On Sun, Dec 2, 2018 at 10:54 PM Ben Goertzel <[email protected]> wrote:
> Nil implemented a diversity measure in the Aidyia period and one can tune > the weight of diversity vs accuracy in the fitness function ..Not sure if > this does what u need... > Ah yes, that's right. There are more than a few ways to measure diversity; being who I am, I recall thinking that there should be other, better ways of measuring diversity, but never found the time to try others. The correct architecture would be to make the diversity-measure be a plug-in module, so you could code up and try different measures of diversity. --linas > > On Mon, 3 Dec 2018, 11:33 Linas Vepstas <[email protected] wrote: > >> >> >> On Sun, Dec 2, 2018 at 2:05 PM Michael Duncan <[email protected]> >> wrote: >> >>> can any moses gurus comment on using the diversity scoring options? for >>> some bioinformatics work we are more interested in evolving a minimal >>> ensemble of models to maximally represent the patterns distinguishing two >>> sample sets rather than just maximizing out of sample prediction accuracy. >>> in other words, we want to maximize the number of unique features in a >>> model ensemble of a given size and accuracy. more generally, are there >>> procedures for choosing optimal ensemble models beyond combining the top n >>> models from different cross-validation runs? >>> >> >> Wow. At one point, I wanted to add something like that as a feature: >> Namely, when keeping a deme (ensemble) of some finite size, I wanted it to >> be as diverse as possible (while still having acceptable fitness). I was >> excited, ready to implement this, but could not quite figure out a good way >> of measuring how "different" two models were. What constitutes diversity, >> and how do you measure it? Without that, I couldn't move forward. >> > -- cassette tapes - analog TV - film cameras - you -- You received this message because you are subscribed to the Google Groups "opencog" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/opencog. To view this discussion on the web visit https://groups.google.com/d/msgid/opencog/CAHrUA34-b_%3Di0m4z86qR3CAezMA8xmFTGZPxqkhXZec4kDSkuQ%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
