alicez

6 ms·

1.
Autotagging Hacker News(blog.dato.com)
5 points by alicez 11y ago|0 comments
2.
by alicez 11y ago
Well, you can use the bootstrap to calculate the variance. It costs computation. But it works. Cosma Shalizi wrote a really nice introduction to it: http://www.americanscientist.org/issues/pub/2010/3/the-
3.
by alicez 11y ago
Great idea. Will run some experiments to see how it performs. It sounds analogous to kmeans++ initialization. Sobol sequences ring a bell. Some of the Bayesian optimization software libraries may in fact use a Sobol sequence of initial eval
4.
by alicez 11y ago
Don't be sad. I'm happy to update the blog post as needed. By overfitting, do you mean over-optimizing the results on the validation set? Based on what I understand about nested CV, it is only necessary if 1. the hold-out validati
5.
Practical deep text learning on blogs(blog.dato.com)
5 points by alicez 11y ago|0 comments
6.
by alicez 11y ago
Dato targets MS-DOS as the next Big Data platform, processes terabytes of data in 640KB of RAM http://blog.dato.com/dato-on-ms-dos
7.
Deep Learning Made Doubly Easy with GraphLab Create(blog.dato.com)
15 points by alicez 12y ago|0 comments
8.
by alicez 12y ago
Linear algebra is one of my favorite subjects. Its design is beautifully simple, yet extremely powerful. Half of modern machine learning (and all of Matlab) is built on matrix algebra. And the existence of fast software for numeric linear a
9.
by alicez 12y ago
Cool! Thanks for the link. I've not read that book, but it looks very interesting and related. It great to see others' perspective on math as a human construction. Another book on the topic of history of mathematics is "Journ