Convex optimization for statistics and machine learning.
To make this more manageable, I decided to separate out this effort into two books: book 1 for analysis, and book 2 for algorithms. Book 2 has not GitHub repo at the moment.
(Will I ever finish? One can be hopeful.)
This repo is for book 1. Currently I'm done with Parts 2 and 3.
Tentative plan:
- Write Part 4 (duality and optimality)
- Lagrange duality
- KKT conditions
- Duality correspondences
- Write Part 5 (case studies on lasso and SVMs)
- Write Part 1 (introduction) and technical appendices
- Collect ideas for Part 6 (advanced topics). Some ideas:
- Uniqueness without strict convexity?
- Caratheodory theorems on sparsity?
- Bregman divergences, projections, proximals?
- Perturbation/sensitivity analysis?
- Or variational analysis (for mere mortals)?