public void ErrorWhenLearningAfterInferenceMode()
{
var node = new SpatialNode2DGaussian();
var mat = new SparseMatrix(4, 4, 4.0);
var learnAfterInferFails = false;
var learnAfterTBInferFails = false;
node.Learn(mat);
node.Infer(mat);
try
{
node.Learn(mat);
}
catch (HtmRuleException e)
{
learnAfterInferFails = true;
Debug.WriteLine(e.Message);
}
node.TimeInfer(mat);
try
{
node.Learn(mat);
}
catch (HtmRuleException e)
{
learnAfterTBInferFails = true;
Debug.WriteLine(e.Message);
}
Assert.IsTrue(learnAfterInferFails);
Assert.IsTrue(learnAfterTBInferFails);
}
public override SpatialNode<SparseMatrix, Vector> GetClone()
{
var node = new SpatialNode2DGaussian();
node.SquaredSigma = this.SquaredSigma;
return node;
}
public void DontLearnBlankInputs()
{
var node = new SpatialNode2DGaussian();
node.Learn(new SparseMatrix(4));
Assert.AreEqual(0, node.CoincidencesFrequencies.Count);
}
public void ErrorWhenLearningVaryingSizeInputs()
{
var node = new SpatialNode2DGaussian();
node.Learn(new SparseMatrix(5, 5, 3.0));
try
{
node.Learn(new SparseMatrix(4, 4, 2.0));
Assert.Inconclusive("Should have fired an exception");
}
catch (Exception e)
{
Debug.WriteLine(e.Message);
Assert.IsInstanceOfType(e, typeof(HtmRuleException));
}
}
public void InferCorrectlyOnIdenticalMatricesWithMaxDistance0()
{
var node = new SpatialNode2DGaussian();
var ones = new SparseMatrix(4, 5, 1.0);
var twos = new SparseMatrix(4, 5, 2.0);
var expectedLower = Math.Exp(-(ones.RowCount * ones.ColumnCount) / (2 * node.SquaredSigma)) - 0.0001;
var expectedUpper = Math.Exp(-(ones.RowCount * ones.ColumnCount) / (2 * node.SquaredSigma)) + 0.0001;
node.Learn(ones);
node.Learn(twos);
var output = node.Infer(ones);
Assert.IsTrue(output[0] == 1.0 && output[1] > expectedLower && output[1] < expectedUpper);
}
public void TwoIdenticalInputsYieldOneCoincidenceOfFrequency2WithMaxDistance0()
{
// Arrange
double maxDistance = 0.0;
SpatialNode2DGaussian node = new SpatialNode2DGaussian(maxDistance, 0, 1);
int nbInputs = 2;
int width = 4;
int height = 4;
// Act
for (int i = 0; i < nbInputs; ++i)
node.Learn(new SparseMatrix(width, height, 1.0));
// Assert
Assert.AreEqual(1, node.CoincidencesFrequencies.Count);
Assert.AreEqual(nbInputs, node.CoincidencesFrequencies[new SparseMatrix(width, height, 1.0)]);
}
public void StopLearningWhenMaxOutputSizeReached()
{
int max = 1;
var node = new SpatialNode2DGaussian(0.0, 0.0, max);
var ones = SparseMatrix.Identity(5);
var twos = SparseMatrix.Identity(5) * 2.0;
node.Learn(ones);
node.Learn(twos);
Assert.AreEqual(max, node.CoincidencesFrequencies.Count);
}
public void Several5By5InputsYieldSeveralCoincidenceWithMaxDistance4_999()
{
var node = new SpatialNode2DGaussian(4.999);
var ones = new SparseMatrix(5, 5, 1.0);
var twos = new SparseMatrix(5, 5, 2.0);
node.Learn(ones);
node.Learn(twos);
Assert.AreNotEqual(1, node.CoincidencesFrequencies.Count);
Assert.AreEqual(2, node.CoincidencesFrequencies.Count);
}
public void Several1DistantInputsYieldOneCoincidenceWithMaxDistance1()
{
// Arrange
double maxDistance = 1.0;
SpatialNode2DGaussian node = new SpatialNode2DGaussian(maxDistance);
var matrices = new List<SparseMatrix>();
matrices.Add(new SparseMatrix(new double[,] { { 3.0, 2.0, 1.0, 0.0 } }));
matrices.Add(new SparseMatrix(new double[,] { { 3.0, 2.0, 0.0, 0.0 } }));
matrices.Add(new SparseMatrix(new double[,] { { 3.0, 2.0, 2.0, 0.0 } }));
matrices.Add(new SparseMatrix(new double[,] { { 3.0, 1.0, 1.0, 0.0 } }));
matrices.Add(new SparseMatrix(new double[,] { { 2.0, 2.0, 1.0, 0.0 } }));
int nbInputs = matrices.Count ;
// Act
//node.Learn();
foreach (var matrix in matrices)
node.Learn(matrix);
// Assert
Assert.AreEqual(1, node.CoincidencesFrequencies.Count);
Assert.AreNotEqual(nbInputs, node.CoincidencesFrequencies.Count);
}
public void RandomInputsYieldOnlyOneCoincidenceWithLargeMaxDistance()
{
// Arrange
double maxDistance = 10000.0;
SpatialNode2DGaussian node = new SpatialNode2DGaussian(maxDistance);
int nbInputs = 100;
int width = 4;
int height = 4;
var random = new Random(10);
// Act
for (int i = 0; i < nbInputs; ++i)
node.Learn(new SparseMatrix(width, height, random.NextDouble()));
// Assert
Assert.AreEqual(1, node.CoincidencesFrequencies.Count);
Assert.AreNotEqual(nbInputs, node.CoincidencesFrequencies.Count);
}
public void InferEmptyVectorIfNoCoincidencesLearned()
{
var node = new SpatialNode2DGaussian(1.0);
var output = node.Infer(new SparseMatrix(5, 5, 2.0));
Assert.AreEqual(new SparseVector(1), output);
output = node.TimeInfer(new SparseMatrix(5, 5, 2.0));
Assert.AreEqual(new SparseVector(1), output);
}
