@Override
public Object compute(FloatMatrix params, int flag) {
x = params.getRange(0, rows * features);
FloatMatrix theta = params.getRange(rows * features, params.length);
x = x.reshape(rows, features);
theta = theta.reshape(columns, features);
if (flag == 1 || flag == 3) {
FloatMatrix M = MatrixFunctions.pow(x.mmul(theta.transpose()).sub(y), 2);
this.cost = M.mul(r).columnSums().rowSums().get(0) / 2;
if (lambda != 0) {
float cost1 =
(lambda / 2)
* (MatrixFunctions.pow(theta, 2).columnSums().rowSums().get(0)
+ MatrixFunctions.pow(x, 2).columnSums().rowSums().get(0));
this.cost += cost1;
}
}
if (flag == 2 || flag == 3) {
FloatMatrix xGrad = FloatMatrix.zeros(x.rows, x.columns);
FloatMatrix thetaGrad = FloatMatrix.zeros(theta.rows, theta.columns);
int[] indices;
FloatMatrix thetaTemp;
FloatMatrix xTemp;
FloatMatrix yTemp;
for (int i = 0; i < rows; i++) {
indices = r.getRow(i).eq(1).findIndices();
if (indices.length == 0) continue;
thetaTemp = theta.getRows(indices);
yTemp = y.getRow(i).get(indices);
xGrad.putRow(i, x.getRow(i).mmul(thetaTemp.transpose()).sub(yTemp).mmul(thetaTemp));
}
xGrad = xGrad.add(x.mmul(lambda));
for (int i = 0; i < columns; i++) {
indices = r.getColumn(i).eq(1).findIndices();
if (indices.length == 0) continue;
xTemp = x.getRows(indices);
yTemp = y.getColumn(i).get(indices);
thetaGrad.putRow(
i, xTemp.mmul(theta.getRow(i).transpose()).sub(yTemp).transpose().mmul(xTemp));
}
thetaGrad = thetaGrad.add(theta.mmul(lambda));
this.gradient = MatrixUtil.merge(xGrad.data, thetaGrad.data);
}
return flag == 1 ? cost : gradient;
}
private FloatTensor getFloatTensorGradient(
FloatMatrix deltaFull, FloatMatrix leftVector, FloatMatrix rightVector) {
int size = deltaFull.length;
FloatTensor Wt_df = new FloatTensor(size * 2, size * 2, size);
FloatMatrix fullVector = FloatMatrix.concatHorizontally(leftVector, rightVector);
for (int slice = 0; slice < size; ++slice) {
Wt_df.setSlice(
slice, SimpleBlas.scal(deltaFull.get(slice), fullVector).mmul(fullVector.transpose()));
}
return Wt_df;
}
private FloatMatrix computeFloatTensorDeltaDown(
FloatMatrix deltaFull,
FloatMatrix leftVector,
FloatMatrix rightVector,
FloatMatrix W,
FloatTensor Wt) {
FloatMatrix WTDelta = W.transpose().mmul(deltaFull);
FloatMatrix WTDeltaNoBias = WTDelta.get(interval(0, 1), interval(0, deltaFull.rows * 2));
int size = deltaFull.length;
FloatMatrix deltaFloatTensor = new FloatMatrix(size * 2, 1);
FloatMatrix fullVector = FloatMatrix.concatHorizontally(leftVector, rightVector);
for (int slice = 0; slice < size; ++slice) {
FloatMatrix scaledFullVector = SimpleBlas.scal(deltaFull.get(slice), fullVector);
deltaFloatTensor =
deltaFloatTensor.add(
Wt.getSlice(slice).add(Wt.getSlice(slice).transpose()).mmul(scaledFullVector));
}
return deltaFloatTensor.add(WTDeltaNoBias);
}
private void backpropDerivativesAndError(
Tree tree,
MultiDimensionalMap<String, String, FloatMatrix> binaryTD,
MultiDimensionalMap<String, String, FloatMatrix> binaryCD,
MultiDimensionalMap<String, String, FloatTensor> binaryFloatTensorTD,
Map<String, FloatMatrix> unaryCD,
Map<String, FloatMatrix> wordVectorD,
FloatMatrix deltaUp) {
if (tree.isLeaf()) {
return;
}
FloatMatrix currentVector = tree.vector();
String category = tree.label();
category = basicCategory(category);
// Build a vector that looks like 0,0,1,0,0 with an indicator for the correct class
FloatMatrix goldLabel = new FloatMatrix(numOuts, 1);
int goldClass = tree.goldLabel();
if (goldClass >= 0) {
goldLabel.put(goldClass, 1.0f);
}
Float nodeWeight = classWeights.get(goldClass);
if (nodeWeight == null) nodeWeight = 1.0f;
FloatMatrix predictions = tree.prediction();
// If this is an unlabeled class, set deltaClass to 0. We could
// make this more efficient by eliminating various of the below
// calculations, but this would be the easiest way to handle the
// unlabeled class
FloatMatrix deltaClass =
goldClass >= 0
? SimpleBlas.scal(nodeWeight, predictions.sub(goldLabel))
: new FloatMatrix(predictions.rows, predictions.columns);
FloatMatrix localCD = deltaClass.mmul(appendBias(currentVector).transpose());
float error = -(MatrixFunctions.log(predictions).muli(goldLabel).sum());
error = error * nodeWeight;
tree.setError(error);
if (tree.isPreTerminal()) { // below us is a word vector
unaryCD.put(category, unaryCD.get(category).add(localCD));
String word = tree.children().get(0).label();
word = getVocabWord(word);
FloatMatrix currentVectorDerivative = activationFunction.apply(currentVector);
FloatMatrix deltaFromClass = getUnaryClassification(category).transpose().mmul(deltaClass);
deltaFromClass =
deltaFromClass.get(interval(0, numHidden), interval(0, 1)).mul(currentVectorDerivative);
FloatMatrix deltaFull = deltaFromClass.add(deltaUp);
wordVectorD.put(word, wordVectorD.get(word).add(deltaFull));
} else {
// Otherwise, this must be a binary node
String leftCategory = basicCategory(tree.children().get(0).label());
String rightCategory = basicCategory(tree.children().get(1).label());
if (combineClassification) {
unaryCD.put("", unaryCD.get("").add(localCD));
} else {
binaryCD.put(
leftCategory, rightCategory, binaryCD.get(leftCategory, rightCategory).add(localCD));
}
FloatMatrix currentVectorDerivative = activationFunction.applyDerivative(currentVector);
FloatMatrix deltaFromClass =
getBinaryClassification(leftCategory, rightCategory).transpose().mmul(deltaClass);
FloatMatrix mult = deltaFromClass.get(interval(0, numHidden), interval(0, 1));
deltaFromClass = mult.muli(currentVectorDerivative);
FloatMatrix deltaFull = deltaFromClass.add(deltaUp);
FloatMatrix leftVector = tree.children().get(0).vector();
FloatMatrix rightVector = tree.children().get(1).vector();
FloatMatrix childrenVector = appendBias(leftVector, rightVector);
// deltaFull 50 x 1, childrenVector: 50 x 2
FloatMatrix add = binaryTD.get(leftCategory, rightCategory);
FloatMatrix W_df = deltaFromClass.mmul(childrenVector.transpose());
binaryTD.put(leftCategory, rightCategory, add.add(W_df));
FloatMatrix deltaDown;
if (useFloatTensors) {
FloatTensor Wt_df = getFloatTensorGradient(deltaFull, leftVector, rightVector);
binaryFloatTensorTD.put(
leftCategory,
rightCategory,
binaryFloatTensorTD.get(leftCategory, rightCategory).add(Wt_df));
deltaDown =
computeFloatTensorDeltaDown(
deltaFull,
leftVector,
rightVector,
getBinaryTransform(leftCategory, rightCategory),
getBinaryFloatTensor(leftCategory, rightCategory));
} else {
deltaDown = getBinaryTransform(leftCategory, rightCategory).transpose().mmul(deltaFull);
}
FloatMatrix leftDerivative = activationFunction.apply(leftVector);
FloatMatrix rightDerivative = activationFunction.apply(rightVector);
FloatMatrix leftDeltaDown = deltaDown.get(interval(0, deltaFull.rows), interval(0, 1));
FloatMatrix rightDeltaDown =
deltaDown.get(interval(deltaFull.rows, deltaFull.rows * 2), interval(0, 1));
backpropDerivativesAndError(
tree.children().get(0),
binaryTD,
binaryCD,
binaryFloatTensorTD,
unaryCD,
wordVectorD,
leftDerivative.mul(leftDeltaDown));
backpropDerivativesAndError(
tree.children().get(1),
binaryTD,
binaryCD,
binaryFloatTensorTD,
unaryCD,
wordVectorD,
rightDerivative.mul(rightDeltaDown));
}
}
