Personalized TV on single-board computers
In contrast to kernel SVMs, linear SVMs are pretty versatile and lightweight. The required memory is usually neglictable and the cost function is convex, in case of the squared hinge loss. 502 more words
5 hours, 12 minutes
The Non-Canon Awards
We are on a new author. One of the first I read when I signed up on WordPress. I think I actually signed up for her. 843 more words
12 hours, 26 minutes
Well, I’m still alive.
I’ve been here a week now and for some reason I’ve kept going. I still don’t know what that reason is however. 3,455 more words
4 days, 13 hours
I fell behind, but you guys don’t know that. Shhh!!!
I know the post for the interview has posted, and I am thrilled for you guys to get to know this awesome and sweet woman!!! 672 more words
The last post was about benefits of kernel machines to classify sparse data with lots of rare features that would drop out if we use a top-k feature selection approach. 509 more words
1 week, 1 day
This is a teaser chapter for my new fic Save Me. I swore I was going to finish Love Eternal before starting this, but my muse decided otherwise. 4,019 more words
1 week, 2 days
Nowadays, where everybody is into Deep Learning, kernel machines got a little out of fashion. One reason might be that you cannot stack multiple SVMs to form something like a Deep SVM. 598 more words