Skip to main content
14 events
when toggle format what by license comment
Feb 26 at 22:19 comment added user522465 @AlirezaKhayatian: I changed the variable names and my solution still applies to E(Y).
Feb 26 at 22:17 history edited user522465 CC BY-SA 4.0
change variable names
Feb 26 at 20:33 comment added Alireza Khayatian Thank you for your efforts, @Damien. You defined two parameters, $Y = \sup_{X \in \mathcal{X}_k} |\langle Z, X \rangle|^2$ and $Y(X) = \sum_{i=1}^{n} w_i x_i $ that may be confusing. your solution is for $E\left[\sup_{X \in \mathcal{X}_k} Y(X)\right] $ but I am looking for $E(Y)$. .
Feb 26 at 0:33 comment added user522465 @AlirezaKhayatian: I updated the solution.
Feb 26 at 0:20 history edited user522465 CC BY-SA 4.0
Added a more precise proof
Feb 22 at 10:33 comment added Alireza Khayatian Thank you for your efforts, @Damien. But, we are looking for $Y = \max_{\mathbf{X} \in \mathcal{X}_k} \| \mathbf{Z}^T \mathbf{X} \|_2^2$. I mean the squared version of what you have considered. It may be related to the squared version of the Gaussian width.
Feb 21 at 3:16 comment added user522465 @AlirezaKhayatian: I updates the solution
Feb 21 at 3:13 history edited user522465 CC BY-SA 4.0
Added a more consise proof
Feb 20 at 12:05 comment added Alireza Khayatian I appreciate your effort, but we are looking for $Y = \max_{X \in X_k} |Z^TX|^2$. I mean, the problem is about top-k selection, not random-k selection.
Feb 20 at 7:44 history edited user522465 CC BY-SA 4.0
update typo
S Feb 20 at 7:07 review First answers
Feb 20 at 8:49
S Feb 20 at 7:07 history edited user522465 CC BY-SA 4.0
remove unecessary lines
S Feb 20 at 5:21 review First answers
Feb 20 at 5:33
S Feb 20 at 5:21 history answered user522465 CC BY-SA 4.0