01.11.2014 Views

MACHINE LEARNING TECHNIQUES - LASA

MACHINE LEARNING TECHNIQUES - LASA

MACHINE LEARNING TECHNIQUES - LASA

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

124<br />

6.3 Perceptron<br />

The earliest neural network models go back to 1940 with the McCulloch & Pitts perceptron. The<br />

perceptron consists of a simple threshold unit taking binary inputs X = { x x }<br />

r<br />

1 ,..., n<br />

. Its output y is<br />

the product of its entries multiplied by a (1-dim) weight matrix W, modulated by a transfer function<br />

or an activity function f.<br />

n<br />

⎛⎛ ⎞⎞<br />

y= f ⎜⎜ w ⋅ x = f W⋅X<br />

⎝⎝ i=<br />

0 ⎠⎠<br />

∑ i i⎟⎟<br />

( )<br />

(6.1)<br />

where wxis the bias and is generally negative.<br />

0 0<br />

Figure 6-1: Perceptron with constant bias<br />

Classical transfer functions are:<br />

The linear function f( x)<br />

The step function<br />

= x<br />

⎧⎧1 if x ≥ 0<br />

f( x)<br />

= ⎨⎨<br />

⎩⎩0<br />

if x < 0<br />

© A.G.Billard 2004 – Last Update March 2011

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!