Closed pansanity666 closed 1 year ago
Thank you for your interest in our work! In A+-B
, B
is actually 2*std
which is the 95% confidence interval. This was calculated using multiple testing runs of the same trained model.
Thank you for your interest in our work! In
A+-B
,B
is actually2*std
which is the 95% confidence interval. This was calculated using multiple testing runs of the same trained model.
Thanks for reply. So, is it a common practice that A+-B
represents 95% confidence interval? Should I clarify that I am using 95% instead of 68% etc when using this formulation? (I am not familiar with machine learning, and it seems not commonly used in deep learning papers.)
Best,
yea many papers today don't report confidence interval, maybe it's better to add some clarification (I'll add it to the preprint)
That will be nice!
Hi, I have a naive question about Tab. 1 in the paper. Does
A+-B
representmean+-std
, which means you run the experiments for multiple times and report the mean and std value?Best,