Package aima.core.util.math

Examples of aima.core.util.math.Matrix

50
51
52
53
54
55
56
57
58
59
60
61
62
63
  public Layer(int numberOfNeurons, int numberOfInputs,
      double lowerLimitForWeights, double upperLimitForWeights,
      ActivationFunction af) {

    activationFunction = af;
    this.weightMatrix = new Matrix(numberOfNeurons, numberOfInputs);
    lastWeightUpdateMatrix = new Matrix(weightMatrix.getRowDimension(),
        weightMatrix.getColumnDimension());
    penultimateWeightUpdateMatrix = new Matrix(
        weightMatrix.getRowDimension(),
        weightMatrix.getColumnDimension());

    this.biasVector = new Vector(numberOfNeurons);
    lastBiasUpdateVector = new Vector(biasVector.getRowDimension());
View Full Code Here
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
    initializeVector(biasVector, lowerLimitForWeights, upperLimitForWeights);
  }

  public Vector feedForward(Vector inputVector) {
    lastInput = inputVector;
    Matrix inducedField = weightMatrix.times(inputVector).plus(biasVector);

    Vector inducedFieldVector = new Vector(numberOfNeurons());
    for (int i = 0; i < numberOfNeurons(); i++) {
      inducedFieldVector.setValue(i, inducedField.get(i, 0));
    }

    lastInducedField = inducedFieldVector.copyVector();
    Vector resultVector = new Vector(numberOfNeurons());
    for (int i = 0; i < numberOfNeurons(); i++) {
View Full Code Here
147
148
149
150
151
152
153
154
155
156
157
158
  public void updateWeights() {
    weightMatrix.plusEquals(lastWeightUpdateMatrix);
  }

  public void updateBiases() {
    Matrix biasMatrix = biasVector.plusEquals(lastBiasUpdateVector);
    Vector result = new Vector(biasMatrix.getRowDimension());
    for (int i = 0; i < biasMatrix.getRowDimension(); i++) {
      result.setValue(i, biasMatrix.get(i, 0));
    }
    biasVector = result;
  }
View Full Code Here
33
34
35
36
37
38
39
40
    List<Double> values = new ArrayList<Double>();
    // for (String state : aBelief.states()) {
    for (String state : states) {
      values.add(get(state, perception));
    }
    Matrix OMatrix = Matrix.createDiagonalMatrix(values);
    return OMatrix;
  }
View Full Code Here
29
30
31
32
33
34
35
36
37
38
39
  }

  public VarDistribution predict(VarDistribution aBelief, String action) {
    VarDistribution newBelief = aBelief.duplicate();

    Matrix beliefMatrix = aBelief.asMatrix();
    Matrix transitionMatrix = transitionModel.asMatrix(action);
    Matrix predicted = transitionMatrix.transpose().times(beliefMatrix);
    newBelief.updateFrom(predicted);
    return newBelief;
  }
View Full Code Here
41
42
43
44
45
46
47
48
49
50
51
52
53
  public VarDistribution perceptionUpdate(VarDistribution aBelief,
      String perception) {
    VarDistribution newBelief = aBelief.duplicate();

    // one way - use matrices
    Matrix beliefMatrix = aBelief.asMatrix();
    Matrix o_matrix = sensorModel.asMatrix(perception);
    Matrix updated = o_matrix.times(beliefMatrix);
    newBelief.updateFrom(updated);
    newBelief.normalize();
    return newBelief;

    // alternate way of doing this. clearer in intent.
View Full Code Here
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
  public VarDistribution calculate_next_backward_message(
      VarDistribution present_backward_message, String perception) {
    VarDistribution result = present_backward_message.duplicate();
    // System.out.println("fb :-calculating new backward message");
    // System.out.println("fb :-diagonal matrix from sens model = ");
    Matrix oMatrix = sensorModel.asMatrix(perception);
    // System.out.println(oMatrix);
    Matrix transitionMatrix = transitionModel.asMatrix();// action
    // should
    // be
    // passed
    // in
    // here?
    // System.out.println("fb :-present backward message = "
    // +present_backward_message);
    Matrix backwardMatrix = transitionMatrix.times(oMatrix
        .times(present_backward_message.asMatrix()));

    result.updateFrom(backwardMatrix);
    result.normalize();
    // System.out.println("fb :-normalized new backward message = "
View Full Code Here
80
81
82
83
84
85
86
87
88
89
90
91
      distribution.put(states.get(i), newProbs.get(i));
    }
  }

  public Matrix asMatrix() {
    Matrix m = new Matrix(states.size(), 1);
    for (int i = 0; i < states.size(); i++) {
      m.set(i, 0, distribution.get(states.get(i)));
    }
    return m;

  }
View Full Code Here
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
  }

  public VarDistribution smooth(String perception) {

    evidenceFromSmoothedStepToPresent.add(perception);
    Matrix O_t = hmm.sensorModel().asMatrix(perception);
    Matrix transitionMatrix = hmm.transitionModel().asMatrix();
    if (time > timelag) {

      forwardMessage = hmm.forward(forwardMessage, perception); // This
      // seems
      // WRONG
      // I think this should be
      // forwardMessage = hmm.forward(forwardMessage,
      // evidenceFromSmoothedStepToPresent.get(0));
      // this the perception at t-d. the book's algorithm
      // uses the latest perception.
      evidenceFromSmoothedStepToPresent.remove(0);
      Matrix O_t_minus_d = hmm.sensorModel().asMatrix(
          evidenceFromSmoothedStepToPresent.get(0));

      B = O_t_minus_d.inverse().times(
          transitionMatrix.inverse().times(
              B.times(transitionMatrix.times(O_t))));

    } else {

      B = B.times(transitionMatrix.times(O_t));

    }
    time += 1;
    if (time > timelag) {

      Matrix one = hmm.prior().createUnitBelief().asMatrix();
      Matrix forwardMatrix = forwardMessage.asMatrix();
      VarDistribution result = hmm.prior().duplicate();
      Matrix backwardMessage = (B.times(one));

      result.updateFrom(forwardMatrix.arrayTimes(backwardMessage));

      result.normalize();
      return result;
View Full Code Here
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
  public double get(String old_state_action, String newState) {
    return table.get(old_state_action, newState);
  }

  public Matrix asMatrix(String action) {
    Matrix transitionMatrix = new Matrix(states.size(), states.size());
    for (int i = 0; i < states.size(); i++) {
      String oldState = states.get(i);
      String old_state_action = oldState.concat(action);
      for (int j = 0; j < states.size(); j++) {
        String newState = states.get(j);
        double transitionProbability = get(old_state_action, newState);
        transitionMatrix.set(i, j, transitionProbability);
      }
    }
    return transitionMatrix;
  }
View Full Code Here
TOP

Related Classes of aima.core.util.math.Matrix

  • aima.core.learning.neural.BackPropLearning
  • aima.core.learning.neural.Layer
  • aima.core.learning.neural.LayerSensitivity
  • aima.core.learning.neural.Perceptron
  • aima.core.probability.example.HMMExampleFactory
  • aima.core.probability.hmm.exact.FixedLagSmoothing
  • aima.core.probability.hmm.exact.HMMForwardBackward
  • aima.core.probability.hmm.exact.HMMForwardBackwardConstantSpace
  • aima.core.probability.hmm.FixedLagSmoothing
  • aima.core.probability.hmm.HiddenMarkovModel
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.