links with various tricks to get things working.
http://andrewgoldstone.com/blog/2014/12/24/slides/
https://github.com/agoldst/tex/tree/master/lecture-slides
http://pandoc.org/demo/example9/producing-slide-shows-with-pandoc.html
https://github.com/sebastianbarfort/numedig/blob/master/2-workflow/slides/slides.md
look up "metadata blocks" in http://pandoc.org/demo/example9/pandocs-markdown.html
- maybe open class with regression to see CVXPY examples
- students won't know convexity at this point
- but this gives us more examples to work with when explaining convex functions, sets, problems in lecture 2
- with regression, can explain multiobjective problems
- give examples with convex problem transformations
- clean up SVM examples so less code is duplicated
- but each problem into a function of its own: SVM, SpVC, SVC..
- make it more explicit that convex <= concave gives a convex set
- add component-wise composition rule to make more interesting examples
- adding features
- polynomial discrimination any different?
- rbf kernel? radial basis function
- kernel methods and relationship with convex opt...
- note that so far, we have said nothing about how to compute a supporting vector
- we have focused on modeling
- that's OK, we're focusing on modeling
- algorithms involve duality and optimality conditions
- make sure it matches up with python SVM formulation
- maybe even do a timing comparison...
- cleaning and centering data
- sparse predictors
- multiclass SVM
- max flow, multicommodity flow
- optimal evacuation plan
- given list of spheres (point in R^n and radius), find a point in the intersection
- find a point in intersection furthest in x direction
- intersection of "diamonds" l1-ball?
- ooh: list of norm balls: center, norm, raidus. find intersection
- polytopes: (need to introduce convex hull of points) given vertices of 2 convex polyhedra, determine if they intersect.
- hyperplane version of polytope?
- list of foods with nutrients
- min and max healthy range for each nutrient
- find a healthy diet
- NMF
- min max rps
- max flow
- portfolio opt
- opt parade route
- evacuation
- structural opt
- rockets/trajectories
- control
- model predictive control
rps max flow parade route opt evac nmf portfolio