Timeline for Why is loss displayed as a parabola in mean squared error with gradient descent?
Current License: CC BY-SA 4.0
11 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 20, 2021 at 8:48 | vote | accept | Kokodoko | ||
| Apr 20, 2021 at 6:40 | answer | added | SmarArror | timeline score: 1 | |
| Apr 19, 2021 at 21:39 | comment | added | Kokodoko | I added "mean squared error" to the question for clarity. The question is still why would all the possible "wrong" loss values happen to be points on a parabola (say, 40, 80, 14... all wrong values)... That connection is not explained in most tutorials. | |
| Apr 19, 2021 at 21:38 | history | edited | Kokodoko | CC BY-SA 4.0 |
added mean squared error for clarity
|
| Apr 19, 2021 at 14:07 | comment | added | David Hoelzer | I suspect the illustration isn't meant to be taken literally; instead, I suspect the author is intending to illustrate that gradient descent attempts to solve the minimization problem by moving downward toward what is (hopefully) a global minimum on some complex surface. | |
| S Apr 19, 2021 at 13:59 | history | suggested | Amazon Dies In Darkness | CC BY-SA 4.0 |
Fixed a minor error
|
| Apr 18, 2021 at 12:14 | review | Suggested edits | |||
| S Apr 19, 2021 at 13:59 | |||||
| Apr 18, 2021 at 12:02 | comment | added | Neil Slater | @Kostya: Yes, but you probably would not draw a parabola for e.g. $\mathcal{L}(\hat{y}, y) = |\hat{y} - y|$ | |
| Apr 18, 2021 at 10:18 | comment | added | Kostya | Every minimum point can be approximated by a parabola. | |
| Apr 18, 2021 at 9:38 | comment | added | Neil Slater | Are you using a specific loss function here? Please use edit to explain which one. I expect it is MSE, and if so that should cover everything needed to answer your question | |
| Apr 18, 2021 at 9:35 | history | asked | Kokodoko | CC BY-SA 4.0 |