I am new to vector calculus and even though I was able to understand how to derive simpler functions like the product of two vectors in $\mathbb{R}^m$, I am confused about this specific scenario.

Lets say we have a column vector $x \in \mathbb{R}^m, A \in \mathbb{R}^{m\times m}$.

How can I determine the gradient of $x^TA$ with respect to $x$?

I am not able to grasp how you derive something with respect to a vector.

From what I can see $x^TA$ is equal to $[A_1 \cdot x$, $A_2 \cdot x$, ..., $A_m \cdot x]$ for all columns of $A$. Which means the operation yields a row vector in $\mathbb{R}^m$.

I tried to derive this solution with respect to the individual scalars of $x$ but I am lost since I am clearly not approaching it correctly.

How would one solve this?

  • $\begingroup$ Do all partial derivatives of all components of the vector function you just defined, no? $\endgroup$ – ir7 Feb 9 '18 at 13:09
  • $\begingroup$ Also, look up Jacobian, as generalization of gradient. $\endgroup$ – ir7 Feb 9 '18 at 13:23

If you now look at $x^TA$ as a linear map $\mathbb{R}^m\to \mathbb{R}^m$ as $$f(x_1,\ldots, x_m)=(\sum_{i=1}^m a_{i1}x_i,\ldots,\sum_{i=1}^m a_{im}x_i)$$ so we don't have to think about $f$ in columns and rows or matrices, we know that the gradient of the linear map is written as $$\frac{\partial f}{\partial x}=\left(\begin{array} \ a_{11} & a_{21} & \dots & a_{m1} \\ a_{12} & a_{22} & \dots & a_{m2} \\ \vdots & \vdots & \ddots & \vdots \\ a_{1m} & a_{2m} & \dots & a_{mm} \end{array}\right)=A^T$$ by convention of derivatives are from left to right and $f$-elements from the top down.

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.