|
Members in Shoutbox
None.
Shoutbox Search
Shoutbox Commands
/w [name] > Whisper
/r > Reply to last whisper /me > Marks as action Shoutbox Information
Moderators may delete any and all shouts at will.
|
Global Shoutbox
Please log in to shout.
[2017-4-05. : 2:27 am]  lil-Inferno -- Dem0nDem0n shouted: Mini Moose 2707 Great, so we're gonna get something that's rushed and shitty LMAO pantaloons in a bunch[2017-4-05. : 2:08 am]  Corbo -- Dem0nDem0n shouted: Mini Moose 2707 Great, so we're gonna get something that's rushed and shitty LMAO and two decades later[2017-4-05. : 2:07 am] Dem0n -- Mini Moose 2707Mini Moose 2707 shouted: That's what they said about the remaster! Great, so we're gonna get something that's rushed and shitty LMAO[2017-4-05. : 1:54 am] Dem0n -- Moose, where's this special meme that Ex was talking about with the staff creating something special for us?[2017-4-05. : 1:20 am] Excalibur -- Brb hitting a page on the site through a google search to put speed back to 56k nostalgia levels.[2017-4-05. : 1:17 am]  Moose -- I didn't do anything. We're under less load and nothing is hitting the terrible parts of the code LMAO[2017-4-05. : 1:16 am] Corbo -- I actually edited my post in your site update to take into account sen was faster[2017-4-04. : 11:19 pm] lil-Inferno -- I can't wait to get on SC so we can sit in a channel and be like "w/e" "tru" "ok" "fite me" "what map?" "idk" "tru"[2017-4-04. : 11:10 pm] lil-Inferno -- Voyager7456Voyager7456 shouted: lil-Inferno shut it, nerd irritated[2017-4-04. : 11:08 pm] Voyager7456 -- lil-Infernolil-Inferno shouted: I never heard people say "shut up" anymore. I don't know if that means I surround myself with positive people that wanna hear what others have to say more, or people just got more apathetic and now they're like "w/e" lmao shut it, nerd[2017-4-04. : 11:08 pm] lil-Inferno -- I never heard people say "shut up" anymore. I don't know if that means I surround myself with positive people that wanna hear what others have to say more, or people just got more apathetic and now they're like "w/e" lmao[2017-4-04. : 9:48 pm] TiKels -- TiKelsTiKels shouted: There should never be a local maximum anywhere other than at the edge ... if I'm thinking about this the right way using your terminology I think minimum was more appropriate. Extrema in general, however should suffice[2017-4-04. : 9:46 pm] Vrael -- it depends on the goal of your problem really, if you have some sort of more closed form solution to your problem with specific target metrics then by all means use them[2017-4-04. : 9:45 pm] TiKels -- Wouldn't I be applying an arbitrary weight to it then? Like oh, try and converge to about this accuracy, effectively?[2017-4-04. : 9:44 pm] Vrael -- so if you have some total accuracy function F, instead of merely minimizing F you minimize F+C, where C represents some cost function or complexity penalty[2017-4-04. : 9:44 pm] TiKels -- There should never be a local maximum anywhere other than at the edge ... if I'm thinking about this the right way[2017-4-04. : 9:44 pm] Vrael -- the more general idea behind this is that you introduce some sort of 'cost function' to your problem/equations, where you can specify the relationships between your variables or other parameters[2017-4-04. : 9:42 pm] Vrael -- if it happens to have a convex global minimum, like in linear regression, there might be a closed form solution as well (like in linear regression)[2017-4-04. : 9:42 pm] Vrael -- you can still use gradient descent if its not true, but you'll want to randomly initialize it and test it out a bunch of times to ensure you don't get stuck in a local minima[2017-4-04. : 9:41 pm] Vrael -- if the problem 'happens' to have a convex global minimum you can use an algorithm like gradient descent |
