People don't put enough effort into their baselines. You should use a solid library and do hyperparameter tuning. http://m-mitchell.com/NAACL-2016/SemEval/pdf/SemEval147.pdf …
-
"The importance of designing your Neural Networks right".pic.twitter.com/51ZDLcWEOL
Many times I have come across papers showing baselines for LSTM that are below that the of the default Keras LSTM on the same dataset.
-
-
i always like those google papers "we just tried the simple approach, but with larger nets, works better than funny architectures"
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
by putting weak baselines, we can always keep reporting "improvements". Forever.
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.