Timeline for Why they use KL divergence in Natural gradient?
Current License: CC BY-SA 4.0
3 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 31, 2019 at 8:35 | comment | added | Phizaz | It is true if the assumption is that the neural network is a probability function. However, the assumption is arbitrary. I could do equally well if I assume that the neural net is a deterministic function. The square distance makes much more sense now. | |
| Oct 30, 2019 at 19:45 | review | First posts | |||
| Oct 30, 2019 at 20:46 | |||||
| Oct 30, 2019 at 19:43 | history | answered | Srivatsan Ramesh | CC BY-SA 4.0 |