Negating an XOR topology by manipulating weights and thresholds?

See more:

artificial-intelligence

, +

Hello,

If you are given a neural network having an XOR topology:

x1 [A]   weight(A,C)=1           weight(A,D)=1
                          [C]          weight (C,D)=-2                [D]
x2 [B]  weight(B,C)=1            weight(B,D)=1
                    T=1.5                                      T=0.5

Now XOR produces this truth table: T+T=F; T+F=T; F+T=T; F+F=F;

I am not allowed to change the network topology.
How should the weights and thresholds change to produce the negation of XOR? i.e T+T=T; T+F=F; F+T=F; F+F=T;

Do i just negate the weights and interchange the thresholds?