Lifts of Convex Sets and Cone Factorizations
Add to your list(s)
Download to your calendar using vCal
If you have a question about this talk, please contact Microsoft Research Cambridge Talks Admins.
This event may be recorded and made available internally or externally via http://research.microsoft.com. Microsoft will own the copyright of any recordings made. If you do not wish to have your image/voice recorded please consider this before attending
The representation of a convex set is crucial for the efficiency of linear optimization algorithms. A common idea to optimize a linear function over a complicated convex set is to express the set as the projection of a much simpler convex set in a higher dimension, called
a ``lift’’ of the original set. In the early 1990s Yannakakis showed that there is a remarkable connection between the size of the smallest polyhedral lift of a polytope and the nonnegative rank of the slack
matrix of the polytope. I will show how this theorem can be generalized to convex sets via cone factorizations of nonnegative operators. In practice, one usually only has a numerical approximation to a cone factorization. I will also show how such
approximate factorizations can be used to construct efficient approximations of polytopes, and mention some of the many open questions in this area.
Joint work with Joao Gouveia (University of Coimbra) and Pablo Parrilo
(MIT)
This talk is part of the Microsoft Research Machine Learning and Perception Seminars series.
This talk is included in these lists:
Note that ex-directory lists are not shown.
|