I am taking a course in Machine Learning and the Professor introduced us to the XOR problem.

I understand the XOR problem is not linearly separable and we need to employ Neural Network for this problem.

However, he mentioned XOR works better with Bipolar representation(-1, +1) which I have not really understand.

I am wondering what Bipolar representation would be better than Binary Representation? Whats the rationale for saying so?

  • $\begingroup$ If bipolar means values +1 and -1 only, this scheme is totally equivalent to {0, +1}, being the mapping (x+1)/2. The NN can handle both, even doing the mapping at a first layer with bias. Some activation functions could be more valid in one case or another thinking in learning performance. $\endgroup$ – pasaba por aqui Oct 21 at 9:07
  • $\begingroup$ xor is not separable with a single hyperplane, but it is with two. There are lots of methods different to NN able to handle this subject. $\endgroup$ – pasaba por aqui Oct 21 at 9:07
  • $\begingroup$ But why Bipolar representation works better for this problem? $\endgroup$ – Exploring Oct 21 at 9:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.