Timeline for convergence in distribution of stochastic gradient descent.
Current License: CC BY-SA 3.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 27, 2011 at 12:59 | answer | added | Johannes L | timeline score: 1 | |
| May 29, 2011 at 12:11 | comment | added | vedarun | @peter Sarkoci you are right that the question is about the behaviour of a randomized algorithm. However, i would want the minimizer to be a non-trivial random variable rather than a constant. Any r.v with expected value as the actual minimizer would do. For instance, a Gaussian with mean at the minimizer and low variance would be fine. I think for such a condition to be satisfied depends only on the possible values of the gradients of the original function and not on the choice of step sizes. Am i right? | |
| May 28, 2011 at 19:30 | comment | added | Peter Sarkoci | I do not understand your question very well. What you asked seems to be a question about behavior of certain randomized minimization algorithm. Am I right? As such, one would expect the minimizer to be a constant value (rather than a nontrivial r.v.); am I still right? If so, are you aware of the fact that, in such case the convergence in distribution turns up into convergence in probability? | |
| May 24, 2011 at 19:45 | comment | added | fedja | I'm not sure what exactly is asked. Are you looking for an example when the convergence in distribution holds and the a.s. one fails, or what? | |
| May 24, 2011 at 12:12 | history | edited | Theo Buehler | CC BY-SA 3.0 |
edited title
|
| May 24, 2011 at 9:47 | history | asked | Vedarun | CC BY-SA 3.0 |