We need to disambiguate meta learning. How about leave that for the arch. search and use hyper learning/net for the dynamic/model as output?
Show this thread
There is no clear distinction between hyperparameter tuning and architecture search... it's a continuum, in theory and in practice
-
-
I agree! But networks that output networks at execution time is almost something else IMHO...
-
Some brand more trivial work as meta-learning to get more recognition/citations etc. Writing about hyperparam tuning is important but unsexy
End of conversation
New conversation -
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.