Hacker News new | comments | ask | show | jobs | submit login
The Matrix Cookbook (2012) [pdf] (dtu.dk)
103 points by nabla9 on July 8, 2017 | hide | past | web | favorite | 9 comments

I clicked wondering what kind of food they were making from the movies.

Must confess, before clicking my mind immediately went to "wonder if it contains a recipe for the mush they ate on the nebuchadnezzar?"

Since we're on the topic of matrices and linear algebra, here is a tutorial on the basics: https://minireference.com/static/tutorials/linear_algebra_in...

This PDF literally got me through grad school. It's an amazing reference.

Similar, but more "cheat-sheet" style: http://www.cs.nyu.edu/~roweis/notes/matrixid.pdf

This is an awesome resource that I keep coming back to again and again. Save this somewhere on your computer so you'll have it handy whenever you see some weird matrix derivative...

I used to use this a bunch when calculating gradient descent algorithms, but I've since moved to using libraries with auto-differentiation (numpy with autodiff or theano) and it's a huge force multiplier for my work (not to mention GPU implementations basically for free).

Of course there's always some derivation that's easiest done by hand.

Isn't tensorflow the most popular auto differential library these days?

Probably you could look at # Watches, Stars, forks etc to gauge that, but if you want to look at vs theano, torch, pyTorch, Caffe2, start with cs231N lecture: http://cs231n.stanford.edu/slides/2017/cs231n_2017_lecture8....

(I think there's vid on youtube)

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact