[Amdnl] Data deletion is "not a fault"?

[Juyang Weng]

Dear Alexandre and Daniela, ICDL 2022 Program Chairs:

Thank you very much for doing the program chair for a conference in which
some people apparently have lost the direction of "development".

I would like to bring this important event to your attention:
Reviewer 491 wrote "I find the point that CNNs train multiple networks not
a fault but a feature --- BECAUSE typically and scientifically
appropriately the performance also of the average, median and standard
deviations is typically reported as well."

The reviewer is obviously out of touch with what Prof. Geoffrey Hinton and
his coworkers have done in their papers (e.g., in Communications of ACM).
They reported the "luckiest network" only and deleted data from all other
trained networks. They reported no average, no median, no standard
deviations (across trained networks).  Please forward this email to
reviewer 491 since he missed "data deletion".

By the way, the EIC of Communications of ACM (CCed) has decided "to do
nothing".

In order to raise attention, I feel it is necessary to let all those on the
CCed lists know.

Best regards,
-John
-- 
Juyang (John) Weng
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cse.msu.edu/pipermail/amdnl/attachments/20220614/467c2521/attachment.html>