Derivative of eigenvalues and eigenvectors of Hermitian Matrix by automatic differentiation

Once you assume a normalization, you can define derivatives. The simplest way is to use implicit differentiation and add the normalization to the constraints. This can be tedious, but I find the Matrix Cookbook and Old and new matrix algebra useful for statistics helpful.

I have worked out a subcase of SVD, I am planning to make a PR when I make it work generally.

1 Like