Author: tommaso
Date: Thu Jan 14 16:19:32 2016
New Revision: 1724646
URL: http://svn.apache.org/viewvc?rev=1724646&view=rev
Log:
pruned old api, minimal shallow ff nn impl
Added:
labs/yay/trunk/core/src/main/java/org/apache/yay/ActivationFunction.java
(with props)
labs/yay/trunk/core/src/main/java/org/apache/yay/IdentityActivationFunction.java
(with props)
labs/yay/trunk/core/src/main/java/org/apache/yay/ShallowFeedForwardNeuralNetwork.java
(with props)
labs/yay/trunk/core/src/main/java/org/apache/yay/SigmoidFunction.java
- copied, changed from r1714192,
labs/yay/trunk/core/src/main/java/org/apache/yay/core/SigmoidFunction.java
labs/yay/trunk/core/src/main/java/org/apache/yay/SoftmaxActivationFunction.java
- copied, changed from r1715735,
labs/yay/trunk/core/src/main/java/org/apache/yay/core/SoftmaxActivationFunction.java
labs/yay/trunk/core/src/main/java/org/apache/yay/StepActivationFunction.java
- copied, changed from r1714192,
labs/yay/trunk/core/src/main/java/org/apache/yay/core/StepActivationFunction.java
labs/yay/trunk/core/src/main/java/org/apache/yay/TanhFunction.java
- copied, changed from r1714192,
labs/yay/trunk/core/src/main/java/org/apache/yay/core/TanhFunction.java
labs/yay/trunk/core/src/test/java/org/apache/yay/ShallowFeedForwardNeuralNetworkTest.java
(with props)
labs/yay/trunk/core/src/test/java/org/apache/yay/SigmoidFunctionTest.java
- copied, changed from r1707760,
labs/yay/trunk/core/src/test/java/org/apache/yay/core/SigmoidFunctionTest.java
Removed:
labs/yay/trunk/api/
labs/yay/trunk/core/src/main/java/org/apache/yay/core/
labs/yay/trunk/core/src/test/java/org/apache/yay/core/
Modified:
labs/yay/trunk/core/pom.xml
labs/yay/trunk/pom.xml
Modified: labs/yay/trunk/core/pom.xml
URL:
http://svn.apache.org/viewvc/labs/yay/trunk/core/pom.xml?rev=1724646&r1=1724645&r2=1724646&view=diff
==============================================================================
--- labs/yay/trunk/core/pom.xml (original)
+++ labs/yay/trunk/core/pom.xml Thu Jan 14 16:19:32 2016
@@ -22,17 +22,12 @@
<parent>
<groupId>org.apache.yay</groupId>
<artifactId>yay-parent</artifactId>
- <version>0.1-SNAPSHOT</version>
+ <version>0.2-SNAPSHOT</version>
<relativePath>../</relativePath>
</parent>
<name>Yay core</name>
<dependencies>
<dependency>
- <groupId>org.apache.yay</groupId>
- <artifactId>yay-api</artifactId>
- <version>${project.version}</version>
- </dependency>
- <dependency>
<groupId>org.mockito</groupId>
<artifactId>mockito-core</artifactId>
<version>1.9.5</version>
Added: labs/yay/trunk/core/src/main/java/org/apache/yay/ActivationFunction.java
URL:
http://svn.apache.org/viewvc/labs/yay/trunk/core/src/main/java/org/apache/yay/ActivationFunction.java?rev=1724646&view=auto
==============================================================================
--- labs/yay/trunk/core/src/main/java/org/apache/yay/ActivationFunction.java
(added)
+++ labs/yay/trunk/core/src/main/java/org/apache/yay/ActivationFunction.java
Thu Jan 14 16:19:32 2016
@@ -0,0 +1,41 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied. See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+package org.apache.yay;
+
+import org.apache.commons.math3.linear.RealMatrix;
+
+/**
+ * An activation function AF : S -* S receives a signal and generates a new
signal.
+ * An activation function AF has horizontal asymptotes at 0 and 1 and a non
+ * decreasing first derivative AF' with AF and AF' both being computable.
+ * These are usually used in neurons in order to propagate the "signal"
+ * throughout the whole network.
+ */
+public interface ActivationFunction {
+
+ /**
+ * Apply this <code>ActivationFunction</code> to the given matrix of
signals, generating a new matrix of transformed
+ * signals.
+ *
+ * @param weights the matrix of weights the activation should be applied to
+ * @return the output signal generated as a {@link RealMatrix}
+ */
+ RealMatrix applyMatrix(RealMatrix weights);
+
+}
Propchange:
labs/yay/trunk/core/src/main/java/org/apache/yay/ActivationFunction.java
------------------------------------------------------------------------------
svn:eol-style = native
Added:
labs/yay/trunk/core/src/main/java/org/apache/yay/IdentityActivationFunction.java
URL:
http://svn.apache.org/viewvc/labs/yay/trunk/core/src/main/java/org/apache/yay/IdentityActivationFunction.java?rev=1724646&view=auto
==============================================================================
---
labs/yay/trunk/core/src/main/java/org/apache/yay/IdentityActivationFunction.java
(added)
+++
labs/yay/trunk/core/src/main/java/org/apache/yay/IdentityActivationFunction.java
Thu Jan 14 16:19:32 2016
@@ -0,0 +1,34 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied. See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+package org.apache.yay;
+
+import org.apache.commons.math3.linear.RealMatrix;
+import org.apache.yay.ActivationFunction;
+
+/**
+ * An {@link org.apache.yay.ActivationFunction} which just replicates the
signal without changing it
+ */
+public class IdentityActivationFunction implements ActivationFunction {
+
+ @Override
+ public RealMatrix applyMatrix(RealMatrix weights) {
+ return weights;
+ }
+
+}
Propchange:
labs/yay/trunk/core/src/main/java/org/apache/yay/IdentityActivationFunction.java
------------------------------------------------------------------------------
svn:eol-style = native
Added:
labs/yay/trunk/core/src/main/java/org/apache/yay/ShallowFeedForwardNeuralNetwork.java
URL:
http://svn.apache.org/viewvc/labs/yay/trunk/core/src/main/java/org/apache/yay/ShallowFeedForwardNeuralNetwork.java?rev=1724646&view=auto
==============================================================================
---
labs/yay/trunk/core/src/main/java/org/apache/yay/ShallowFeedForwardNeuralNetwork.java
(added)
+++
labs/yay/trunk/core/src/main/java/org/apache/yay/ShallowFeedForwardNeuralNetwork.java
Thu Jan 14 16:19:32 2016
@@ -0,0 +1,357 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied. See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+package org.apache.yay;
+
+import org.apache.commons.lang3.SystemUtils;
+import org.apache.commons.math3.linear.ArrayRealVector;
+import org.apache.commons.math3.linear.MatrixUtils;
+import org.apache.commons.math3.linear.RealMatrix;
+import org.apache.commons.math3.linear.RealMatrixChangingVisitor;
+import org.apache.commons.math3.linear.RealVector;
+
+import java.util.Arrays;
+import java.util.Random;
+
+/**
+ * A shallow feed forward neural network.
+ * It learns its weights through backpropagation algorithm via stochastic
gradient descent applied to a collection of
+ * training samples.
+ * Each example is a real vectors whose first N elements (identified by the
no. of network input units) are the actual
+ * outputs and the remaining elements are the input features.
+ */
+public class ShallowFeedForwardNeuralNetwork {
+
+ private final Configuration configuration;
+
+ /**
+ * Each RealMatrix maps weights between two layers.
+ * E.g.: weights[0] controls function mapping from layer 0 to layer 1.
+ * If network has 4 units in layer 1 and 5 units in layer 2, then weights[0]
will be of dimension 5x4, plus bias terms.
+ * A network having layers with 3, 4 and 2 units each will have the
following weights matrix dimensions:
+ * - weights[0] : 4x3
+ * - weights[1] : 2x4
+ * <p>
+ * the first row of weighs[0]Â matrix holds the weights of each neuron in
the first neuron of the second layer,
+ * the second row of weighs[0]Â holds the weights of each neuron in the
second neuron of the second layer, etc.
+ */
+ private RealMatrix[] weights;
+
+ public ShallowFeedForwardNeuralNetwork(Configuration configuration) {
+ this.configuration = configuration;
+ initialize();
+ }
+
+ private void initialize() {
+ weights = createRandomWeights();
+ }
+
+ private RealMatrix[] createRandomWeights() {
+ Random r = new Random();
+ int[] layers = new int[configuration.layers.length];
+ for (int i = 0; i < layers.length; i++) {
+ layers[i] = configuration.layers[i] + (i < layers.length - 1 ? 1 : 0);
+ }
+ int weightsCount = layers.length - 1;
+
+ RealMatrix[] initialWeights = new RealMatrix[weightsCount];
+
+ for (int i = 0; i < weightsCount; i++) {
+
+ RealMatrix matrix = MatrixUtils.createRealMatrix(layers[i + 1],
layers[i]);
+ final int finalI = i;
+ matrix.walkInOptimizedOrder(new RealMatrixChangingVisitor() {
+ @Override
+ public void start(int rows, int columns, int startRow, int endRow, int
startColumn, int endColumn) {
+
+ }
+
+ @Override
+ public double visit(int row, int column, double value) {
+ if (finalI != weightsCount - 1 && row == 0) {
+ return 0d;
+ } else if (column == 0) {
+ return 1d;
+ }
+ return r.nextInt(100) / 101d;
+ }
+
+ @Override
+ public double end() {
+ return 0;
+ }
+ });
+
+ initialWeights[i] = matrix;
+ }
+ return initialWeights;
+
+ }
+
+ // --- stochastic gradient descent ---
+
+ /**
+ * perform weights learning from the training examples using stochastic
gradient descent algorithm
+ *
+ * @param samples the training examples
+ * @return the final cost with the updated weights
+ * @throws Exception if SGD fails to converge or any numerical error happens
+ */
+ public double learnWeights(double[]... samples) throws Exception {
+ double newCost;
+ int iterations = 0;
+
+ double cost = Double.MAX_VALUE;
+ long start = System.currentTimeMillis();
+ while (true) {
+ if (iterations % (1 + (configuration.maxIterations / 100)) == 0) {
+ long time = (System.currentTimeMillis() - start) / 1000;
+ if (time > 60) {
+ System.out.println("cost is " + cost + " after " + iterations + "
iterations in " + (time / 60) + " minutes (" + ((double) iterations / time) + "
ips)");
+ }
+ }
+ // current training example
+ double[] sample = samples[iterations % samples.length];
+
+ int outputLayerSize = configuration.layers[configuration.layers.length -
1];
+ double[] expectedOutput = getSampleOutput(sample, outputLayerSize);
+ double[] input = getSampleInput(sample, outputLayerSize);
+
+ double[] predictedOutput = predictOutput(input); // TODO : use
debugOutput to avoid performing it again when calculating derivatives
+
+ // calculate cost
+ newCost = calculateCost(expectedOutput, predictedOutput, samples.length);
+
+ if (Double.POSITIVE_INFINITY == newCost) {
+ throw new Exception("failed to converge at iteration " + iterations +
" with alpha " + configuration.alpha + " : cost going from " + cost + " to " +
newCost);
+ } else if (iterations > 1 && (cost == newCost || newCost <
configuration.threshold || iterations > configuration.maxIterations)) {
+ System.out.println("successfully converged after " + (iterations - 1)
+ " iterations (alpha:" + configuration.alpha + ",threshold:" +
configuration.threshold + ") with cost " + newCost);
+ break;
+ } else if (Double.isNaN(newCost)) {
+ throw new Exception("failed to converge at iteration " + iterations +
" with alpha " + configuration.alpha + " : cost calculation underflow");
+ }
+
+ // update registered cost
+ cost = newCost;
+
+ // calculate the derivatives to update the parameters
+ RealMatrix[] derivatives = calculateDerivatives(input, expectedOutput,
sample.length);
+
+ // update the weights
+ weights = getUpdatedWeights(derivatives);
+
+ iterations++;
+ }
+ return newCost;
+ }
+
+ // --- sample parsing ---
+
+ private double[] getSampleInput(double[] sample, int outputLayerSize) {
+ double[] input = Arrays.copyOfRange(sample, outputLayerSize,
sample.length);
+ double[] result = new double[input.length + 1];
+ result[0] = 1d;
+ System.arraycopy(input, 0, result, 1, input.length);
+ return result;
+ }
+
+ private double[] getSampleOutput(double[] sample, int outputLayerSize) {
+ return Arrays.copyOfRange(sample, 0, outputLayerSize);
+ }
+
+ // --- backpropagation ---
+
+ private RealMatrix[] calculateDerivatives(double[] input, double[] output,
int size) throws Exception {
+ RealVector[] deltaVectors = new RealVector[weights.length];
+ RealMatrix[] deltas = new RealMatrix[weights.length];
+ RealMatrix[] ds = new RealMatrix[weights.length];
+
+ // compute delta vectors
+
+ // get activations from feed forward propagation
+ RealVector[] activations = applyFeedForward(input);
+
+ // calculate output error (corresponding to the last delta vector)
+ RealVector nextLayerDelta = calculateOutputError(output, activations);
+
+ deltaVectors[weights.length - 1] = nextLayerDelta;
+
+ for (int l = weights.length - 1; l > 0; l--) {
+ RealVector currentActivationsVector = activations[l - 1];
+ nextLayerDelta = calculateDeltaVector(weights[l],
currentActivationsVector, nextLayerDelta);
+
+ // collect delta vectors for this example
+ deltaVectors[l - 1] = nextLayerDelta;
+ }
+
+ RealVector[] newActivations = new RealVector[activations.length];
+ newActivations[0] = MatrixUtils.createRealVector(input);
+ System.arraycopy(activations, 0, newActivations, 1, activations.length -
1);
+
+ // compute deltas
+ for (int l = deltas.length - 1; l >= 0; l--) {
+ RealMatrix realMatrix = deltaVectors[l].outerProduct(newActivations[l]);
+ if (deltas[l] == null) {
+ deltas[l] = realMatrix;
+ } else {
+ deltas[l] = deltas[l].add(realMatrix);
+ }
+ }
+
+ // compute derivatives
+ for (int i = 0; i < deltas.length; i++) {
+ ds[i] = deltas[i].scalarMultiply(1d / size);
+ }
+
+ // regularization
+ int l = 0;
+ for (RealMatrix d : ds) {
+ final int finalL = l;
+ d.walkInOptimizedOrder(new RealMatrixChangingVisitor() {
+ @Override
+ public void start(int rows, int columns, int startRow, int endRow, int
startColumn, int endColumn) {
+
+ }
+
+ @Override
+ public double visit(int row, int column, double value) {
+ if (column != 0) {
+ return value + configuration.alpha * weights[finalL].getEntry(row,
column);
+ } else {
+ return value;
+ }
+ }
+
+ @Override
+ public double end() {
+ return 0;
+ }
+ });
+ l++;
+ }
+
+ return ds;
+ }
+
+ private RealVector calculateDeltaVector(RealMatrix weight, RealVector
activationsVector, RealVector nextLayerDelta) {
+ double[] ones = new double[activationsVector.getDimension()];
+ Arrays.fill(ones, 1d);
+ return
weight.preMultiply(nextLayerDelta).ebeMultiply(activationsVector.ebeMultiply(MatrixUtils.createRealVector(ones).subtract(activationsVector)));
+ }
+
+ private RealVector calculateOutputError(double[] expectedOutput,
RealVector[] activations) {
+ RealVector output = activations[activations.length - 1];
+ RealVector learnedOutputRealVector = new ArrayRealVector(expectedOutput);
+ return output.subtract(learnedOutputRealVector);
+ }
+
+ private RealMatrix[] getUpdatedWeights(final RealMatrix[] derivatives) {
+ RealMatrix[] updatedParameters = new RealMatrix[weights.length];
+
+ for (int l = 0; l < weights.length; l++) {
+ RealMatrix realMatrix = weights[l].copy();
+ final int finalL = l;
+ RealMatrixChangingVisitor visitor = new RealMatrixChangingVisitor() {
+
+ @Override
+ public void start(int rows, int columns, int startRow, int endRow, int
startColumn, int endColumn) {
+
+ }
+
+ @Override
+ public double visit(int row, int column, double value) {
+ if (!(row == 0 && value == 0d) && !(column == 0 && value == 1d)) {
+ return value - configuration.alpha *
derivatives[finalL].getEntry(row, column);
+ } else {
+ return value;
+ }
+ }
+
+ @Override
+ public double end() {
+ return 0;
+ }
+ };
+ realMatrix.walkInOptimizedOrder(visitor);
+ updatedParameters[l] = realMatrix;
+ }
+ return updatedParameters;
+ }
+
+ // --- cost ---
+
+ private double calculateCost(double[] expectedOutput, double[]
predictedOutput, int size) {
+ // neural network cost function
+ double res = 0d;
+ for (int i = 0; i < predictedOutput.length; i++) {
+ Double yo = expectedOutput[i];
+ Double ho = predictedOutput[i];
+ res += yo * Math.log(ho) + (1d - yo)
+ * Math.log(1d - ho);
+ }
+ return (-1d / size) * res;
+ }
+
+ // --- feed forward ---
+
+ public double[] predictOutput(double[] input) throws Exception {
+ if (input.length == configuration.layers[0]) {
+ double[] i2 = new double[input.length + 1];
+ i2[0] = 1d;
+ System.arraycopy(input, 0, i2, 1, input.length);
+ input = i2;
+ }
+ RealVector[] debugOutput = applyFeedForward(input);
+ RealVector d = debugOutput[debugOutput.length - 1];
+ return d.toArray();
+ }
+
+ private RealVector[] applyFeedForward(double[] input) throws Exception {
+ if (weights == null) {
+ throw new Exception("weights undefined, perform learning first");
+ }
+
+ RealVector[] debugOutput = new RealVector[weights.length];
+
+ RealMatrix x = MatrixUtils.createRowRealMatrix(input);
+ for (int w = 0; w < weights.length; w++) {
+ // compute matrix multiplication
+ x = x.multiply(weights[w].transpose());
+
+ // get activation function for w-th layer
+ int idx = configuration.activationFunctions.length == weights.length ? w
: 0;
+
+ // apply the activation function to each element in the matrix
+ x =
configuration.activationFunctions[idx].applyMatrix(x.getRowMatrix(0));
+ x = new SigmoidFunction().applyMatrix(x.getRowMatrix(0));
+
+ debugOutput[w] = x.getRowVector(0);
+ }
+ return debugOutput;
+ }
+
+ // --- neural network configuration ---
+
+ public static class Configuration {
+ protected int maxIterations;
+ protected double alpha;
+ protected double threshold;
+ protected int[] layers;
+ protected ActivationFunction[] activationFunctions;
+ }
+}
\ No newline at end of file
Propchange:
labs/yay/trunk/core/src/main/java/org/apache/yay/ShallowFeedForwardNeuralNetwork.java
------------------------------------------------------------------------------
svn:eol-style = native
Copied: labs/yay/trunk/core/src/main/java/org/apache/yay/SigmoidFunction.java
(from r1714192,
labs/yay/trunk/core/src/main/java/org/apache/yay/core/SigmoidFunction.java)
URL:
http://svn.apache.org/viewvc/labs/yay/trunk/core/src/main/java/org/apache/yay/SigmoidFunction.java?p2=labs/yay/trunk/core/src/main/java/org/apache/yay/SigmoidFunction.java&p1=labs/yay/trunk/core/src/main/java/org/apache/yay/core/SigmoidFunction.java&r1=1714192&r2=1724646&rev=1724646&view=diff
==============================================================================
--- labs/yay/trunk/core/src/main/java/org/apache/yay/core/SigmoidFunction.java
(original)
+++ labs/yay/trunk/core/src/main/java/org/apache/yay/SigmoidFunction.java Thu
Jan 14 16:19:32 2016
@@ -16,7 +16,7 @@
* specific language governing permissions and limitations
* under the License.
*/
-package org.apache.yay.core;
+package org.apache.yay;
import org.apache.commons.math3.linear.RealMatrix;
import org.apache.commons.math3.linear.RealMatrixChangingVisitor;
@@ -25,13 +25,9 @@ import org.apache.yay.ActivationFunction
/**
* See <a href="http://en.wikipedia.org/wiki/Sigmoid_function";>here</a>.
*/
-public class SigmoidFunction implements ActivationFunction<Double> {
+public class SigmoidFunction implements ActivationFunction {
- public Double apply(RealMatrix matrix, final Double input) {
- return sigmoid(input);
- }
-
- private double sigmoid(Double input) {
+ public double apply(Double input) {
return 1d / (1d + Math.exp(-1d * input));
}
@@ -46,7 +42,7 @@ public class SigmoidFunction implements
@Override
public double visit(int row, int column, double value) {
- return sigmoid(value);
+ return apply(value);
}
@Override
Copied:
labs/yay/trunk/core/src/main/java/org/apache/yay/SoftmaxActivationFunction.java
(from r1715735,
labs/yay/trunk/core/src/main/java/org/apache/yay/core/SoftmaxActivationFunction.java)
URL:
http://svn.apache.org/viewvc/labs/yay/trunk/core/src/main/java/org/apache/yay/SoftmaxActivationFunction.java?p2=labs/yay/trunk/core/src/main/java/org/apache/yay/SoftmaxActivationFunction.java&p1=labs/yay/trunk/core/src/main/java/org/apache/yay/core/SoftmaxActivationFunction.java&r1=1715735&r2=1724646&rev=1724646&view=diff
==============================================================================
---
labs/yay/trunk/core/src/main/java/org/apache/yay/core/SoftmaxActivationFunction.java
(original)
+++
labs/yay/trunk/core/src/main/java/org/apache/yay/SoftmaxActivationFunction.java
Thu Jan 14 16:19:32 2016
@@ -16,7 +16,7 @@
* specific language governing permissions and limitations
* under the License.
*/
-package org.apache.yay.core;
+package org.apache.yay;
import org.apache.commons.math3.linear.RealMatrix;
import org.apache.commons.math3.linear.RealMatrixChangingVisitor;
@@ -26,14 +26,7 @@ import org.apache.yay.ActivationFunction
/**
* Softmax activation function
*/
-public class SoftmaxActivationFunction implements ActivationFunction<Double> {
-
- @Override
- public Double apply(RealMatrix weights, Double signal) {
- double num = Math.exp(signal);
- double den = expDen(weights);
- return num / den;
- }
+public class SoftmaxActivationFunction implements ActivationFunction {
public RealMatrix applyMatrix(RealMatrix weights) {
RealMatrix matrix = weights.copy();
Copied:
labs/yay/trunk/core/src/main/java/org/apache/yay/StepActivationFunction.java
(from r1714192,
labs/yay/trunk/core/src/main/java/org/apache/yay/core/StepActivationFunction.java)
URL:
http://svn.apache.org/viewvc/labs/yay/trunk/core/src/main/java/org/apache/yay/StepActivationFunction.java?p2=labs/yay/trunk/core/src/main/java/org/apache/yay/StepActivationFunction.java&p1=labs/yay/trunk/core/src/main/java/org/apache/yay/core/StepActivationFunction.java&r1=1714192&r2=1724646&rev=1724646&view=diff
==============================================================================
---
labs/yay/trunk/core/src/main/java/org/apache/yay/core/StepActivationFunction.java
(original)
+++
labs/yay/trunk/core/src/main/java/org/apache/yay/StepActivationFunction.java
Thu Jan 14 16:19:32 2016
@@ -16,16 +16,15 @@
* specific language governing permissions and limitations
* under the License.
*/
-package org.apache.yay.core;
+package org.apache.yay;
import org.apache.commons.math3.linear.RealMatrix;
import org.apache.commons.math3.linear.RealMatrixChangingVisitor;
-import org.apache.yay.ActivationFunction;
/**
* An {@link org.apache.yay.ActivationFunction} implementing a step function
of the input
*/
-public class StepActivationFunction implements ActivationFunction<Double> {
+public class StepActivationFunction implements ActivationFunction {
private final double center;
@@ -33,11 +32,6 @@ public class StepActivationFunction impl
this.center = center;
}
- @Override
- public Double apply(RealMatrix matrix, Double signal) {
- return step(signal);
- }
-
private double step(Double signal) {
return signal >= center ? 1d : 0d;
}
Copied: labs/yay/trunk/core/src/main/java/org/apache/yay/TanhFunction.java
(from r1714192,
labs/yay/trunk/core/src/main/java/org/apache/yay/core/TanhFunction.java)
URL:
http://svn.apache.org/viewvc/labs/yay/trunk/core/src/main/java/org/apache/yay/TanhFunction.java?p2=labs/yay/trunk/core/src/main/java/org/apache/yay/TanhFunction.java&p1=labs/yay/trunk/core/src/main/java/org/apache/yay/core/TanhFunction.java&r1=1714192&r2=1724646&rev=1724646&view=diff
==============================================================================
--- labs/yay/trunk/core/src/main/java/org/apache/yay/core/TanhFunction.java
(original)
+++ labs/yay/trunk/core/src/main/java/org/apache/yay/TanhFunction.java Thu Jan
14 16:19:32 2016
@@ -16,7 +16,7 @@
* specific language governing permissions and limitations
* under the License.
*/
-package org.apache.yay.core;
+package org.apache.yay;
import org.apache.commons.math3.linear.RealMatrix;
import org.apache.commons.math3.linear.RealMatrixChangingVisitor;
@@ -25,11 +25,7 @@ import org.apache.yay.ActivationFunction
/**
* Tanh activation function
*/
-public class TanhFunction implements ActivationFunction<Double> {
- @Override
- public Double apply(RealMatrix matrix, Double signal) {
- return Math.tanh(signal);
- }
+public class TanhFunction implements ActivationFunction {
@Override
public RealMatrix applyMatrix(RealMatrix weights) {
Added:
labs/yay/trunk/core/src/test/java/org/apache/yay/ShallowFeedForwardNeuralNetworkTest.java
URL:
http://svn.apache.org/viewvc/labs/yay/trunk/core/src/test/java/org/apache/yay/ShallowFeedForwardNeuralNetworkTest.java?rev=1724646&view=auto
==============================================================================
---
labs/yay/trunk/core/src/test/java/org/apache/yay/ShallowFeedForwardNeuralNetworkTest.java
(added)
+++
labs/yay/trunk/core/src/test/java/org/apache/yay/ShallowFeedForwardNeuralNetworkTest.java
Thu Jan 14 16:19:32 2016
@@ -0,0 +1,65 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied. See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+package org.apache.yay;
+
+import org.junit.Test;
+
+import static org.junit.Assert.assertEquals;
+import static org.junit.Assert.assertNotNull;
+import static org.junit.Assert.assertTrue;
+
+/**
+ * Tests for {@link ShallowFeedForwardNeuralNetwork}
+ */
+public class ShallowFeedForwardNeuralNetworkTest {
+
+ @Test
+ public void testLearnAndPredict() throws Exception {
+ ShallowFeedForwardNeuralNetwork.Configuration configuration = new
ShallowFeedForwardNeuralNetwork.Configuration();
+ configuration.alpha = 0.0001d;
+ configuration.layers = new int[]{3, 4, 1};
+ configuration.maxIterations = 10000;
+ configuration.threshold = 0.004d;
+ configuration.activationFunctions = new ActivationFunction[]{new
SigmoidFunction()};
+
+ ShallowFeedForwardNeuralNetwork neuralNetwork = new
ShallowFeedForwardNeuralNetwork(configuration);
+
+ assertNotNull(neuralNetwork);
+ double[][] samples = new double[3][4];
+ samples[0][0] = 0.1d;
+ samples[0][1] = 0.2d;
+ samples[0][2] = 0.3d;
+ samples[0][3] = 0.4d;
+ samples[1][0] = 0.5d;
+ samples[1][1] = 0.6d;
+ samples[1][2] = 0.7d;
+ samples[1][3] = 0.8d;
+ samples[2][0] = 0.9d;
+ samples[2][1] = 0.1d;
+ samples[2][2] = 0.2d;
+ samples[2][3] = 0.3d;
+ double cost = neuralNetwork.learnWeights(samples);
+ assertTrue(cost > 0 && cost < 10);
+
+ double[] doubles = neuralNetwork.predictOutput(new double[]{0.7d, 0.8d,
0.9d});
+ assertNotNull(doubles);
+
+ assertEquals(0.4d, doubles[0], 0.4d);
+ }
+}
\ No newline at end of file
Propchange:
labs/yay/trunk/core/src/test/java/org/apache/yay/ShallowFeedForwardNeuralNetworkTest.java
------------------------------------------------------------------------------
svn:eol-style = native
Copied:
labs/yay/trunk/core/src/test/java/org/apache/yay/SigmoidFunctionTest.java (from
r1707760,
labs/yay/trunk/core/src/test/java/org/apache/yay/core/SigmoidFunctionTest.java)
URL:
http://svn.apache.org/viewvc/labs/yay/trunk/core/src/test/java/org/apache/yay/SigmoidFunctionTest.java?p2=labs/yay/trunk/core/src/test/java/org/apache/yay/SigmoidFunctionTest.java&p1=labs/yay/trunk/core/src/test/java/org/apache/yay/core/SigmoidFunctionTest.java&r1=1707760&r2=1724646&rev=1724646&view=diff
==============================================================================
---
labs/yay/trunk/core/src/test/java/org/apache/yay/core/SigmoidFunctionTest.java
(original)
+++ labs/yay/trunk/core/src/test/java/org/apache/yay/SigmoidFunctionTest.java
Thu Jan 14 16:19:32 2016
@@ -16,33 +16,34 @@
* specific language governing permissions and limitations
* under the License.
*/
-package org.apache.yay.core;
+package org.apache.yay;
+import org.apache.yay.SigmoidFunction;
import org.junit.Test;
import static org.junit.Assert.assertEquals;
/**
- * Tests for {@link org.apache.yay.core.SigmoidFunction}
+ * Tests for {@link SigmoidFunction}
*/
public class SigmoidFunctionTest {
@Test
public void testCorrectOutput() throws Exception {
SigmoidFunction sigmoidFunction = new SigmoidFunction();
- Double output = sigmoidFunction.apply(null, 38d);
+ Double output = sigmoidFunction.apply(38d);
assertEquals(Double.valueOf(1d), output);
- output = sigmoidFunction.apply(null, 6d);
+ output = sigmoidFunction.apply(6d);
assertEquals(Double.valueOf(0.9975273768433653d), output);
- output = sigmoidFunction.apply(null, 2.5d);
+ output = sigmoidFunction.apply(2.5d);
assertEquals(Double.valueOf(0.9241418199787566d), output);
- output = sigmoidFunction.apply(null, -2.5d);
+ output = sigmoidFunction.apply(-2.5d);
assertEquals(Double.valueOf(0.07585818002124355d), output);
- output = sigmoidFunction.apply(null, 0d);
+ output = sigmoidFunction.apply(0d);
assertEquals(Double.valueOf(0.5d), output);
}
}
Modified: labs/yay/trunk/pom.xml
URL:
http://svn.apache.org/viewvc/labs/yay/trunk/pom.xml?rev=1724646&r1=1724645&r2=1724646&view=diff
==============================================================================
--- labs/yay/trunk/pom.xml (original)
+++ labs/yay/trunk/pom.xml Thu Jan 14 16:19:32 2016
@@ -5,7 +5,7 @@
<groupId>org.apache.yay</groupId>
<artifactId>yay-parent</artifactId>
<packaging>pom</packaging>
- <version>0.1-SNAPSHOT</version>
+ <version>0.2-SNAPSHOT</version>
<name>Yay parent</name>
<url>http://svn.apache.org/repos/asf/labs/yay</url>
<organization>
@@ -206,7 +206,6 @@
</resources>
</build>
<modules>
- <module>api</module>
<module>core</module>
</modules>
<profiles>
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@labs.apache.org
For additional commands, e-mail: commits-h...@labs.apache.org
