Hacker News new | ask | show | jobs
by microtonal 4 days ago
It doesn't though. Wx + b is an affine transformation, which is just a linear transformation + translation: https://en.wikipedia.org/wiki/Affine_transformation
1 comments

Yes, thanks for the correction. The opening line was not meant as an excerpt. It was describing the neuron’s weighted sum plus bias, not softmax.

Activation functions give non-linearity. I have fixed it. I believe the rest was correct.