public void TestKernelSummation()
{
double log_length = System.Math.Log(1.234);
double log_sig_sd = System.Math.Log(2.345);
double log_nse_sd = System.Math.Log(3.456);
SummationKernel kf = new SummationKernel(new SquaredExponential(log_length, log_sig_sd));
kf += new WhiteNoise(log_nse_sd);
Assert.Equal(log_length, kf[0]);
Assert.Equal(log_sig_sd, kf[1]);
Assert.Equal(log_nse_sd, kf[2]);
// Try resetting directly on the summation, using the name indexer
log_length = 4.321;
log_sig_sd = 5.432;
log_nse_sd = 6.543;
kf["Length"] = log_length;
kf["SignalSD"] = log_sig_sd;
kf["NoiseSD"] = log_nse_sd;
Assert.Equal(log_length, kf[0]);
Assert.Equal(log_sig_sd, kf[1]);
Assert.Equal(log_nse_sd, kf[2]);
}
public static void TestWithNoise(Vector[] inputs, double[] data)
{
var n = new Range(data.Length);
//var kf = new SummationKernel(new ARD(new double[]{ 0 }, 0))+new WhiteNoise();
var kf = new SummationKernel(new SquaredExponential()) + new WhiteNoise(-3);
var y = Variable <Vector> .Factor <double, Vector[], int[], KernelFunction>(MyFactors.GP, 1.0 /*Variable.GammaFromShapeAndRate(1,1)*/, inputs, new int[] { 0, 1 },
kf);
GPFactor.settings = new Settings
{
solverMethod = Settings.SolverMethod.GradientDescent,
};
var kf_noise = new SummationKernel(new SquaredExponential()) + new WhiteNoise(-3);
var noiseFunction = Variable <Vector> .Factor <double, Vector[], int[], KernelFunction>(MyFactors.GP, 1.0 /*Variable.GammaFromShapeAndRate(1,1)*/, inputs, new int[] { 0, 1 },
kf_noise);
GPFactor.settings = new Settings
{
solverMethod = Settings.SolverMethod.GradientDescent,
};
noiseFunction.AddAttribute(new MarginalPrototype(new VectorGaussian(n.SizeAsInt)));
var noiseFunctionValues = Variable.ArrayFromVector(noiseFunction, n);
var noisePrecisionValues = Variable.Array <double>(n);
//noisePrecisionValues[n] = Variable.Exp(noiseFunctionValues[n] + 2.0);
noisePrecisionValues[n] = Variable.Exp(noiseFunctionValues[n] + Variable.GaussianFromMeanAndPrecision(0, 1));
y.AddAttribute(new MarginalPrototype(new VectorGaussian(n.SizeAsInt)));
var y2noiseless = Variable.ArrayFromVector(y, n);
var y2 = Variable.Array <double>(n);
y2[n] = Variable.GaussianFromMeanAndPrecision(y2noiseless[n], noisePrecisionValues[n]);
y2.ObservedValue = data;
var ypredictiveNoiseless = Variable.ArrayFromVector(y, n);
var ypredictive = Variable.Array <double>(n);
ypredictive[n] = Variable.GaussianFromMeanAndPrecision(ypredictiveNoiseless[n], noisePrecisionValues[n]);
var ie = new InferenceEngine(new VariationalMessagePassing());
var post = ie.Infer <Gaussian[]>(ypredictive);
var mplWrapper = new MatplotlibWrapper();
mplWrapper.AddArray("x", inputs.Select(j => j[0]).ToArray());
mplWrapper.AddArray("y", data);
var f = post.Select(i => i.GetMean()).ToArray();
var e = post.Select(i => 2.0 * Math.Sqrt(i.GetVariance())).ToArray();
mplWrapper.AddArray("f", f);
mplWrapper.AddArray("e", e);
mplWrapper.Plot(new string[] {
"fill_between(x,f-e,f+e,color=\"gray\")",
"plot(x,f,'k')",
"scatter(x,y)"
});
}
public static (Vector[] dataX, double[] dataY) GenerateRandomData(int numData, double proportionCorrupt)
{
int randomSeed = 9876;
Random rng = new Random(randomSeed);
Rand.Restart(randomSeed);
InferenceEngine engine = Utilities.GetInferenceEngine();
// The points to evaluate
Vector[] randomInputs = Utilities.VectorRange(0, 1, numData, null);
var gaussianProcessGenerator = new GaussianProcessRegressor(randomInputs);
// The basis
Vector[] basis = Utilities.VectorRange(0, 1, 6, rng);
// The kernel
var kf = new SummationKernel(new SquaredExponential(-1)) + new WhiteNoise();
// Fill in the sparse GP prior
GaussianProcess gp = new GaussianProcess(new ConstantFunction(0), kf);
gaussianProcessGenerator.Prior.ObservedValue = new SparseGP(new SparseGPFixed(gp, basis));
// Infer the posterior Sparse GP, and sample a random function from it
SparseGP sgp = engine.Infer <SparseGP>(gaussianProcessGenerator.F);
var randomFunc = sgp.Sample();
double[] randomOutputs = new double[randomInputs.Length];
int numCorrupted = (int)Math.Ceiling(numData * proportionCorrupt);
var subset = Enumerable.Range(0, randomInputs.Length + 1).OrderBy(x => rng.Next()).Take(numCorrupted);
// get random data
for (int i = 0; i < randomInputs.Length; i++)
{
double post = randomFunc.Evaluate(randomInputs[i]);
// corrupt data point if it we haven't exceed the proportion we wish to corrupt
if (subset.Contains(i))
{
double sign = rng.NextDouble() > 0.5 ? 1 : -1;
double distance = rng.NextDouble() * 1;
post = (sign * distance) + post;
}
randomOutputs[i] = post;
}
Console.WriteLine("Model complete: Generated {0} points with {1} corrupted", numData, numCorrupted);
return(randomInputs, randomOutputs);
}
public void TestSummmationReadWrite()
{
double log_length = System.Math.Log(1.234);
double log_sig_sd = System.Math.Log(2.345);
double log_nse_sd = System.Math.Log(3.456);
double[] log_var = { System.Math.Log(0.987), System.Math.Log(0.876), System.Math.Log(0.765) };
SummationKernel kf = new SummationKernel(new SquaredExponential(log_length, log_sig_sd));
kf += new LinearKernel(log_var);
kf += new WhiteNoise(log_nse_sd);
TestReadWrite(kf);
}
static void FitDataset(bool useSynthetic)
{
Vector[] trainingInputs;
double[] trainingOutputs;
if (!useSynthetic)
{
var trainingData = Utilities.LoadAISDataset();
trainingInputs = trainingData.Select(tup => Vector.FromArray(new double[1] {
tup.x
})).ToArray();
trainingOutputs = trainingData.Select(tup => tup.y).ToArray();
}
else
{
(trainingInputs, trainingOutputs) = GaussianProcessDataGenerator.GenerateRandomData(30, 0.3);
}
InferenceEngine engine = Utilities.GetInferenceEngine();
// First fit standard GP, then fit Student-T GP
foreach (var useStudentTLikelihood in new[] { false, true })
{
var gaussianProcessRegressor = new GaussianProcessRegressor(trainingInputs, useStudentTLikelihood, trainingOutputs);
// Log length scale estimated as -1
var noiseVariance = 0.8;
var kf = new SummationKernel(new SquaredExponential(-1)) + new WhiteNoise(Math.Log(noiseVariance) / 2);
GaussianProcess gp = new GaussianProcess(new ConstantFunction(0), kf);
// Convert SparseGP to full Gaussian Process by evaluating at all the training points
gaussianProcessRegressor.Prior.ObservedValue = new SparseGP(new SparseGPFixed(gp, trainingInputs.ToArray()));
double logOdds = engine.Infer <Bernoulli>(gaussianProcessRegressor.Evidence).LogOdds;
Console.WriteLine("{0} evidence = {1}", kf, logOdds.ToString("g4"));
// Infer the posterior Sparse GP
SparseGP sgp = engine.Infer <SparseGP>(gaussianProcessRegressor.F);
#if NETFULL
string datasetName = useSynthetic ? "Synthetic" : "AIS";
Utilities.PlotPredictions(sgp, trainingInputs, trainingOutputs, useStudentTLikelihood, datasetName);
#endif
}
}
public void TestSumKDerivs()
{
double[] log_length = { System.Math.Log(0.543), System.Math.Log(0.432), System.Math.Log(0.321) };
double log_sig_sd = System.Math.Log(2.345);
double[] log_var = { System.Math.Log(0.987), System.Math.Log(0.876), System.Math.Log(0.765) };
double log_nse_sd = System.Math.Log(3.456);
SummationKernel kf = new SummationKernel(new ARD(log_length, log_sig_sd));
kf += new LinearKernel(log_var);
kf += new WhiteNoise(log_nse_sd);
double[] x1 = { 0.1, 0.2, 0.3 };
double[] x2 = { 0.9, 0.7, 0.5 };
Vector x1Vec = Vector.FromArray(x1);
Vector x2Vec = Vector.FromArray(x2);
TestDerivatives(kf, x1Vec, x1Vec);
TestDerivatives(kf, x1Vec, x2Vec);
}
public void BasicGPC()
{
Vector[] inputs = new Vector[]
{
Vector.FromArray(new double[2] {
0, 0
}),
Vector.FromArray(new double[2] {
0, 1
}),
Vector.FromArray(new double[2] {
1, 0
}),
Vector.FromArray(new double[2] {
0, 0.5
}),
Vector.FromArray(new double[2] {
1.5, 0
}),
Vector.FromArray(new double[2] {
0.5, 1.0
})
};
bool[] outputs = { true, true, false, true, false, false };
var kf = new SummationKernel(new SquaredExponential(0));
kf += new WhiteNoise(System.Math.Log(0.1));
var K = GramMatrix(kf, inputs);
var n = new Range(inputs.Length);
var x = Variable.VectorGaussianFromMeanAndVariance(Vector.Zero(inputs.Length), K);
var g = Variable.ArrayFromVector(x, n);
var p = Variable.Array <bool>(n);
p[n] = Variable.IsPositive(g[n]);
p.ObservedValue = outputs;
var ie = new InferenceEngine();
Console.WriteLine(ie.Infer(x));
}
public static void Test()
{
var inputs = Enumerable.Range(0, 50).Select(i => Vector.Constant(1, i)).ToArray();
var data = inputs.Select(j => Math.Cos(2 * j[0] / 10.0)).ToArray();
var n = new Range(data.Length);
//var kf = new SummationKernel(new ARD(new double[]{ 0 }, 0))+new WhiteNoise();
var kf = new SummationKernel(new SquaredExponential()) + new WhiteNoise();
var y = Variable <Vector> .Factor <double, Vector[], int[], KernelFunction>(MyFactors.GP, 1.0 /*Variable.GammaFromShapeAndRate(1,1)*/, inputs, new int[] { 0, 1 },
kf);
GPFactor.settings = new Settings
{
solverMethod = Settings.SolverMethod.GradientDescent,
};
y.AddAttribute(new MarginalPrototype(new VectorGaussian(n.SizeAsInt)));
var y2 = Variable.ArrayFromVector(y, n);
y2.ObservedValue = data;
var ypredictive = Variable.ArrayFromVector(y, n);
var ie = new InferenceEngine(new VariationalMessagePassing());
var post = ie.Infer <Gaussian[]>(ypredictive);
var mplWrapper = new MatplotlibWrapper();
mplWrapper.AddArray("x", inputs.Select(j => j[0]).ToArray());
mplWrapper.AddArray("y", data);
var f = post.Select(i => i.GetMean()).ToArray();
var e = post.Select(i => Math.Sqrt(i.GetVariance())).ToArray();
mplWrapper.AddArray("f", f);
mplWrapper.AddArray("e", e);
mplWrapper.Plot(new string[] {
"fill_between(x,f-e,f+e,color=\"gray\")",
"scatter(x,y)"
});
}
/// <summary>
/// Primary definition of the GPRN model as an Infer.NET model.
/// </summary>
/// <param name="inputs">Covariates X</param>
/// <param name="data">Outputs Y</param>
/// <param name="Q">Number of latent functions</param>
/// <param name="missing">Which elements of Y are missing</param>
/// <param name="nodeFunctionNoise">Whether to include node noise</param>
/// <param name="constrainWpositive">Whether to constrain W to be positive [experimental]</param>
/// <param name="isotropicNoise">Whether to use isotropic observation noise</param>
/// <param name="meanFunctions">Whether to include a per output mean function</param>
/// <param name="initLoglengthscales">Initial values for the length scales of the kernels</param>
/// <param name="sw">An output file for logging</param>
public void GPRN_InferNET_model(Vector[] inputs,
double[,] data,
int Q,
bool grid = false,
bool[,] missing = null,
bool nodeFunctionNoise = false,
bool constrainWpositive = false,
bool isotropicNoise = true,
bool meanFunctions = false,
double[] initLoglengthscales = null,
StreamWriter sw = null)
{
var toInfer = new List <IVariable>();
SummationKernel kf_node = new SummationKernel(new SquaredExponential(0)) + new WhiteNoise(-3);
var K_node = Utils.GramMatrix(kf_node, inputs);
SummationKernel kf_weights = new SummationKernel(new SquaredExponential(1)) + new WhiteNoise(-3);
var K_weights = Utils.GramMatrix(kf_weights, inputs);
var D = Variable.Observed <int>(data.GetLength(0)).Named("D");
var d = new Range(D).Named("d");
var Qvar = Variable.Observed <int>(Q).Named("Q");
var q = new Range(Qvar).Named("q");
var N = Variable.Observed <int>(data.GetLength(1)).Named("N");
var n = new Range(N).Named("n");
if (missing == null)
{
missing = new bool[D.ObservedValue, N.ObservedValue]; // check this is all false
}
var ev = Variable.Bernoulli(.5).Named("ev");
var modelBlock = Variable.If(ev);
var nodeSignalPrecisions = Variable.Array <double>(q).Named("nodeSignalPrecisions");
// set this to 1 if not learning signal variance
var nodeSignalPrecisionsPrior = Variable.Observed(Enumerable.Range(0, Q).Select(_ => Gamma.FromShapeAndRate(.1, .1)).ToArray(), q).Named("nodeSignalPrecisionsPrior");
nodeSignalPrecisions[q] = Variable.Random <double, Gamma>(nodeSignalPrecisionsPrior[q]);
var nodeFunctions = Variable.Array <Vector>(q).Named("nodeFunctions");
var K_node_inverse = Variable.Observed(K_node.Inverse()).Named("K_node_inverse");
nodeFunctions[q] = Variable <Vector> .Factor(MyFactors.VectorGaussianScaled, nodeSignalPrecisions[q], K_node_inverse);
nodeFunctions.AddAttribute(new MarginalPrototype(new VectorGaussian(N.ObservedValue)));
var nodeFunctionValues = Variable.Array(Variable.Array <double>(n), q).Named("nodeFunctionValues");
var nodeFunctionValuesPredictive = Variable.Array(Variable.Array <double>(n), q).Named("nodeFunctionValuesPredictive");
VariableArray <double> nodeNoisePrecisions = null;
if (nodeFunctionNoise)
{
var nodeFunctionValuesClean = Variable.Array(Variable.Array <double>(n), q).Named("nodeFunctionValuesClean");
nodeFunctionValuesClean[q] = Variable.ArrayFromVector(nodeFunctions[q], n);
nodeNoisePrecisions = Variable.Array <double>(q).Named("nodeNoisePrecisions");
var nodeNoisePrecisionPrior = Variable.Observed(Enumerable.Range(0, Q).Select(_ => Gamma.FromShapeAndRate(.1, .01)).ToArray(), q).Named("nodeNoisePrecisionPrior");
nodeNoisePrecisions[q] = Variable.Random <double, Gamma>(nodeNoisePrecisionPrior[q]);
toInfer.Add(nodeNoisePrecisions);
nodeFunctionValues[q][n] = Variable.GaussianFromMeanAndPrecision(nodeFunctionValuesClean[q][n], nodeNoisePrecisions[q]);
nodeFunctionValuesPredictive[q][n] = Variable.GaussianFromMeanAndPrecision(nodeFunctionValuesClean[q][n], nodeNoisePrecisions[q]);
}
else
{
nodeFunctionValues[q] = Variable.ArrayFromVector(nodeFunctions[q], n);
nodeFunctionValuesPredictive[q] = Variable.ArrayFromVector(nodeFunctions[q], n);
}
var weightFunctions = Variable.Array <Vector>(d, q).Named("weightFunctions");
var K_weights_inverse = Variable.Observed(K_weights.Inverse()).Named("K_weights_inverse");
weightFunctions[d, q] = Variable <Vector> .Factor(MyFactors.VectorGaussianScaled, Variable.Constant <double>(1), K_weights_inverse).ForEach(d, q);
weightFunctions.AddAttribute(new MarginalPrototype(new VectorGaussian(N.ObservedValue)));
var weightFunctionValues = Variable.Array(Variable.Array <double>(n), d, q).Named("weightFunctionValues");
var weightFunctionValues2 = Variable.Array(Variable.Array <double>(n), d, q).Named("weightFunctionValuesPredictive");
weightFunctionValues[d, q] = Variable.ArrayFromVector(weightFunctions[d, q], n);
if (constrainWpositive)
{
var weightFunctionValuesCopy = Variable.Array(Variable.Array <double>(n), d, q).Named("weightFunctionValuesCopy");
weightFunctionValuesCopy[d, q][n] = Variable.GaussianFromMeanAndPrecision(weightFunctionValues[d, q][n], 100);
Variable.ConstrainPositive(weightFunctionValuesCopy[d, q][n]);
}
weightFunctionValues2[d, q] = Variable.ArrayFromVector(weightFunctions[d, q], n);
var observedData = Variable.Array <double>(d, n).Named("observedData");
var noisePrecisionPrior = Variable.Observed(Gamma.FromShapeAndRate(1, .1)).Named("noisePrecisionPrior");
Variable <double> noisePrecision = null;
VariableArray <double> noisePrecisionArray = null;
if (isotropicNoise)
{
noisePrecision = Variable.Random <double, Gamma>(noisePrecisionPrior).Named("noisePrecision");
toInfer.Add(noisePrecision);
}
else
{
noisePrecisionArray = Variable.Array <double>(d).Named("noisePrecision");
noisePrecisionArray[d] = Variable.Random <double, Gamma>(noisePrecisionPrior).ForEach(d);
toInfer.Add(noisePrecisionArray);
}
var isMissing = Variable.Array <bool>(d, n).Named("isMissing");
isMissing.ObservedValue = missing;
var noiseLessY = Variable.Array <double>(d, n).Named("noiseLessY");
VariableArray <VariableArray <double>, double[][]> meanFunctionValues = null;
if (meanFunctions)
{
GPFactor.settings = new Settings
{
solverMethod = Settings.SolverMethod.GradientDescent,
};
VariableArray <KernelFunction> kf = Variable.Array <KernelFunction>(d);
kf.ObservedValue = Enumerable.Range(0, D.ObservedValue).Select(
o => new SummationKernel(new SquaredExponential()) + new WhiteNoise(-3)).ToArray();
var mf = Variable.Array <Vector>(d).Named("meanFunctions");
mf[d] = Variable <Vector> .Factor <double, Vector[], int[], KernelFunction>(MyFactors.GP, 1.0 /*Variable.GammaFromShapeAndRate(1,1)*/, inputs, new int[] { 0 },
kf[d]);
mf.AddAttribute(new MarginalPrototype(new VectorGaussian(N.ObservedValue)));
meanFunctionValues = Variable.Array(Variable.Array <double>(n), d).Named("meanFunctionValues");
meanFunctionValues[d] = Variable.ArrayFromVector(mf[d], n);
toInfer.Add(meanFunctionValues);
}
using (Variable.ForEach(n))
using (Variable.ForEach(d))
{
var temp = Variable.Array <double>(q).Named("temp");
temp[q] = weightFunctionValues[d, q][n] * nodeFunctionValues[q][n];
if (meanFunctions)
{
noiseLessY[d, n] = Variable.Sum(temp) + meanFunctionValues[d][n];
}
else
{
noiseLessY[d, n] = Variable.Sum(temp);
}
using (Variable.IfNot(isMissing[d, n]))
if (isotropicNoise)
{
observedData[d, n] = Variable.GaussianFromMeanAndPrecision(noiseLessY[d, n], noisePrecision);
}
else
{
observedData[d, n] = Variable.GaussianFromMeanAndPrecision(noiseLessY[d, n], noisePrecisionArray[d]);
}
using (Variable.If(isMissing[d, n]))
observedData[d, n] = Variable.GaussianFromMeanAndPrecision(0, 1);
}
observedData.ObservedValue = data;
var nodeFunctionsInit = Enumerable.Range(0, Q).Select(i =>
VectorGaussian.FromMeanAndVariance(
VectorGaussian.Sample(Vector.Zero(N.ObservedValue), PositiveDefiniteMatrix.IdentityScaledBy(N.ObservedValue, 100)),
PositiveDefiniteMatrix.IdentityScaledBy(N.ObservedValue, 100))).ToArray(); // should put this manually in generated code
var distArray = Distribution <Vector> .Array(nodeFunctionsInit);
var nodeFunctionsInitVar = Variable.Observed(distArray).Named("nodeFunctionsInitVar");
nodeFunctions.InitialiseTo(nodeFunctionsInitVar);
modelBlock.CloseBlock();
toInfer.AddRange(new List <IVariable>()
{
ev, noiseLessY, nodeFunctionValues, nodeSignalPrecisions, nodeFunctionValuesPredictive, weightFunctionValues, weightFunctionValues2
});
var infer = new InferenceEngine(new VariationalMessagePassing());
infer.ModelName = "MeanFunction";
var ca = infer.GetCompiledInferenceAlgorithm(toInfer.ToArray());
var kernel = new SummationKernel(new SquaredExponential(initLoglengthscales[0]));
kernel += new WhiteNoise(-3);
ca.SetObservedValue(K_node_inverse.NameInGeneratedCode, Utils.GramMatrix(kernel, inputs).Inverse());
kernel = new SummationKernel(new SquaredExponential(initLoglengthscales[1]));
kernel += new WhiteNoise(-3);
ca.SetObservedValue(K_weights_inverse.NameInGeneratedCode, Utils.GramMatrix(kernel, inputs).Inverse());
ca.Reset();
double oldML = double.NegativeInfinity;
double ml = 0;
int it = 0;
for (; it < 100; it++)
{
ca.Update(1);
ml = ca.Marginal <Bernoulli>(ev.NameInGeneratedCode).LogOdds;
Console.WriteLine(ml);
if (Math.Abs(oldML - ml) < .1)
{
break;
}
oldML = ml;
}
Console.WriteLine("Finished after " + it);
}
/// <summary>
/// Infer.NET definition of the Semi Parametric Latent Factor Model of
/// Teh, Y., Seeger, M., and Jordan, M. (AISTATS 2005).
/// </summary>
/// <param name="inputs">Covariates X</param>
/// <param name="data">Outputs Y</param>
/// <param name="Q">Number of latent functions</param>
/// <param name="missing">Which elements of Y are missing</param>
/// <param name="nodeFunctionNoise">Whether to include node noise</param>
public void SPLFM(
Vector[] inputs,
double[,] data,
int Q,
bool[,] missing = null,
bool nodeFunctionNoise = false)
{
var toInfer = new List <IVariable>();
SummationKernel kf_node = new SummationKernel(new SquaredExponential(0));
var K_node = Utils.GramMatrix(kf_node, inputs);
var D = Variable.Observed <int>(data.GetLength(0)).Named("D");
var d = new Range(D).Named("d");
var Qvar = Variable.Observed <int>(Q).Named("Q");
var q = new Range(Qvar).Named("q");
var N = Variable.Observed <int>(data.GetLength(1)).Named("N");
var n = new Range(N).Named("n");
if (missing == null)
{
missing = new bool[D.ObservedValue, N.ObservedValue]; // check this is all false
}
var ev = Variable.Bernoulli(.5).Named("ev");
var modelBlock = Variable.If(ev);
var nodeSignalPrecisions = Variable.Array <double>(q).Named("nodeSignalPrecisions");
// set this to 1 if not learning signal variance
var nodeSignalPrecisionsPrior = Variable.Observed(Enumerable.Range(0, Q).Select(_ => Gamma.FromShapeAndRate(.1, .1)).ToArray(), q).Named("nodeSignalPrecisionsPrior");
nodeSignalPrecisions[q] = Variable.Random <double, Gamma>(nodeSignalPrecisionsPrior[q]);
var nodeFunctions = Variable.Array <Vector>(q).Named("nodeFunctions");
var K_node_inverse = Variable.Observed(K_node.Inverse()).Named("K_node_inverse");
nodeFunctions[q] = Variable <Vector> .Factor(MyFactors.VectorGaussianScaled, nodeSignalPrecisions[q], K_node_inverse);
nodeFunctions.AddAttribute(new MarginalPrototype(new VectorGaussian(N.ObservedValue)));
var nodeFunctionValues = Variable.Array(Variable.Array <double>(n), q).Named("nodeFunctionValues");
var nodeFunctionValuesPredictive = Variable.Array(Variable.Array <double>(n), q).Named("nodeFunctionValuesPredictive");
VariableArray <double> nodeNoisePrecisions = null;
if (nodeFunctionNoise)
{
var nodeFunctionValuesClean = Variable.Array(Variable.Array <double>(n), q).Named("nodeFunctionValuesClean");
nodeFunctionValuesClean[q] = Variable.ArrayFromVector(nodeFunctions[q], n);
nodeNoisePrecisions = Variable.Array <double>(q).Named("nodeNoisePrecisions");
var nodeNoisePrecisionPrior = Variable.Observed(Enumerable.Range(0, Q).Select(_ => Gamma.FromShapeAndRate(.1, .01)).ToArray(), q).Named("nodeNoisePrecisionPrior");
nodeNoisePrecisions[q] = Variable.Random <double, Gamma>(nodeNoisePrecisionPrior[q]);
toInfer.Add(nodeNoisePrecisions);
nodeFunctionValues[q][n] = Variable.GaussianFromMeanAndPrecision(nodeFunctionValuesClean[q][n], nodeNoisePrecisions[q]);
nodeFunctionValuesPredictive[q][n] = Variable.GaussianFromMeanAndPrecision(nodeFunctionValuesClean[q][n], nodeNoisePrecisions[q]);
}
else
{
nodeFunctionValues[q] = Variable.ArrayFromVector(nodeFunctions[q], n);
nodeFunctionValuesPredictive[q] = Variable.ArrayFromVector(nodeFunctions[q], n);
}
var weights = Variable.Array <double>(d, q).Named("weights");
weights[d, q] = Variable.GaussianFromMeanAndPrecision(0, 1).ForEach(d, q);
var observedData = Variable.Array <double>(d, n).Named("observedData");
var noisePrecisionPrior = Variable.Observed(Gamma.FromShapeAndRate(1, .1)).Named("noisePrecisionPrior");
var noisePrecision = Variable.Random <double, Gamma>(noisePrecisionPrior).Named("noisePrecision");
var isMissing = Variable.Array <bool>(d, n).Named("isMissing");
isMissing.ObservedValue = missing;
var noiseLessY = Variable.Array <double>(d, n).Named("noiseLessY");
using (Variable.ForEach(n))
using (Variable.ForEach(d))
{
var temp = Variable.Array <double>(q).Named("temp");
temp[q] = weights[d, q] * nodeFunctionValues[q][n];
noiseLessY[d, n] = Variable.Sum(temp);
using (Variable.IfNot(isMissing[d, n]))
observedData[d, n] = Variable.GaussianFromMeanAndPrecision(noiseLessY[d, n], noisePrecision);
using (Variable.If(isMissing[d, n]))
observedData[d, n] = Variable.GaussianFromMeanAndPrecision(0, 1);
}
observedData.ObservedValue = data;
var nodeFunctionsInit = Enumerable.Range(0, Q).Select(i =>
VectorGaussian.FromMeanAndVariance(
VectorGaussian.Sample(Vector.Zero(N.ObservedValue), PositiveDefiniteMatrix.IdentityScaledBy(N.ObservedValue, 100)),
PositiveDefiniteMatrix.IdentityScaledBy(N.ObservedValue, 100))).ToArray(); // should put this manually in generated code
var distArray = Distribution <Vector> .Array(nodeFunctionsInit);
var nodeFunctionsInitVar = Variable.Observed(distArray).Named("nodeFunctionsInitVar");
nodeFunctions.InitialiseTo(nodeFunctionsInitVar);
modelBlock.CloseBlock();
toInfer.AddRange(new List <IVariable>()
{
ev, noiseLessY, noisePrecision, nodeFunctionValues, nodeSignalPrecisions, nodeFunctionValuesPredictive, weights
});
var ie = new InferenceEngine(new VariationalMessagePassing());
ie.ModelName = "SPLFM";
var ca = ie.GetCompiledInferenceAlgorithm(toInfer.ToArray());
ca.Execute(100);
var fvals = ca.Marginal <Gaussian[][]>(nodeFunctionValues.NameInGeneratedCode)[0]; // [q][n]
var x = inputs.Select(i => i[0]).ToArray();
var mplWrapper = new MatplotlibWrapper();
mplWrapper.AddArray("x", x);
mplWrapper.AddArray("y", fvals.Select(i => i.GetMean()).ToArray());
mplWrapper.AddArray("s", fvals.Select(i => Math.Sqrt(i.GetVariance())).ToArray());
mplWrapper.Plot(new string[] {
"fill_between(x,y-s,y+s,color=\"gray\")",
"ylabel(\"node (fitted)\")"
});
}
