Timeline for What are the differences between sparse coding and autoencoder?
Current License: CC BY-SA 3.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 10, 2015 at 18:37 | comment | added | RockTheStar | Sorry to bring this question up again. I think Autoencoder can encourage sparsity as well, i.e. sparse autoencoder. | |
| Dec 2, 2014 at 2:30 | vote | accept | RockTheStar | ||
| Oct 12, 2014 at 18:07 | comment | added | bayerj | Having a fixed map that has been estimated in order to follow some constraint (here: sparse result) is less powerful than having an optimiser that tries to find a solution like that over possibly many iterations. | |
| Oct 10, 2014 at 19:44 | comment | added | RockTheStar | Sorry, can you elaborate on this? | |
| Oct 10, 2014 at 8:18 | comment | added | bayerj | There is no general rule like that. SC has one benefit over AEs: encoding via optimisation can be very powerful. | |
| Oct 9, 2014 at 22:02 | comment | added | RockTheStar | Thanks for your answer! So is that mean sparse coding should not be used in any case, but autoencoder? Also, there should be an extra term in sparse coding which regularize W? | |
| Oct 9, 2014 at 18:51 | history | answered | bayerj | CC BY-SA 3.0 |