Create Presentation
Download Presentation

Download Presentation
## Artificial Neural Networks

- - - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - - -

**Introduction**• Artificial Neural Networks (ANN) • Information processing paradigm inspired by biological nervous systems • ANN is composed of a system of neurons connected by synapses • ANN learn by example • Adjust synaptic connections between neurons**History**• 1943: McCulloch and Pitts model neural networks based on their understanding of neurology. • Neurons embed simple logic functions: • a or b • a and b • 1950s: • Farley and Clark • IBM group that tries to model biological behavior • Consult neuro-scientists at McGill, whenever stuck • Rochester, Holland, Haibit and Duda**History**• Perceptron (Rosenblatt 1958) • Three layer system: • Input nodes • Output node • Association layer • Can learn to connect or associate a given input to a random output unit • Minsky and Papert • Showed that a single layer perceptron cannot learn the XOR of two binary inputs • Lead to loss of interest (and funding) in the field**History**• Perceptron (Rosenblatt 1958) • Association units A1, A2, … extract features from user input • Output is weighted and associated • Function fires if weighted sum of input exceeds a threshold.**History**• Back-propagation learning method (Werbos 1974) • Three layers of neurons • Input, Output, Hidden • Better learning rule for generic three layer networks • Regenerates interest in the 1980s • Successful applications in medicine, marketing, risk management, … (1990) • In need for another breakthrough.**ANN**• Promises • Combine speed of silicon with proven success of carbon artificial brains**Neuron Model**• Natural neurons**Neuron Model**• Neuron collects signals from dendrites • Sends out spikes of electrical activity through an axon, which splits into thousands of branches. • At end of each brand, a synapses converts activity into either exciting or inhibiting activity of a dendrite at another neuron. • Neuron fires when exciting activity surpasses inhibitory activity • Learning changes the effectiveness of the synapses**Neuron Model**• Natural neurons**Neuron Model**• Abstract neuron model:**ANN Forward Propagation**• Bias Nodes • Add one node to each layer that has constant output • Forward propagation • Calculate from input layer to output layer • For each neuron: • Calculate weighted average of input • Calculate activation function**Neuron Model**• Firing Rules: • Threshold rules: • Calculate weighted average of input • Fire if larger than threshold • Perceptron rule • Calculate weighted average of input input • Output activation level is**Neuron Model**• Firing Rules: Sigmoid functions: • Hyperbolic tangent function • Logistic activation function**ANN Forward Propagation**• Apply input vector X to layer of neurons. • Calculate • where Xi is the activation of previous layer neuron i • Wji is the weight of going from node i to node j • p is the number of neurons in the previous layer • Calculate output activation**0**1 2 3 ANN Forward Propagation • Example: ADALINE Neural Network • Calculates and of inputs w0,3=.6 threshold function is step function w1,3=.6 w2,3=-.9 Bias Node**0**1 2 3 4 ANN Forward Propagation • Example: Three layer network • Calculates xor of inputs -4.8 5.9 5.1 4.6 5.2 -5.2 -2.6 -2.7 -3.2 Bias**0**1 2 3 4 ANN Forward Propagation • Input (0,0) -4.8 5.9 5.1 4.6 5.2 -5.2 -2.6 -2.7 -3.2 Bias**0**1 2 3 4 ANN Forward Propagation • Input (0,0) • Node 2 activation is (-4.8 0+4.6 0 - 2.6)= 0.0691 -4.8 5.9 5.1 4.6 5.2 -5.2 -2.6 -2.7 -3.2 Bias**0**1 2 3 4 ANN Forward Propagation • Input (0,0) • Node 3 activation is (5.1 0 - 5.2 0 - 3.2)= 0.0392 -4.8 5.9 5.1 4.6 5.2 -5.2 -2.6 -2.7 -3.2 Bias**0**1 2 3 4 ANN Forward Propagation • Input (0,0) • Node 4 activation is (5.9 0.069 + 5.2 0.069 – 2.7)= 0.110227 -4.8 5.9 5.1 4.6 5.2 -5.2 -2.6 -2.7 -3.2 Bias**0**1 2 3 4 ANN Forward Propagation • Input (0,1) • Node 2 activation is (4.6 -2.6)= 0.153269 -4.8 5.9 5.1 4.6 5.2 -5.2 -2.6 -2.7 -3.2 Bias**0**1 2 3 4 ANN Forward Propagation • Input (0,1) • Node 3 activation is (-5.2 -3.2)= 0.000224817 -4.8 5.9 5.1 4.6 5.2 -5.2 -2.6 -2.7 -3.2 Bias**0**1 2 3 4 ANN Forward Propagation • Input (0,1) • Node 4 activation is (5.9 0.153269 + 5.2 0.000224817 -2.7 )= 0.923992 -4.8 5.9 5.1 4.6 5.2 -5.2 -2.6 -2.7 -3.2 Bias**ANN Forward Propagation**• Density Plot of Output**ANN Forward Propagation**• Network can learn a non-linearly separated set of outputs. • Need to map output (real value) into binary values.**ANN Training**• Weights are determined by training • Back-propagation: • On given input, compare actual output to desired output. • Adjust weights to output nodes. • Work backwards through the various layers • Start out with initial random weights • Best to keep weights close to zero (<<10)**ANN Training**• Weights are determined by training • Need a training set • Should be representative of the problem • During each training epoch: • Submit training set element as input • Calculate the error for the output neurons • Calculate average error during epoch • Adjust weights**ANN Training**• Error is the mean square of differences in output layer y – observed output t – target output**ANN Training**• Error of training epoch is the average of all errors.**ANN Training**• Update weights and thresholds using • Weights • Bias • is a possibly time-dependent factor that should prevent overcorrection**ANN Training**• Using a sigmoid function, we get • Logistics function has derivative ’(t) = (t)(1- (t))**-0,5**0 1 2 3 4 0.1 -0.5 1 -0.5 1 -0.5 1 -1 Bias ANN Training Example • Start out with random, small weights**-0,5**0 1 2 3 4 0.1 -0.5 1 -0.5 1 -0.5 1 -1 Bias ANN Training Example Average Error is 0.283038**-0,5**0 1 2 3 4 0.1 -0.5 1 -0.5 1 -0.5 1 -1 Bias ANN Training Example Average Error is 0.283038**ANN Training Example**• Calculate the derivative of the error with respect to the weights and bias into the output layer neurons**-0,5**0 1 2 3 4 0.1 -0.5 1 -0.5 1 -0.5 1 -1 Bias ANN Training Example New weights going into node 4 We do this for all training inputs, then average out the changes net4 is the weighted sum of input going into neuron 4: net4(0,0)= 0.787754 net4(0,1)= 0.696717 net4(1,0)= 0.838124 net4(1,1)= 0.73877**-0,5**0 1 2 3 4 0.1 -0.5 1 -0.5 1 -0.5 1 -1 Bias ANN Training Example New weights going into node 4 We calculate the derivative of the activation function at the point given by the net-input. Recall our cool formula ’(t) = (t)(1- (t)) ’( net4(0,0)) = ’( 0.787754) = 0.214900 ’( net4(0,1)) = ’( 0.696717) = 0.221957 ’( net4(1,0)) = ’( 0.838124) = 0.210768 ’( net4(1,1)) = ’( 0.738770) = 0.218768**-0,5**0 1 2 3 4 0.1 -0.5 1 -0.5 1 -0.5 1 -1 Bias ANN Training Example New weights going into node 4 We now obtain values for each input separately: Input 0,0: 4= ’( net4(0,0)) *(0-y4(0,0)) = -0.152928 Input 0,1: 4= ’( net4(0,1)) *(1-y4(0,1)) = 0.0682324 Input 1,0: 4= ’( net4(1,0)) *(1-y4(1,0)) = 0.0593889 Input 1,1: 4= ’( net4(1,1)) *(0-y4(1,1)) = -0.153776 Average: 4 = -0.0447706**-0,5**0 1 2 3 4 0.1 -0.5 1 -0.5 1 -0.5 1 -1 Bias ANN Training Example New weights going into node 4 Average: 4 = -0.0447706 We can now update the weights going into node 4: Let’s call: Eji the derivative of the error function with respect to the weight going from neuron i into neuron j. We do this for every possible input: E4,2 = - output(neuron(2)* 4 For (0,0): E4,2 = 0.0577366 For (0,1): E4,2 = -0.0424719 For(1,0): E4,2 = -0.0159721 For(1,1): E4,2 = 0.0768878 Average is 0.0190451**-0,5**0 1 2 3 4 0.1 -0.5 1 -0.5 1 -0.5 1 -1 Bias ANN Training Example New weight from 2 to 4 is now going to be 0.1190451.**-0,5**0 1 2 3 4 0.1 -0.5 1 -0.5 1 -0.5 1 -1 Bias ANN Training Example New weights going into node 4 For (0,0): E4,3 = 0.0411287 For (0,1): E4,3 = -0.0341162 For(1,0): E4,3 = -0.0108341 For(1,1): E4,3 = 0.0580565 Average is 0.0135588 New weight is -0.486441**-0,5**0 1 2 3 4 0.1 -0.5 1 -0.5 1 -0.5 1 -1 Bias ANN Training Example New weights going into node 4: We also need to change the bias node For (0,0): E4,B = 0.0411287 For (0,1): E4,B = -0.0341162 For(1,0): E4,B = -0.0108341 For(1,1): E4,B = 0.0580565 Average is 0.0447706 New weight is 1.0447706**ANN Training Example**• We now have adjusted all the weights into the output layer. • Next, we adjust the hidden layer • The target output is given by the delta values of the output layer • More formally: • Assume that j is a hidden neuron • Assume that k is the delta-value for an output neuron k. • While the example has only one output neuron, most ANN have more. When we sum over k, this means summing over all output neurons. • wkj is the weight from neuron j into neuron k**-0,5**0 1 2 3 4 0.1 -0.5 1 -0.5 1 -0.5 1 -1 Bias ANN Training Example We now calculate the updates to the weights of neuron 2. First, we calculate the net-input into 2. This is really simple because it is just a linear functions of the arguments x1 and x2 net2 = -0.5 x1 + x2 - 0.5 We obtain 2 (0,0) = - 0.00359387 2(0,1) = 0.00160349 2(1,0) = 0.00116766 2(1,1) = - 0.00384439**-0,5**0 1 2 3 4 0.1 -0.5 1 -0.5 1 -0.5 1 -1 Bias ANN Training Example Call E20 the derivative of E with respect to w20. We use the output activation for the neurons in the previous layer (which happens to be the input layer) E20 (0,0) = - (0)2 (0,0) = 0.00179694 E20(0,1) = 0.00179694 E20(1,0) = -0.000853626 E20(1,1) = 0.00281047 The average is 0.00073801 and the new weight is -0.499262**-0,5**0 1 2 3 4 0.1 -0.5 1 -0.5 1 -0.5 1 -1 Bias ANN Training Example Call E21 the derivative of E with respect to w21. We use the output activation for the neurons in the previous layer (which happens to be the input layer) E21 (0,0) = - (1)2 (0,0) = 0.00179694 E21(0,1) = -0.00117224 E21(1,0) = -0.000583829 E21(1,1) = 0.00281047 The average is 0.000712835 and the new weight is 1.00071