Let L(w; x; y) be a convex loss function for a linear classifier w. Can you always express the gradient of L w.r.t w as f(y; wx)*x? I.e, is the gradient always some scalar function f of the gold label y and the prediction wx, times x? This seems to hold for hinge loss, square loss and logistic loss.


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.