Timeline for Error increase on L2 regularization in an NN
Current License: CC BY-SA 3.0
14 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 1, 2018 at 22:03 | comment | added | Prasqui | @JanKukacka No the weight, but the error go to zero, so the penalty term become the most relevant in the formula. Anyway the answer of hxd1011 was very simple and clear, but since usually no one mentions it, i was a bit confused. | |
| Feb 1, 2018 at 2:01 | history | tweeted | twitter.com/StackStats/status/958882875964837889 | ||
| S Jan 31, 2018 at 21:41 | history | suggested | smci | CC BY-SA 3.0 |
clarification
|
| Jan 31, 2018 at 21:38 | review | Suggested edits | |||
| S Jan 31, 2018 at 21:41 | |||||
| S Jan 31, 2018 at 19:14 | history | edited | gung - Reinstate Monica | CC BY-SA 3.0 |
corrected spelling
|
| S Jan 31, 2018 at 19:14 | history | suggested | StatsSorceress | CC BY-SA 3.0 |
corrected spelling
|
| Jan 31, 2018 at 18:56 | review | Suggested edits | |||
| S Jan 31, 2018 at 19:14 | |||||
| Jan 31, 2018 at 18:06 | comment | added | Vladislavs Dovgalecs | You want to trade some error on the training set for a lower error on validation set. You don't want to overfit on the training set. I suggest to do a bit reading on how training, validation and test errors are related and controlled. | |
| Jan 31, 2018 at 18:06 | comment | added | Jan Kukacka | What exactly do you mean by lambda*w? If the weights go to zero, also the L2 penalty term becomes smaller. | |
| Jan 31, 2018 at 18:03 | answer | added | Jan Kukacka | timeline score: 6 | |
| Jan 31, 2018 at 17:52 | vote | accept | Prasqui | ||
| Jan 31, 2018 at 17:31 | answer | added | Haitao Du | timeline score: 12 | |
| Jan 31, 2018 at 17:27 | review | First posts | |||
| Jan 31, 2018 at 19:03 | |||||
| Jan 31, 2018 at 17:26 | history | asked | Prasqui | CC BY-SA 3.0 |