Backpropagation

Networks
 Introduction to
Backpropagation
- In 1969 a method for learning in multi-layer network,
Backpropagation,
Backpropagation was invented by Bryson and Ho.

- The Backpropagation algorithm is a sensible approach

for dividing the contribution of each weight.

- Works basically the same as perceptrons

 Backpropagation Learning Principles:
Hidden Layers and Gradients

There are two differences for the updating rule :

1) The activation of the hidden unit is used instead of

activation of the input value.

2) The rule contains a term for the gradient of the activation

function.
Backpropagation Network
training
• 1. Initialize network with random weights
• 2. For all training cases (called examples):
– a. Present training inputs to network and
calculate output
– b. For all layers (starting with output layer,
back to input layer):
• i. Compare network output with correct output
(error function)
This is
• ii. Adapt weights in current layer what
you
want
 Backpropagation Learning
Details
• Method for learning weights in feed-forward (FF)
nets

• Can’t use Perceptron Learning Rule

– no teacher values are possible for hidden units

• Use gradient descent to minimize the error

– propagate deltas to adjust for errors
backward from outputs
to hidden layers forward
to inputs
backward
 Backpropagation Algorithm – Main
Idea – error in hidden layers

The ideas of the algorithm can be summarized as follows :

1. Computes the error term for the output units using the
observed error.

2. From output layer, repeat

- propagating the error term back to the previous layer
and
- updating the weights between the two layers
until the earliest hidden layer is reached.
 Backpropagation Algorithm
• Initialize weights (typically random!)
• Keep doing epochs
– For each example e in training set do
• forward pass to compute
– O = neural-net-output(network,e)
– miss = (T-O) at each output unit
• backward pass to calculate deltas to weights
• update all weights
– end
• until tuning set error stops improving
Forward pass explained Backward pass explained in next slide
earlier
 Backward Pass
• Compute deltas to weights
– from hidden layer
– to output layer

• Without changing any weights (yet),

compute the actual contributions
– within the hidden layer(s)
– and compute deltas
 Gradient Descent
• Think of the N weights as a point in an N-
dimensional space

• Add a dimension for the observed error

• Try to minimize your position on the “error

surface”
 Error Surface
error

weights

Error as function of
weights in
multidimensional space
Compute
deltas Gradient
• Trying to make error decrease the fastest
• Compute:
• GradE = [dE/dw1, dE/dw2, . . ., dE/dwn]
• Change i-th weight by
• deltawi = -alpha * dE/dwi
Derivatives of error for weights

• We need a derivative!
• Activation function must be continuous,
continuous
differentiable, non-decreasing, and easy to
compute
 Can’t use LTU
• To effectively assign credit / blame to units
in hidden layers, we want to look at the
first derivative of the activation function

• Sigmoid function is easy to differentiate

and easy to compute forward

Linear Threshold Units Sigmoid function

 Updating hidden-to-output
• We have teacher supplied desired values

• deltawji = α * aj * (Ti - Oi) * g’(ini)

= α * aj * (Ti - Oi) * Oi * (1 - Oi)

– for sigmoid the derivative is, g’(x) = g(x) * (1 - g(x))

derivative

alpha Here we have

general formula with
derivative, next we
use for sigmoid miss
 Updating interior weights
• Layer k units provide values to all layer
k+1 units
• “miss” is sum of misses from all units on k+1
• missj = Σ [ ai(1- ai) (Ti - ai) wji ]
• weights coming into this unit are adjusted
based on their contribution
deltakj = α * Ik * aj * (1 - aj) * missj For layer k+1

Compute deltas
 How do we pick α ?
1. Tuning set, or

2. Cross validation, or

3. Small for slow, conservative learning

 How many hidden layers?
• Usually just one (i.e., a 2-layer net)

• How many hidden units in the layer?

– Too few ==> can’t learn
– Too many ==> poor generalization
 How big a training set?
• Determine your target error rate, e
• Success rate is 1- e
• Typical training set approx. n/e, where n is the
number of weights in the net
• Example:
– e = 0.1, n = 80 weights
– training set size 800
trained until 95% correct training set classification
should produce 90% correct classification
on testing set (typical)
 Examples of Backpropagation
Learning

In the restaurant
problem NN was
worse than the
decision tree

Error Decision tree still

decreases better for
with number restaurant
of epochs example
Examples of Backpropagation Learning

Majority example, Restaurant

perceptron better example, DT
better
Backpropagation
Learning Math

See next
slide for
explanation
 Visualization of
Backpropagation
learning

Backprop output layer

 Bias Neurons in Backpropagation
Learning

bias neuron in input layer

Software for Backpropagation Learning
This routine
calculate error for
Training pairs backpropagation

Run network forward.

Was explained earlier

Calculate difference to
desired output

Calculate total error

Software for Backpropagation Learning
continuation
Here we do not use Update output weights
alpha, the learning rate

Calculate hidden difference values

Update input weights

Return total error

 The general Backpropagation Algorithm for updating weights in a multilayer network

Here we use alpha, the

Repeat until
learning rate
convergent

Go through all Run network to

calculate its
examples output for this
example

Compute the
error in output

Update weights
to output layer

Compute error in
each hidden layer

Update weights in
each hidden layer
Return learned network
Examples and
Applications
of ANN
 Neural Network in Practice

NNs are used for classification and function approximation

or mapping problems which are:

- Tolerant of some imprecision.

- Have lots of training data available.

- Hard and fast rules cannot easily be applied.

 NETalk (1987)
• Mapping character strings into phonemes so they
can be pronounced by a computer

• Neural network trained how to pronounce each

letter in a word in a sentence, given the three
letters before and three letters after it in a window

• Output was the correct phoneme

• Results
– 95% accuracy on the training data
– 78% accuracy on the test set
 Other Examples
• Neurogammon (Tesauro & Sejnowski, 1989)
– Backgammon learning program

• Speech Recognition (Waibel, 1989)

• Character Recognition (LeCun et al., 1989)

• Face Recognition (Mitchell)

 ALVINN
• Steer a van down the road
– 2-layer feedforward
• using backpropagation for learning

– Raw input is 480 x 512 pixel image 15x per sec

– Color image preprocessed into 960 input units

– 4 hidden units

– 30 output units, each is a steering direction

 Neural Network Approaches
ALVINN - Autonomous Land Vehicle In a Neural Network
• ALVINN learned as the vehicle
traveled Learning on-the-
– initially by observing a human
driving
fly
– learns from its own driving by
watching for future corrections

– never saw bad driving

• didn’t know what was
dangerous, NOT correct
• computes alternate views of
the road (rotations, shifts, and
fill-ins) to use as “bad”
examples

– keeps a buffer pool of 200 pretty

old examples to avoid overfitting
to only the most recent images
 Feed-forward vs. Interactive
Nets
• Feed-forward
– activation propagates in one direction
– We usually focus on this

• Interactive
– activation propagates forward & backwards
– propagation continues until equilibrium is reached in
the network
– We do not discuss these networks here, complex
training. May be unstable.
 Ways of learning with an ANN
• Add nodes & connections
• Subtract nodes & connections
• Modify connection weights
– current focus
– can simulate first two

• I/O pairs:
– given the inputs, what should the output be?
[“typical” learning problem]
 More Neural Network
Applications
- May provide a model for massive parallel computation.

- More successful approach of “parallelizing” traditional

serial algorithms.

- Can compute any computable function.

- Can do everything a normal digital computer can do.

- Can do even more under some impractical assumptions.

 Neural Network Approaches to driving
•Use special hardware
•ASIC
•FPGA
•analog
- Developed in 1993.
Output units

- Performs driving with

Hidden layer Neural Networks.

- An intelligent VLSI image

sensor for road following.
Input units
- Learns to filter out image
details not relevant to
driving.
 Neural Network Approaches

Input Array Hidden Units Output units

 Actual Products Available

ex1. Enterprise Miner:

- Single multi-layered feed-forward neural networks.

- Provides business solutions for data mining.

ex2. Nestor:
- Uses Nestor Learning System (NLS).
- Several multi-layered feed-forward neural networks.
- Intel has made such a chip - NE1000 in VLSI technology.
 Ex1. Software tool - Enterprise Miner

- Based on SEMMA (Sample, Explore, Modify, Model,

Access) methodology.

- Statistical tools include :

Clustering, decision trees, linear and logistic
regression and neural networks.

- Data preparation tools include :

Outliner detection, variable transformation, random
sampling, and partition of data sets (into training,
testing and validation data sets).
 Ex 2. Hardware Tool - Nestor

- With low connectivity within each layer.

- Minimized connectivity within each layer results in rapid

training and efficient memory utilization, ideal for VLSI.

- Composed of multiple neural networks, each specializing

in a subset of information about the input patterns.

- Real time operation without the need of special computers

or custom hardware DSP platforms

•Software exists.
 Problems with using ANNs
1. Insufficiently characterized development
process compared with conventional software
– What are the steps to create a neural network?

1. How do we create neural networks in a

repeatable and predictable manner?

2. Absence of quality assurance methods for

neural network models and implementations
– How do I verify my implementation?
Solving Problem 1 – The Steps to create
a ANN
Define the process of developing neural networks:
1. Formally capture the specifics of the problem in
a document based on a template

2. Define the factors/parameters for creation

– Neural network creation parameters
– Performance requirements

1. Create the neural network

2. Get feedback on performance

Neural Network Development Process
 Problem Specification Phase
• Some factors to define in problem specification:
– Type of neural networks (based on experience or
published results)
– How to collect and transform problem data
– Potential input/output representations
– Training & testing method and data selection
– Performance targets (accuracy and precision)

• Most important output is the ranked collection of

factors/parameters
 Problem 2 –
How to create a Neural Network

• Predictability (with regard to resources)

– Depending on creation approach used, record time
for one iteration
– Use timing to predict maximum and minimum times
for all of the combinations specified

• Repeatability
– Relevant information must be captured in problem
specification and combinations of parameters
 Problem 3 - Quality Assurance
• Specification of generic neural network software
(models and learning)

• Prototype of specification

• Comparison of a given implementation with

specification prototype

• Allows practitioners to create arbitrary neural

networks verified against models
Two Methods for Comparison
• Direct comparison of outputs:
20-10-5 (with particular connections and input)
Prototype <0.123892, 0.567442, 0.981194, 0.321438, 0.699115>
Implementation <0.123892, 0.567442, 0.981194, 0.321438, 0.699115>

• Verification of weights generated by learning

algorithm:
20-10-5 Iteration 100 Iteration 200 ………. Iteration n

Prototype Weight state 1 Weight state 2 ………. Weight state n

Implementation Weight state 1 Weight state 2 ………. Weight state n
 Further Work on
improvements
• Practitioners to use the development process or
at least document in problem specification
• Feedback from neural network development
community on the content of the problem
specification template
• Collect problem specifications and analyse to
look for commonalities in problem domains and
improve predictability (eg. control)
• More verification of specification prototype
 Further Work (2)
• Translation methods for formal specification
• Extend formal specification to new types
• Fully prove aspects of the specification
• Cross discipline data analysis methods (eg. ICA,
statistical analysis)
• Implementation of learning on distributed
systems
– Peer-to-peer network systems (farm each
combination of parameters to a peer)
• Remain unfashionable
Summary
- Neural network is a computational model that simulate
some properties of the human brain.

- The connections and nature of units determine the

behavior of a neural network.

- Perceptrons are feed-forward networks that can only

represent linearly separable functions.
 Summary

- Given enough units, any function can be represented

by Multi-layer feed-forward networks.

- Backpropagation learning works on multi-layer

feed-forward networks.

- Neural Networks are widely used in developing

artificial learning systems.
 References
- Russel, S. and P. Norvig (1995). Artificial Intelligence - A
Modern Approach. Upper Saddle River, NJ, Prentice
Hall.

- Sarle, W.S., ed. (1997), Neural Network FAQ, part 1 of 7:

Introduction, periodic posting to the Usenet newsgroup
comp.ai.neural-nets,
URL: ftp://ftp.sas.com/pub/neural/FAQ.html
Sources

Eric Wong
Eddy Li

Martin Ho

Kitty Wong