Scribd
Upload
9 views39 pages

13 Ann

The document discusses the evolution and principles of artificial neural networks, highlighting their historical context, structure, and learning mechanisms. It covers key concepts such as the limitations of perceptrons, the significance of hidden units, and various training methods including backpropagation and regularization. Additionally, it touches on advanced topics like spiking nets, recurrent networks, and hyper-networks, emphasizing the importance of network architecture and design decisions in AI development.

Uploaded by

TAMAL BANIK
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
9 views39 pages

13 Ann

The document discusses the evolution and principles of artificial neural networks, highlighting their historical context, structure, and learning mechanisms. It covers key concepts such as the limitations of perceptrons, the significance of hidden units, and various training methods including backpropagation and regularization. Additionally, it touches on advanced topics like spiking nets, recurrent networks, and hyper-networks, emphasizing the importance of network architecture and design decisions in AI development.

Uploaded by

TAMAL BANIK
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 39

Artificial Neural Networks

The future of AI
Restaurant Data Set
Limited Expressiveness of Perceptrons
The XOR affair
• Minsky and Papert (1969) showed
certain simple functions cannot be
represented (e.g. Boolean XOR).
Killed the field!
• Mid 80th: Non-linear Neural
Networks (Rumelhart et al. 1986)
The XOR affair
Neural Networks
• Rich history, starting in the early forties
(McCulloch and Pitts 1943).
• Two views:
– Modeling the brain
– “Just” representation of complex functions
(Continuous; contrast decision trees)
• Much progress on both fronts.
• Drawn interest from: Neuroscience,
Cognitive science, AI, Physics, Statistics, and
CS/EE.
Neuron
Neural Structure
1. Cell body; one axon (delivers output to other connect neurons); many
dendrites (provide surface area for connections from other neurons).

2. Axon is a single long fiber. 100 or more times the diameter of cell body.
Axon connects via synapses to dendrites of other cells.

3. Signals propagated via complicated electrochemical reaction.

4. Each neuron is a “threshold unit”. Neurons do nothing unless the collective


influence from all inputs reaches a threshold level.

5. Produces full-strength output. “fires”. Stimulation at some synapses


encourages neurons to fire; some discourage from firing.

6. Synapses can increase (excitatory) or decrease (inhibitory) potential (signal


Why Neural Nets?
Motivation:
Solving problems under the constraints similar to
those of the brain may lead to solutions to AI
problems that would otherwise be overlooked.
• Individual neurons operate very slowly
But the brain does complex tasks fast:  massively parallel algorithms
• Neurons are failure-prone devices
But brain is reliable anyway  distributed representations
• Neurons promote approximate matching
less brittle  learnable
Connectionist Models of Learning
Characterized by:

• A large number of very simple neuron-like processing elements.

• A large number of weighted connections between the elements.

• Highly parallel, distributed control.

• An emphasis on learning internal representations automatically.


Artificial Neurons

Activation Functions:

stept (x) = 1, if x ≥ t; otherwise 0. sign(x) = +1, if x ≥ 0; otherwise -1 sigmoid(x) = 1/(1+e-x)


Example: Perceptron
Perceptrons
Single Layer Feed Forward Neural Networks

Can be easily trained using perceptron algorithm


2-Layer Feedforward Networks
Boolean functions:
x1 o1
• Every boolean function can be
represented by network with single x2 o2
hidden layer
• But might require exponential (in number
of inputs) hidden units
Continuous functions:
• Every bounded continuous function can
be approximated with arbitrarily small
error, by network with one hidden layer
[Cybenko 1989; Hornik et al. 1989]
xN oO
Any function can be approximated to
arbitrary accuracy by a network with two
hidden layers [Cybenko 1988].
Multi-Layer Nets
• Fully connected, two layer, feedforward
Activation function: g(x) = (1 if greater than threshold, 0 otherwise)

Jonathan

Mary

Joe

Elizabeth

Alice

Bart How are Mary and Elizabeth related?


A=Acquaintances B=Family
Multi-Layer Nets
• Fully connected, two layer, feedforward
Ofer Melnik, http://www.demo.cs.brandeis.edu/pr/DIBA
Ofer Melnik, http://www.demo.cs.brandeis.edu/pr/DIBA
Ofer Melnik, http://www.demo.cs.brandeis.edu/pr/DIBA
How can we train perceptrons?
Hebbian learning
• D. O. Hebb:
– The general idea is an old one, that any two cells or systems of
cells that are repeatedly active at the same time will tend to
become 'associated', so that activity in one facilitates activity in
the other." (Hebb 1949, p. 70)
– "When one cell repeatedly assists in firing another, the axon of
the first cell develops synaptic knobs (or enlarges them if they
already exist) in contact with the soma of the second cell."
(Hebb 1949, p. 63)
• Cells that fire together, wire together
– If error is small, increase magnitude of connections that
contributed.
– If error is large, decrease magnitude of connections that
contributed.
Backpropagation
• Classical measure of error
– Sum of square errors
– hw(x) is output on perceptron on x.
• Gradient decent using partial derivatives

• Update weights
Backpropagation Training (Overview)
Training data:
– (x1,y1),…, (xn,yn), with target labels yz {0,1}
Optimization Problem (single output neuron):
– Variables: network weights wij
– Obj.:E=minw∑z=1..n(yz–o(xz))2,
– Constraints: none
Algorithm: local search via gradient descent.
• Randomly initialize weights.
• Until performance is satisfactory,
– Compute partial derivatives ( E /  wi j) of objective
function E for each weight wi j
– Update each weight by wi j à wi j +  ( E /  wi j)
Smooth and Differentiable Threshold Function

• Replace sign function by a differentiable


activation function
 sigmoid function:
Slope of Sigmoid Function
Backpropagation Training (Detail)
• Input: training data (x1,y1),…, (xn,yn), learning rate parameter α.
• Initialize weights.
• Until performance is satisfactory
– For each training instance,
• Compute the resulting output
• Compute βz = (yz – oz) for nodes in the output layer
• Compute βj = ∑k wjk ok (1 – ok) βk for all other nodes.
• Compute weight changes for all weights using
∆wi j(l) = oi oj (1 – oj) βj
– Add up weight changes for all training instances, and
update the weights accordingly.
wi,j ← wi,j + α ∑l ∆wi,j(l)
Summary: Hidden Units
• Hidden units are nodes that are situated between the input nodes
and the output nodes.

• Hidden units allow a network to learn non-linear functions.

• Hidden units allow the network to represent combinations of the


input features.

• Given too many hidden units, a neural net will simply memorize the
input patterns (overfitting).

• Given too few hidden units, the network may not be able to
represent all of the necessary generalizations (underfitting).
How long should you train the net?

A B C D E

When would you stop training?


How long should you train the net?
• The goal is to achieve a balance between correct
responses for the training patterns and correct responses
for new patterns.
– That is, a balance between memorization and generalization)

• If you train the net for too long, then you run the risk of
overfitting.
– Select number of training iterations via cross-validation on a
holdout set.
Regularization
• Simpler models are better
• NN with smaller/fewer weights are better
– Add penalty to total sum of absolute weights
– Pareto optimize
Design Decisions
• Choice of learning rate 
• Stopping criterion – when should training stop?
• Network architecture
– How many hidden layers? How many hidden units
per layer?
– How should the units be connected? (Fully? Partial?
Use domain knowledge?)
• How many restarts (local optima) of search to
find good optimum of objective function?
Spiking Nets
• Represent continues values using rates
– Output spike if # of incoming spikes > threshold
– Leaky counter

http://www.ine-news.org
Spiking

From http://www.cs.uu.nl/research/techreps/repo/CS-2003/2003-008.pdf
Recurrent networks
• Nodes connect
– Laterally
– Backwards,
– To themselves
• Complex behavior
– Dynamics, Memory

www.stowa-nn.ihe.nl/ANN.htm
Learning Network Topology
• Optimal Brain Damage algorithm
– Trains a fully connected network
– Removes connections and nodes that contribute least
to the performance
• Using information-theoretic criteria
– Repeats until performance starts decreasing
• Tiling algorithm: Grows networks
– Start with a small network that classifies many
examples
– Repeatedly add more nodes to classify remaining
examples
Hyper-Networks
• Use a network to generate a network
– E.g. to determine connection wij use network that
takes in i and j and produces w.
– In 2D:

Ken Stanley, eplex.cs.ucf.edu


Hyper-Networks

Ken Stanley, eplex.cs.ucf.edu

You might also like