Timeline for Counterexamples to differentiation under integral sign, revisited
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 11, 2022 at 19:51 | comment | added | Jiaji Huang | Thank you both! I guess the original question stems from here. In fact, I was trying to make sense of this question for kernel approximation in machine learning: Let $k(x, y)$ be a kernel. It is often possible to construct (random) feature maps: $\phi(x; Z)$, $\phi(y; Z)$ such that $k(x, y) = \mathbb E_Z[\phi(x; Z)^\top \phi(y; Z)]$. Here $Z$ is some random variable. But, does the following necessarily hold? ${\partial k(x, y) \over\partial x} = \mathbb E_Z [{\phi(x; Z)^\top \phi(y; Z) \over \partial x}]$ | |
| Dec 11, 2022 at 16:03 | comment | added | Iosif Pinelis | @DanieleTampieri : Thank you very much for your kind words and for your help. | |
| Dec 11, 2022 at 14:32 | comment | added | Daniele Tampieri | You are welcome, Iosif: I follow almost all your Q&A on the MathOverflow, and I like when I can improve a bit such high quality writings. | |
| Dec 11, 2022 at 13:30 | comment | added | Iosif Pinelis | @DanieleTampieri : Thank you for your edit and the useful information about it. | |
| Dec 11, 2022 at 6:54 | history | edited | Daniele Tampieri | CC BY-SA 4.0 |
Formula hyperlinking (despite not being shown correctly in the preview, links work correctly on the whole page, for all questions and answers).
|
| Dec 11, 2022 at 3:50 | history | answered | Iosif Pinelis | CC BY-SA 4.0 |