Timeline for Computing gradients via Gaussian Process Regression
Current License: CC BY-SA 4.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 18 at 3:50 | answer | added | Sudipto Banerjee | timeline score: 0 | |
| Nov 14, 2020 at 10:57 | answer | added | David Brandes | timeline score: 2 | |
| Sep 5, 2020 at 13:37 | answer | added | MichalK | timeline score: 2 | |
| Jul 28, 2020 at 16:34 | answer | added | kuberry | timeline score: 3 | |
| Oct 27, 2018 at 20:18 | vote | accept | Mathews24 | ||
| Oct 27, 2018 at 20:13 | history | edited | user20160 |
[Edit removed during grace period]; edited tags
|
|
| Oct 27, 2018 at 11:00 | answer | added | user20160 | timeline score: 19 | |
| Oct 25, 2018 at 12:20 | comment | added | Mathews24 | @gg Awesome, that's almost exactly what I was searching for. If that can be supplemented with a simple example code implementation (e.g. computing derivatives of a GP with SE kernel on some random data set), that would answer my question. | |
| Oct 24, 2018 at 20:36 | comment | added | g g | Ah and I forgot here you find more theory: link.springer.com/chapter/10.1007/978-3-642-21738-8_20 | |
| Oct 24, 2018 at 20:24 | comment | added | g g | You should not use finite differences because there are better ways to do this. If your kernel is twice differentiable then your GP is differentiable. Have a look at this question and its answer: stats.stackexchange.com/questions/180823/… | |
| Oct 24, 2018 at 7:40 | comment | added | Frans Rodenburg | Is a 'Gaussian Process Regression' with a posterior mean any different from Bayesian linear regression? You might attract more useful replies if you use simpler language. | |
| Oct 24, 2018 at 0:35 | history | asked | Mathews24 | CC BY-SA 4.0 |