Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

vanilla bp

2 views
Skip to first unread message

Caren Marzban

unread,
Dec 2, 2002, 8:02:17 PM12/2/02
to

Hi,

I'm finding that in standard backprop, the overall magnitude of the
weights drops as a power-law with the number of training cycles. I.e.

w ~ NCYCLE ^ -alpha

where alpha is some real number (depending on data, etc.). Is there
anything in updating of the weights that causes them to shrink with
NCYCLE?

Caren

--
http://www.nhn.ou.edu/~marzban

0 new messages