You are on page 1of 1

NN&GA Exam Subjects

One problem similar with:


1. Consider a neuron with two binary inputs x1 , x2 1, 1 and one binary output y 1, 1
that has the weights w1 1, w2 1 and an activation bipolar step function
1, n 0
y
,
1, n 0
where n w1 x1 w2 x 2 b is the internal activation potential (the net input) of the neuron.
1.1 What logical function models this neuron for a bias b = 1?
1.2 What is the range of the bias b for which the logic function modeled by the neuron remains the
same if the input signals can be in the range 0.9 1.1 for the logical value TRUE, and in
range 1.1 0.9 for FALSE, while the incertitude (noise) range for the internal activation
potential is n 0.1 0.1 ?
1.3 Find out the weighs and the bias for a unipolar neuron, x1 , x 2 , y 0, 1 that models the same
function as the one from point 1.1.
2. A problem of implementation

2 or 3 theory subjects from:


1.
2.
3.
4.
5.
6.

Perceptrons with perceptron training rule


Feed-forward NNs with the backpropagation algorithm
Linear Networks with Adaline algorithm
Self organizing NNs with competitive learning algorithm
Hopfield NNs with training
Genetic Algorithms with training

You might also like