Turned up to eleven: Fair and Balanced 
Front page
Random thoughts, sometimes deep, mostly not, about politics, war, science, religion, life in general
by Paul Orwin

Wednesday, May 15, 2002
Posted
11:20 PM
by Paul
In order to multiply a vector V by a Matrix A, the dimension of the vector n has to be equivalent the dimension of the matrix A nxn (only square matrices are used, I think). In any case, the multiplication is done A*V=V', where entry V'(i)=Sum(Ai*Vi) as i goes from 1 to n. I had it backwards in the original statement. Using this mechanism, the ith entry in V' is the input value into neuron i, and the thresholding rule can be applied. Just a math error, but it makes me hang my head in shame!
