Alternative ways to write the gradient

189 Views Asked by At

This may not be the right forum to ask this question, but suppose that I have a multi-dimensional function $L$, and I want to compute its gradient w.r.t. a set of parameters $\theta$, where $[|\theta| > 1]$. I saw in several texts that they write it down as:

$$ \nabla_{\theta}L $$

However, $\nabla_{\theta}L$ occupies too much page width, i.e. if $\theta$ includes other qualifications such as $\theta_{i,t}$. Since am writing using a two column paper format, I want to conserve on text width, but I can take up more spaces height-wise. Is it correct to write the gradient this way?

$$ \frac{\nabla_{L}}{\nabla{\theta}} $$

or does it make sense if I stick to using partial derivative notations, i.e.

$$ \frac{\partial{L}}{\partial{\theta}} $$

1

There are 1 best solutions below

0
On BEST ANSWER

You write it as $$\frac{\partial L}{\partial \theta},$$ because that's what the $\theta$ component of the gradient means, by definition.