Question about the CCA protocol

[PeiqinZhuang]

Hi, guys. I am also using CCA algorithm to perform dimension reduction, i.e. using the api from sklearn. However, this operation may require more than a whole day to achieve that goal, from 2048-dims to 1048-dims. So I wonder if there is another solution to accelerate the procedure of this operation?

[nhynes]

What kind of hardware are you using? How big is the dataset? In either case, it might be helpful to subsample the data if you have a lot of it; the linear model won't have enough capacity to model all of the data. Also, 1048 is still a lot of dimensions, which suggests the use of PCA. You'll have an easier time doing retrieval when the space isn't so large.

[PeiqinZhuang]

Hi, nhynes. Thank you for your reply. If possible, could I inquriy you some details about the baseline protocols through email. Previously, I sent an enquiry email to the address(nhynes at mit dot edu), but, unfortunately, the email was rejected.

[nhynes]

I've since graduated so that email is no longer fresh. If you have questions that you don't want to ask on github, please feel free to resend to the.nhynes@<popular, definitely evil mail provider>.com