)}");
int counter = 0;
bool isInStableState = false;
// HTM parameters
HtmConfig htmConfig = new HtmConfig(new int[] { imgSize, imgSize }, new int[] { 64, 64 })
{
PotentialRadius = 10,
PotentialPct = 1,
GlobalInhibition = true,
LocalAreaDensity = -1.0,
NumActiveColumnsPerInhArea = 0.02 * numOfCols,
StimulusThreshold = 0.0,
SynPermInactiveDec = 0.008,
SynPermActiveInc = 0.05,
SynPermConnected = 0.10,
MinPctOverlapDutyCycles = 1.0,
MinPctActiveDutyCycles = 0.001,
DutyCyclePeriod = 100,
MaxBoost = 10.0,
RandomGenSeed = 42,
Random = new ThreadSafeRandom(42)
};
Connections connections = new Connections(htmConfig);
HomeostaticPlasticityController hpa = new HomeostaticPlasticityController(connections, trainingImages.Length * 50, (isStable, numPatterns, actColAvg, seenInputs) =>
{
isInStableState = true;
Debug.WriteLine($"Entered STABLE state: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
});
SpatialPooler sp = new SpatialPoolerMT(hpa);
sp.Init(connections);
string outFolder = nameof(CreateSdrsTest);
Directory.CreateDirectory(outFolder);
while (true)
{
counter++;
Dictionary <string, int[]> sdrs = new Dictionary <string, int[]>();
Dictionary <string, int[]> inputVectors = new Dictionary <string, int[]>();
foreach (var trainingImage in trainingImages)
{
FileInfo fI = new FileInfo(trainingImage);
string outputHamDistFile = $"{outFolder}\\image-{fI.Name}_hamming.txt";
string outputActColFile = $"{outFolder}\\image{fI.Name}_activeCol.txt";
string outputActColFile1 = $"{outFolder}\\image{fI.Name}_activeCol.csv";
using (StreamWriter swActCol = new StreamWriter(outputActColFile))
{
using (StreamWriter swActCol1 = new StreamWriter(outputActColFile1))
{
int[] activeArray = new int[numOfCols];
string testName = $"{outFolder}\\{fI.Name}";
string inputBinaryImageFile = NeoCortexUtils.BinarizeImage($"{trainingImage}", imgSize, testName);
// Read input csv file into array
int[] inputVector = NeoCortexUtils.ReadCsvIntegers(inputBinaryImageFile).ToArray();
List <double[, ]> overlapArrays = new List <double[, ]>();
List <double[, ]> bostArrays = new List <double[, ]>();
sp.compute(inputVector, activeArray, true);
var activeCols = ArrayUtils.IndexWhere(activeArray, (el) => el == 1);
if (isInStableState)
{
CalculateResult(sdrs, inputVectors, numOfCols, activeCols, outFolder, trainingImage, inputVector);
overlapArrays.Add(ArrayUtils.Make2DArray <double>(ArrayUtils.ToDoubleArray(connections.Overlaps), colDims[0], colDims[1]));
bostArrays.Add(ArrayUtils.Make2DArray <double>(connections.BoostedOverlaps, colDims[0], colDims[1]));
var activeStr = Helpers.StringifyVector(activeArray);
swActCol.WriteLine("Active Array: " + activeStr);
int[,] twoDimenArray = ArrayUtils.Make2DArray <int>(activeArray, colDims[0], colDims[1]);
twoDimenArray = ArrayUtils.Transpose(twoDimenArray);
List <int[, ]> arrays = new List <int[, ]>();
arrays.Add(twoDimenArray);
arrays.Add(ArrayUtils.Transpose(ArrayUtils.Make2DArray <int>(inputVector, (int)Math.Sqrt(inputVector.Length), (int)Math.Sqrt(inputVector.Length))));
//Calculating the max value of the overlap in the OverlapArray
int max = SdrRepresentation.TraceColumnsOverlap(overlapArrays, swActCol1, fI.Name);
int red = Convert.ToInt32(max * 0.80); // Value above this threshould would be red and below this will be yellow
int green = Convert.ToInt32(max * 0.50); // Value above this threshould would be yellow and below this will be green
string outputImage = $"{outFolder}\\cycle-{counter}-{fI.Name}";
NeoCortexUtils.DrawBitmaps(arrays, outputImage, Color.Yellow, Color.Gray, OutImgSize, OutImgSize);
NeoCortexUtils.DrawHeatmaps(overlapArrays, $"{outputImage}_overlap.png", 1024, 1024, red, red, green);
if (sdrs.Count == trainingImages.Length)
{
return;
}
}
}
}
}
}
}
//[DataRow("Box")]
//[DataRow("Horizontal")]
public void ImageSimilarityExperiment(string inputPrefix)
{
//int stableStateCnt = 100;
double minOctOverlapCycles = 1.0;
double maxBoost = 10.0;
//int inputBits = 100;
var colDims = new int[] { 64, 64 };
int numOfCols = 64 * 64;
//int numColumns = colDims[0];
string trainingFolder = "Similarity\\TestFiles";
int imgSize = 28;
//var colDims = new int[] { 64, 64 };
//int numOfActCols = colDims[0] * colDims[1];
string TestOutputFolder = $"Output-{nameof(ImageSimilarityExperiment)}";
var trainingImages = Directory.GetFiles(trainingFolder, $"{inputPrefix}*.png");
Directory.CreateDirectory($"{nameof(ImageSimilarityExperiment)}");
int counter = 0;
//var parameters = GetDefaultParams();
////parameters.Set(KEY.DUTY_CYCLE_PERIOD, 20);
////parameters.Set(KEY.MAX_BOOST, 1);
////parameters.setInputDimensions(new int[] { imageSize[imSizeIndx], imageSize[imSizeIndx] });
////parameters.setColumnDimensions(new int[] { topologies[topologyIndx], topologies[topologyIndx] });
////parameters.setNumActiveColumnsPerInhArea(0.02 * numOfActCols);
//parameters.Set(KEY.NUM_ACTIVE_COLUMNS_PER_INH_AREA, 0.06 * 4096); // TODO. Experiment with different sizes
//parameters.Set(KEY.POTENTIAL_RADIUS, imgSize * imgSize);
//parameters.Set(KEY.POTENTIAL_PCT, 1.0);
//parameters.Set(KEY.GLOBAL_INHIBITION, true); // TODO: Experiment with local inhibition too. Note also the execution time of the experiment.
//// Num of active synapces in order to activate the column.
//parameters.Set(KEY.STIMULUS_THRESHOLD, 50.0);
//parameters.Set(KEY.SYN_PERM_INACTIVE_DEC, 0.008);
//parameters.Set(KEY.SYN_PERM_ACTIVE_INC, 0.05);
//parameters.Set(KEY.INHIBITION_RADIUS, (int)0.02 * imgSize * imgSize); // TODO. check if this has influence in a case of the global inhibition. ALso check how this parameter influences the similarity of SDR.
//parameters.Set(KEY.SYN_PERM_CONNECTED, 0.2);
//parameters.Set(KEY.MIN_PCT_OVERLAP_DUTY_CYCLES, 0.001);
//parameters.Set(KEY.MIN_PCT_ACTIVE_DUTY_CYCLES, 0.001);
//parameters.Set(KEY.DUTY_CYCLE_PERIOD, 1000);
//parameters.Set(KEY.MAX_BOOST, 100);
//parameters.Set(KEY.WRAP_AROUND, true);
//parameters.Set(KEY.SEED, 1969);
//parameters.setInputDimensions(new int[] { imgSize, imgSize });
//parameters.setColumnDimensions(colDims);
Parameters p = Parameters.getAllDefaultParameters();
p.Set(KEY.RANDOM, new ThreadSafeRandom(42));
p.Set(KEY.INPUT_DIMENSIONS, new int[] { imgSize, imgSize });
p.Set(KEY.COLUMN_DIMENSIONS, colDims);
p.Set(KEY.CELLS_PER_COLUMN, 10);
p.Set(KEY.MAX_BOOST, maxBoost);
p.Set(KEY.DUTY_CYCLE_PERIOD, 50);
p.Set(KEY.MIN_PCT_OVERLAP_DUTY_CYCLES, minOctOverlapCycles);
// Global inhibition
// N of 40 (40= 0.02*2048 columns) active cells required to activate the segment.
p.Set(KEY.GLOBAL_INHIBITION, true);
p.setNumActiveColumnsPerInhArea(0.02 * numOfCols);
p.Set(KEY.POTENTIAL_RADIUS, (int)(0.8 * imgSize * imgSize));
p.Set(KEY.LOCAL_AREA_DENSITY, -1); // In a case of global inhibition.
//p.setInhibitionRadius( Automatically set on the columns pace in a case of global inhibition.);
// Activation threshold is 10 active cells of 40 cells in inhibition area.
p.setActivationThreshold(10);
// Max number of synapses on the segment.
p.setMaxNewSynapsesPerSegmentCount((int)(0.02 * numOfCols));
bool isInStableState = false;
var mem = new Connections();
p.apply(mem);
HomeostaticPlasticityController hpa = new HomeostaticPlasticityController(mem, trainingImages.Length * 50, (isStable, numPatterns, actColAvg, seenInputs) =>
{
// Event should only be fired when entering the stable state.
// Ideal SP should never enter unstable state after stable state.
Assert.IsTrue(isStable);
Assert.IsTrue(numPatterns == trainingImages.Length);
isInStableState = true;
Debug.WriteLine($"Entered STABLE state: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}, requiredSimilarityThreshold: 0.975);
SpatialPooler sp = new SpatialPoolerMT(hpa);
sp.Init(mem, UnitTestHelpers.GetMemory());
string outFolder = $"{TestOutputFolder}\\{inputPrefix}";
Directory.CreateDirectory(outFolder);
string outputHamDistFile = $"{outFolder}\\digit{inputPrefix}_hamming.txt";
string outputActColFile = $"{outFolder}\\digit{inputPrefix}_activeCol.txt";
using (StreamWriter swHam = new StreamWriter(outputHamDistFile))
{
using (StreamWriter swActCol = new StreamWriter(outputActColFile))
{
int cycle = 0;
Dictionary <string, int[]> sdrs = new Dictionary <string, int[]>();
Dictionary <string, int[]> inputVectors = new Dictionary <string, int[]>();
while (true)
{
foreach (var trainingImage in trainingImages)
{
int[] activeArray = new int[numOfCols];
FileInfo fI = new FileInfo(trainingImage);
string outputImage = $"{outFolder}\\{inputPrefix}_cycle_{counter}_{fI.Name}";
string testName = $"{outFolder}\\{inputPrefix}_{fI.Name}";
string inputBinaryImageFile = NeoCortexUtils.BinarizeImage($"{trainingImage}", imgSize, testName);
// Read input csv file into array
int[] inputVector = NeoCortexUtils.ReadCsvIntegers(inputBinaryImageFile).ToArray();
int[] oldArray = new int[activeArray.Length];
List <double[, ]> overlapArrays = new List <double[, ]>();
List <double[, ]> bostArrays = new List <double[, ]>();
sp.compute(inputVector, activeArray, true);
var activeCols = ArrayUtils.IndexWhere(activeArray, (el) => el == 1);
Debug.WriteLine($"Cycle: {cycle++} - Input: {trainingImage}");
Debug.WriteLine($"{Helpers.StringifyVector(activeCols)}\n");
if (isInStableState)
{
if (sdrs.Count == trainingImages.Length)
{
CalculateSimilarity(sdrs, inputVectors);
return;
}
var distance = MathHelpers.GetHammingDistance(oldArray, activeArray, true);
//var similarity = MathHelpers.CalcArraySimilarity(oldArray, activeArray, true);
sdrs.Add(trainingImage, activeCols);
inputVectors.Add(trainingImage, inputVector);
swHam.WriteLine($"{counter++}|{distance} ");
oldArray = new int[numOfCols];
activeArray.CopyTo(oldArray, 0);
overlapArrays.Add(ArrayUtils.Make2DArray <double>(ArrayUtils.ToDoubleArray(mem.Overlaps), colDims[0], colDims[1]));
bostArrays.Add(ArrayUtils.Make2DArray <double>(mem.BoostedOverlaps, colDims[0], colDims[1]));
var activeStr = Helpers.StringifyVector(activeArray);
swActCol.WriteLine("Active Array: " + activeStr);
int[,] twoDimenArray = ArrayUtils.Make2DArray <int>(activeArray, colDims[0], colDims[1]);
twoDimenArray = ArrayUtils.Transpose(twoDimenArray);
List <int[, ]> arrays = new List <int[, ]>();
arrays.Add(twoDimenArray);
arrays.Add(ArrayUtils.Transpose(ArrayUtils.Make2DArray <int>(inputVector, (int)Math.Sqrt(inputVector.Length), (int)Math.Sqrt(inputVector.Length))));
NeoCortexUtils.DrawBitmaps(arrays, outputImage, Color.Yellow, Color.Gray, OutImgSize, OutImgSize);
NeoCortexUtils.DrawHeatmaps(overlapArrays, $"{outputImage}_overlap.png", 1024, 1024, 150, 50, 5);
NeoCortexUtils.DrawHeatmaps(bostArrays, $"{outputImage}_boost.png", 1024, 1024, 150, 50, 5);
}
}
}
}
}
}
private void RunExperiment(int inputBits, HtmConfig cfgL4, EncoderBase encoder, List <double> inputValues, HtmConfig cfgL2)
{
Stopwatch swL2 = new Stopwatch();
int maxMatchCnt = 0;
bool learn = true;
bool isSP4Stable = false;
bool isSP2STable = false;
var memL4 = new Connections(cfgL4);
var memL2 = new Connections(cfgL2);
var numInputs = inputValues.Distinct <double>().ToList().Count;
HtmClassifier <string, ComputeCycle> cls = new HtmClassifier <string, ComputeCycle>();
layerL4 = new CortexLayer <object, object>("L4");
layerL2 = new CortexLayer <object, object>("L2");
tm4 = new TemporalMemoryMT();
tm2 = new TemporalMemoryMT();
// HPC for Layer 4 SP
HomeostaticPlasticityController hpa_sp_L4 = new HomeostaticPlasticityController(memL4, numInputs * 50, (isStable, numPatterns, actColAvg, seenInputs) =>
{
if (isStable)
{
Debug.WriteLine($"SP L4 STABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
else
{
Debug.WriteLine($"SP L4 INSTABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
learn = isSP4Stable = isStable;
}, numOfCyclesToWaitOnChange: 50);
// HPC for Layer 2 SP
HomeostaticPlasticityController hpa_sp_L2 = new HomeostaticPlasticityController(memL2, numInputs * 50, (isStable, numPatterns, actColAvg, seenInputs) =>
{
if (isStable)
{
Debug.WriteLine($"SP L2 STABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
else
{
Debug.WriteLine($"SP L2 INSTABLE: Patterns: {numPatterns}, Inputs: {seenInputs}, iteration: {seenInputs / numPatterns}");
}
learn = isSP2STable = isStable;
cls.ClearState();
}, numOfCyclesToWaitOnChange: 50);
SpatialPooler sp4 = new SpatialPoolerMT(hpa_sp_L4);
SpatialPooler sp2 = new SpatialPoolerMT(hpa_sp_L2);
sp4.Init(memL4);
sp2.Init(memL2);
// memL2.TraceInputPotential();
tm4.Init(memL4);
tm2.Init(memL2);
layerL4.HtmModules.Add("encoder", encoder);
layerL4.HtmModules.Add("sp", sp4);
layerL4.HtmModules.Add("tm", tm4);
layerL2.HtmModules.Add("sp", sp2);
layerL2.HtmModules.Add("tm", tm2);
int[] inpCellsL4ToL2 = new int[cfgL4.CellsPerColumn * cfgL4.NumColumns];
double[] inputs = inputValues.ToArray();
int[] prevActiveCols = new int[0];
int cycle = 0;
int matches = 0;
string lastPredictedValue = "0";
int maxCycles = 3500;
int maxPrevInputs = inputValues.Count - 1;
List <string> previousInputs = new List <string>();
//
// Training SP at Layer 4 to get stable. New-born stage.
//
using (StreamWriter swL4Sdrs = new StreamWriter($"L4-SDRs-in_{cfgL2.NumInputs}-col_{cfgL2.NumColumns}-r_{cfgL2.PotentialRadius}.txt"))
{
using (StreamWriter sw = new StreamWriter($"in_{cfgL2.NumInputs}-col_{cfgL2.NumColumns}-r_{cfgL2.PotentialRadius}.txt"))
{
for (int i = 0; i < maxCycles; i++)
{
matches = 0;
cycle = i;
Debug.WriteLine($"-------------- Newborn Cycle {cycle} at L4 SP region ---------------");
foreach (var input in inputs)
{
Debug.WriteLine($" INPUT: '{input}'\tCycle:{cycle}");
Debug.Write("L4: ");
//L4 SP pre train
layerL4.Compute(input, learn);
InitArray(inpCellsL4ToL2, 0);
if (isSP4Stable)
{
var cellSdrL4Indexes = memL4.ActiveCells.Select(c => c.Index).ToArray();
// Write SDR as output of L4 and input of L2
swL4Sdrs.WriteLine($"{input} - {Helpers.StringifyVector(cellSdrL4Indexes)}");
// Set the output active cell array
ArrayUtils.SetIndexesTo(inpCellsL4ToL2, cellSdrL4Indexes, 1);
Debug.WriteLine($"L4 out sdr: {Helpers.StringifyVector(cellSdrL4Indexes)}");
Debug.WriteLine("L2: ");
swL2.Restart();
/// <summary>
/// This part is for to make SP of Layer2 stable thourgh help
/// of HPC Boosting Algo from new born stage.
/// </summary>
layerL2.Compute(inpCellsL4ToL2, true);
swL2.Stop();
Debug.WriteLine($"{swL2.ElapsedMilliseconds / 1000}");
sw.WriteLine($"{swL2.ElapsedMilliseconds / 1000}");
sw.Flush();
var overlaps = ArrayUtils.IndexWhere(memL2.Overlaps, o => o > 0);
var strOverlaps = Helpers.StringifyVector(overlaps);
Debug.WriteLine($"Potential columns: {overlaps.Length}, overlaps: {strOverlaps}");
}
}
if (isSP4Stable && isSP2STable)
{
break;
}
}
}
}
// SP+TM at L4
for (int i = 0; i < maxCycles; i++)
{
matches = 0;
cycle = i;
Debug.WriteLine($"-------------- L4 TM Train region Cycle {cycle} ---------------");
foreach (var input in inputs)
{
Debug.WriteLine($"-------------- {input} ---------------");
var layerL4Out = layerL4.Compute(input, learn) as ComputeCycle;
previousInputs.Add(input.ToString());
if (previousInputs.Count > (maxPrevInputs + 1))
{
previousInputs.RemoveAt(0);
}
if (previousInputs.Count < maxPrevInputs)
{
continue;
}
string key = GetKey(previousInputs, input);
List <Cell> actCells;
if (layerL4Out.ActiveCells.Count == layerL4Out.WinnerCells.Count)
{
// SP+TM at L2
Debug.WriteLine($"-------------- L2 TM Train region Cycle {cycle} ---------------");
// Reset tha array
InitArray(inpCellsL4ToL2, 0);
var cellSdrL4Indexes = memL4.ActiveCells.Select(c => c.Index).ToArray();
// Set the output active cell array
ArrayUtils.SetIndexesTo(inpCellsL4ToL2, cellSdrL4Indexes, 1);
var layerL2Out = layerL2.Compute(inpCellsL4ToL2, true) as ComputeCycle;
int[] overlaps = ArrayUtils.IndexWhere(memL2.Overlaps, o => o > 0);
var strOverlaps = Helpers.StringifyVector(overlaps);
Debug.WriteLine($"Potential columns: {overlaps.Length}, overlaps: {strOverlaps}");
if (layerL2Out.ActiveCells.Count == layerL2Out.WinnerCells.Count)
{
actCells = layerL2Out.ActiveCells;
}
else
{
actCells = layerL2Out.WinnerCells;
}
cls.Learn(key, actCells.ToArray());
if (key == lastPredictedValue)
{
matches++;
Debug.WriteLine($"Match. Actual value: {key} - Predicted value: {lastPredictedValue}");
}
else
{
Debug.WriteLine($"Missmatch! Actual value: {key} - Predicted value: {lastPredictedValue}");
}
if (layerL2Out.PredictiveCells.Count > 0)
{
var predictedInputValue = cls.GetPredictedInputValue(layerL2Out.PredictiveCells.ToArray());
Debug.WriteLine($"Current Input: {input} \t| Predicted Input: {predictedInputValue}");
lastPredictedValue = predictedInputValue;
}
else
{
Debug.WriteLine($"NO CELLS PREDICTED for next cycle.");
lastPredictedValue = String.Empty;
}
}
}
double accuracy = (double)matches / (double)inputs.Length * 100.0;
Debug.WriteLine($"Cycle: {cycle}\tMatches={matches} of {inputs.Length}\t {accuracy}%");
if (accuracy >= 100.0)
{
maxMatchCnt++;
Debug.WriteLine($"100% accuracy reched {maxMatchCnt} times.");
//
// Experiment is completed if we are 20 cycles long at the 100% accuracy.
if (maxMatchCnt >= 20)
{
Debug.WriteLine($"Exit experiment in the stable state after 20 repeats with 100% of accuracy.");
learn = false;
break;
}
}
else if (maxMatchCnt > 0)
{
Debug.WriteLine($"At 100% accuracy after {maxMatchCnt} repeats we get a drop of accuracy with {accuracy}. This indicates instable state. Learning will be continued.");
maxMatchCnt = 0;
}
}
}
public void RunGaussianNoiseExperiment()
{
const int E_outBits = 423;
const int columnsNumber = 2048;
int[] SP_Result = null;
int[] SP_NoisyResult = null;
List <int[]> SP_Result_List = new List <int[]>();
List <int[]> SP_NoisyResult_List = new List <int[]>();
double[] ham_total = { 0, 0, 0, 0, 0, 0, 0, 0, 0 };
double[] ham_avg = { 0, 0, 0, 0, 0, 0, 0, 0, 0 };
double ham_upper;
double ham_lower;
int number_training_set = 41; // Integer numbers range from -20 to 20 with step of 1.
int number_testing_set = 401; // Decimal numbers range from -20 to 20 with step of 0.1.
string experimentFolder = nameof(GaussianNoiseExperiment);
// string SP_noisyinFolder = nameof(GoussianNoiseExperiment);
Directory.CreateDirectory(experimentFolder);
// Directory.CreateDirectory(SP_noisyinFolder);
string SP_inFile = $"{experimentFolder}\\MyEncoderOut.csv";
string SP_noisyinFile = $"{experimentFolder}\\MyNoisyEncoderOut.csv";
//string SP_outFolder = "MySPOutput";
//string SP_noisyoutFolder = "MyNoisySPOutput";
//Directory.CreateDirectory(SP_outFolder);
//Directory.CreateDirectory(SP_noisyoutFolder);
string SP_outFile = $"{experimentFolder}\\MySPOut.csv";
string SP_noisyoutFile = $"{experimentFolder}\\MyNoisySPOut.csv";
//string testFolder = "MyDraftFoler";
//Directory.CreateDirectory(testFolder);
//-------------------------------------------------------
//| PARAMETERS |
//-------------------------------------------------------
Parameters p = Parameters.getAllDefaultParameters();
p.Set(KEY.RANDOM, new ThreadSafeRandom(42));
//------------------SPATIAL POOLER PARAMETERS-----------------
p.Set(KEY.INPUT_DIMENSIONS, new int[] { E_outBits });
p.Set(KEY.POTENTIAL_RADIUS, -1);
p.Set(KEY.POTENTIAL_PCT, 0.75);
p.Set(KEY.GLOBAL_INHIBITION, true);
p.Set(KEY.INHIBITION_RADIUS, 15);
p.Set(KEY.LOCAL_AREA_DENSITY, -1.0);
p.Set(KEY.NUM_ACTIVE_COLUMNS_PER_INH_AREA, 0.02 * columnsNumber);
p.Set(KEY.STIMULUS_THRESHOLD, 5);
p.Set(KEY.SYN_PERM_INACTIVE_DEC, 0.008);
p.Set(KEY.SYN_PERM_ACTIVE_INC, 0.05);
p.Set(KEY.SYN_PERM_CONNECTED, 0.10);
p.Set(KEY.SYN_PERM_BELOW_STIMULUS_INC, 0.01);
p.Set(KEY.SYN_PERM_TRIM_THRESHOLD, 0.05);
p.Set(KEY.MIN_PCT_OVERLAP_DUTY_CYCLES, 1);
p.Set(KEY.MIN_PCT_ACTIVE_DUTY_CYCLES, 0.001);
p.Set(KEY.DUTY_CYCLE_PERIOD, 100);
// These values activate powerfull boosting.
p.Set(KEY.MAX_BOOST, 5);
p.Set(KEY.DUTY_CYCLE_PERIOD, 100);
p.Set(KEY.MIN_PCT_OVERLAP_DUTY_CYCLES, 1);
p.Set(KEY.MAX_BOOST, 10);
p.Set(KEY.WRAP_AROUND, true);
p.Set(KEY.LEARN, true);
//-------------------TEMPORAL MEMORY PARAMETERS----------------
p.Set(KEY.COLUMN_DIMENSIONS, new int[] { columnsNumber });
p.Set(KEY.CELLS_PER_COLUMN, 32);
p.Set(KEY.ACTIVATION_THRESHOLD, 10);
p.Set(KEY.LEARNING_RADIUS, 10);
p.Set(KEY.MIN_THRESHOLD, 9);
p.Set(KEY.MAX_NEW_SYNAPSE_COUNT, 20);
p.Set(KEY.MAX_SYNAPSES_PER_SEGMENT, 225);
p.Set(KEY.MAX_SEGMENTS_PER_CELL, 225);
p.Set(KEY.INITIAL_PERMANENCE, 0.21);
p.Set(KEY.CONNECTED_PERMANENCE, 0.1);
p.Set(KEY.PERMANENCE_INCREMENT, 0.10);
p.Set(KEY.PERMANENCE_DECREMENT, 0.10);
p.Set(KEY.PREDICTED_SEGMENT_DECREMENT, 0.1);
p.Set(KEY.LEARN, true);
//Initiating components of a Cortex Layer
SpatialPoolerMT sp1 = new SpatialPoolerMT();
TemporalMemory tm1 = new TemporalMemory();
var mem = new Connections();
p.apply(mem);
sp1.Init(mem, UnitTestHelpers.GetMemory());
tm1.Init(mem);
HtmClassifier <double, ComputeCycle> cls = new HtmClassifier <double, ComputeCycle>();
Encoding(E_outBits);
// Can adjust the number of SP learning cycles below
for (int cycle = 0; cycle < 320; cycle++)
{
if (cycle >= 300)
{
// These activates ew-born effect which switch offs the boosting.
//mem.setMaxBoost(0.0);
mem.HtmConfig.MaxBoost = 0.0;
//mem.updateMinPctOverlapDutyCycles(0.0);
mem.HtmConfig.MinPctOverlapDutyCycles = 0.0;
}
using (StreamReader sr = new StreamReader(SP_inFile))
{
string line;
while ((line = sr.ReadLine()) != null)
{
string[] tokens = line.Split(",");
int[] SP_input = new int[E_outBits];
for (int i = 0; i < E_outBits; i++)
{
if (tokens[i + 1] == "0")
{
SP_input[i] = 0;
}
else
{
SP_input[i] = 1;
}
}
SP_Result = sp1.Compute(SP_input, true);
}
}
}
using (StreamReader sr = new StreamReader(SP_inFile))
{
string line;
int lineNumber = 0;
while ((line = sr.ReadLine()) != null)
{
string[] tokens = line.Split(",");
int[] SP_input = new int[E_outBits];
for (int i = 0; i < E_outBits; i++)
{
if (tokens[i + 1] == "0")
{
SP_input[i] = 0;
}
else
{
SP_input[i] = 1;
}
}
SP_Result = sp1.Compute(SP_input, false, false);
SP_Result_List.Add(SP_Result);
int[,] SP_twoDimenArray = ArrayUtils.Make2DArray(SP_Result, 32, 64);
var SP_twoDimArray = ArrayUtils.Transpose(SP_twoDimenArray);
NeoCortexUtils.DrawBitmap(SP_twoDimArray, 1024, 1024, $"{experimentFolder}\\{lineNumber}.png", Color.DimGray, Color.LawnGreen, text: tokens[0]);
lineNumber++;
}
}
using (StreamReader sr = new StreamReader(SP_noisyinFile))
{
string line;
int lineNumber = 0;
while ((line = sr.ReadLine()) != null)
{
string[] tokens = line.Split(",");
int[] SP_input = new int[E_outBits];
for (int i = 0; i < E_outBits; i++)
{
if (tokens[i + 1] == "0")
{
SP_input[i] = 0;
}
else
{
SP_input[i] = 1;
}
}
SP_NoisyResult = sp1.Compute(SP_input, false, false);
SP_NoisyResult_List.Add(SP_NoisyResult);
var ham = MathHelpers.GetHammingDistance(SP_Result_List[lineNumber], SP_NoisyResult_List[lineNumber], true);
Debug.WriteLine($"Noisy input: {tokens[0]} - Hamming NonZ: {ham}");
ham = MathHelpers.GetHammingDistance(SP_Result_List[lineNumber], SP_NoisyResult_List[lineNumber], false);
Debug.WriteLine($"Noisy input: {tokens[0]} - Hamming All: {ham}");
int[,] SP_twoDimenArray = ArrayUtils.Make2DArray(SP_NoisyResult, 32, 64);
var SP_twoDimArray = ArrayUtils.Transpose(SP_twoDimenArray);
NeoCortexUtils.DrawBitmap(SP_twoDimArray, 1024, 1024, $"{experimentFolder}\\{lineNumber}.png", Color.DimGray, Color.LawnGreen, text: tokens[0]);
lineNumber++;
}
}
for (int i = 0; i < number_testing_set - 1; i += 10)
{
int count = 1;
for (int j = i + 1; j < i + 1 + 9; j++)
{
if (i != 0 && i != number_testing_set - 1)
{
ham_upper = MathHelpers.GetHammingDistance(SP_NoisyResult_List[i], SP_NoisyResult_List[j], true);
ham_lower = MathHelpers.GetHammingDistance(SP_NoisyResult_List[i], SP_NoisyResult_List[i - count], true);
ham_total[count - 1] += ham_upper + ham_lower;
count++;
}
else if (i == 0)
{
ham_upper = MathHelpers.GetHammingDistance(SP_NoisyResult_List[i], SP_NoisyResult_List[j], true);
ham_total[count - 1] += ham_upper;
count++;
}
else
{
ham_lower = MathHelpers.GetHammingDistance(SP_NoisyResult_List[i], SP_NoisyResult_List[i - count], true);
ham_total[count - 1] += ham_lower;
count++;
}
}
}
for (int i = 0; i < 9; i++)
{
ham_avg[i] = ham_total[i] / (number_training_set * 2 - 2);
Debug.WriteLine($"0.{i + 1} step avg hamming distance: {ham_avg[i]}");
}
}
public void LongerSequenceExperiment()
{
int inputBits = 1024;
bool learn = true;
Parameters p = Parameters.getAllDefaultParameters();
p.Set(KEY.RANDOM, new ThreadSafeRandom(42));
p.Set(KEY.INPUT_DIMENSIONS, new int[] { inputBits });
p.Set(KEY.CELLS_PER_COLUMN, 10);
p.Set(KEY.COLUMN_DIMENSIONS, new int[] { 2048 });
CortexNetwork net = new CortexNetwork("my cortex");
List <CortexRegion> regions = new List <CortexRegion>();
CortexRegion region0 = new CortexRegion("1st Region");
regions.Add(region0);
SpatialPoolerMT sp1 = new SpatialPoolerMT();
TemporalMemory tm1 = new TemporalMemory();
var mem = new Connections();
p.apply(mem);
sp1.Init(mem, UnitTestHelpers.GetMemory());
tm1.Init(mem);
Dictionary <string, object> settings = new Dictionary <string, object>()
{
{ "W", 21 },
{ "N", inputBits },
{ "Radius", -1.0 },
{ "MinVal", 0.0 },
// { "MaxVal", 20.0 },
{ "Periodic", false },
{ "Name", "scalar" },
{ "ClipInput", false },
};
double max = 50;
List <double> lst = new List <double>();
for (double i = 0; i < max; i++)
{
lst.Add(i);
}
settings["MaxVal"] = max;
EncoderBase encoder = new ScalarEncoder(settings);
CortexLayer <object, object> layer1 = new CortexLayer <object, object>("L1");
//
// NewBorn learning stage.
region0.AddLayer(layer1);
layer1.HtmModules.Add("encoder", encoder);
layer1.HtmModules.Add("sp", sp1);
HtmClassifier <double, ComputeCycle> cls = new HtmClassifier <double, ComputeCycle>();
double[] inputs = lst.ToArray();
//
// This trains SP.
foreach (var input in inputs)
{
Debug.WriteLine($" ** {input} **");
for (int i = 0; i < 3; i++)
{
var lyrOut = layer1.Compute((object)input, learn) as ComputeCycle;
}
}
// Here we add TM module to the layer.
layer1.HtmModules.Add("tm", tm1);
//
// Now, training with SP+TM. SP is pretrained on pattern.
for (int i = 0; i < 200; i++)
{
foreach (var input in inputs)
{
var lyrOut = layer1.Compute(input, learn) as ComputeCycle;
cls.Learn(input, lyrOut.ActiveCells.ToArray());
Debug.WriteLine($"-------------- {input} ---------------");
if (learn == false)
{
Debug.WriteLine($"Inference mode");
}
Debug.WriteLine($"W: {Helpers.StringifyVector(lyrOut.WinnerCells.Select(c => c.Index).ToArray())}");
Debug.WriteLine($"P: {Helpers.StringifyVector(lyrOut.PredictiveCells.Select(c => c.Index).ToArray())}");
Debug.WriteLine($"Current Input: {input} \t| Predicted Input: {cls.GetPredictedInputValue(lyrOut.PredictiveCells.ToArray())}");
}
if (i == 50)
{
Debug.WriteLine("Stop Learning From Here. Entering inference mode.");
learn = false;
}
tm1.Reset(mem);
}
cls.TraceState();
Debug.WriteLine("------------------------------------------------------------------------\n----------------------------------------------------------------------------");
}