0
$\begingroup$

I am reading this paper and came across an equation I don't really understand. I learned the maximum likelihood method last semester.

However the functions that I saw always contained the parameter $\Theta$. Then I took the log of the function, derived it and solved for $\Theta$.

I don't understand what parameter they are trying to find. This is the equation (it is on page 5 of the paper):

enter image description here

$\endgroup$
  • 1
    $\begingroup$ The paper explains the notation a few paragraphs before the equation. $\hat y$ depends on parameters $W_l$ and $b_l$, which they denote collectively as $\theta$. $\endgroup$ – Jirapat Samranvedhya Oct 13 '17 at 7:16
1
$\begingroup$

According to the paper ${\hat y}_i$ and ${\hat \sigma}$ belong to a class of functions parameterized by $\theta$. So, $\theta$ is the vector of all parameters in ${\hat y}_i$ and ${\hat \sigma}$. For instance, if you consider a deep NN, $\theta$ contains all of the elements of $W_i$ and $b_i$ ($i \in \{1,2,\cdots,L\}$) in the expression above your equation.

$\endgroup$

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.