Statistics for machine learning, papers to start? I have a background in computer programming and elementary number theory, but no real statistics training, and have recently "discovered" that the amazing world of a whole range of techniques is actually a statistical world. It seems that matrix factorizations, matrix completion, high dimensional tensors, embeddings, density estimation, Bayesian inference, Markov partitions, eigenvector computation, PageRank are all highly statistical techniques, and that the machine learning algorithms that use such things, use a lot of statistics. 
My goal is to be able to read papers that discuss such things, and implement or create the algorithms, while understanding the notation, "proofs" and statistical arguments used. I guess the hardest thing is to follow all the proofs that involve matrices. 
What basic papers can get me started? Or a good textbook with exercises that are worth working through ? 
Specifically, some papers I would like to understand completely are :


*

*Exact Matrix Completion via Convex Optimization, Candes, Recht, 2008

*The Fast Cauchy Transform and Faster Robust Linear Regression, Clarkson et al, 2013

*Random Projections for Support Vector Machines, Paul et al, 2013

*High-Dimensional Probability Estimation with Deep Density Models, Rippel, Adams, 2013

*Obtaining Error-Minimizing Estimates and Universal Entry-Wise Error Bounds for Low-Rank Matrix Completion, Király, Theran, 2013
 A: I would recommend Andrew Ngs Machine Learning course on Coursera, it does a brilliant coverage on all the basics. If you are studying anything to do with probabilistic graphical models Daphne Kollers course would be good to have a look at too. 
This is a treasure trove for self-study resources too http://ragle.sanukcode.net/articles/machine-learning-self-study-resources/ Herb Grossman's lectures are awesome. 
I've also been recommended this book https://www.openintro.org/stat/textbook.php as I'm always still learning myself and stats is not my background! 
My two cents re the maths side of things and papers though is don't get too caught up on the background maths. Learn the basics and reference the papers that those papers you mentioned are built on and see are they easier (maybe you'll have to go back a few papers to get something you can understand -it's what I do myself) there are a lot of different elements of maths in ML and it's easy to get sucked down a rabbit hole (again something I've done myself more than once!). 
Best of luck, it's a really interesting field!
