The course investigates landmark convex optimization algorithms and their complexity results. We aim to cover both key theory, such as polynomial time algorithms for linear programming and the optimality of Nesterov’s method, while also surveying current practical state-of-the-art methods. Homework will consist of a choice between theory and programming.
The course is 4720-009 / 5720-009.
There is no required textbook for the course; see syllabus for more information on supplemental resources.
Some supplemental resources:
More advanced resources:
More resources are mentioned in the syllabus.
|Stephen Becker||ECCR 151||MWF 1 to 1:50 PM|
The quizzes are found on the Stanford EE364a lecture slides page.
|1||1/18/17||Intro, convexity||[Boyd and Vandenberghe 2004, ch 1]. Take the convex intro quiz.|
|2||1/23/17||Convex sets||[Boyd and Vandenberghe 2004, ch 2]. Take the convex set quiz.|
|3||1/30/17||Convex functions||[Boyd and Vandenberghe 2004, ch 3; and strong convexity and Lipschitz constants handout]. Take the convex function quiz.|
[Boyd and Vandenberghe 2004, ch 3; and P. L. Combettes and V. R. Wajs, "Signal recovery by proximal forward-backward splitting" (2005)].
see also What shape is your Conjugate (Yves Lucet, SIAM Review 2009; free version); and Proximal splitting methods in signal processing (P. L. Combettes and J.-C. Pesquet, 2011) for proximal calculs rules
|Instructor/TA||Room Number||Office Hours|
|Stephen Becker||ECOT 231||Wednesday, 2-4 PM; Thurs 1-2 PM|
Homework solutions are to uploaded to D2L.
Homeworks are due every two weeks.
|Homework 1 and 2||Fri Jan 27 2017||Solutions posted on D2L|
|Homework 3 and 4||Fri Feb 10 2017||Solutions posted on D2L|
|Homework 5 and 6; see also logistic regression handout on D2L||Fri Feb 24 2017||Solutions posted on D2L|
|Homework 7 and 8||Fri Mar 10 2017||Solutions posted on D2L|
|Homework 9 and 10||Fri Mar 24 2017|
There are no exams
There will be one class project, worth 25% of your grade. We suggest working in groups of up to three, but you may work alone if necessary.