What effect does linearly projecting a state space to another dimension have on the Jacobian?

26 Views Asked by At

Suppose I have a real vector $x \in \mathbb{R}^n$ and a real, nonlinear vector-valued function $\sigma(x) \in \mathbb{R}^n$. I denote the matrix Jacobian

$$\partial_x \sigma(x)$$

Suppose I have a linear transformation $P: \mathbb{R}^n \rightarrow \mathbb{R}^k$. What is the Jacobian of the projected state space with respect to the projected input? In other words, what is

$$\partial_{Px} P\sigma(x)$$

How would I go about calculating this?