I want to show directly from the definition of exp(X) that it is Frechet differentiable at every $X \in \mathbb{R^{n\times n}}$. The definition being exp$(X):= \sum \frac{X^k}{k!}$

I know continuous partial derivatives implies the function is frechet differentiable. However, this was proven in my notes in the the context of functions from an n-dimensional Euclidean space to an m-dimensional one and so I'm not confident in extending it to matrix functions (matrix-valued). Can I take partials as: $\frac{\partial(exp(X))}{X_{ij}}$ where the denominator is the i-jth entry of the matrix $X$. Then show that these are continuous and therefore exp$(X)$ is differentiable at $X$? If yes, then this 'object' is a matrix is it not? And so is continuous iff each of it's entries are continuous? And am I correct in thinking $\frac{\partial(X)}{X_{ij}}$ is the matrix with a $1$ in the i-jth position and $0$ elsewhere? I could then use this to show (by product rule?) the higher order terms in the sum are partially differentiable w.r.t $X_{ij}$ and hence so is their sum.

Is the Jacobian matrix of derivative of exp$(X)$ an $n^2 \times n^2$ matrix? Any comments are welcome, just want someone to point out mistakes or perhaps add something you might suspect I don't quite understand. Thank you.


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.