Hyperparameter tuning magicpic.twitter.com/t7X1FLdjXQ
Show this thread
Note that the search space can be dynamic: here you have a 'for' loop controlled by a hyperparameter, inside which you define new hyperparameters. You could push this to any level (e.g. searching over a connectivity graph for a dynamic number of nodes). Even recursion...
-
-
Need AI for hyperparameters...
-
This Tweet is unavailable.
- Show replies
New conversation -
-
-
Would you say that this is a graph search algorithm? Does it fit some sort of function graph→number and finds an optimal graph?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Can it be installed via pip or has to be built by source?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
And adding hyperparameters on top of hyperparameters on top of hyperparameters, Ptolemaic model managed to survive to Keplero-Galileo. But indeed it was not explaining the real world. Kind of ante litteram overtraining/overtuning. Occam Razor stands.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.