11/24/03
12
Goal of Perceptron Learning
¥Suppose we have training patterns x1, x2, É, xP with corresponding desired outputs y1, y2, É, yP 
¥where xp ë {0, 1}n, yp ë {0, 1}
¥We want to find w, q such that
yp = Q(w×xp Ð q) for p = 1, É, P