CfP: "Nonconvex Optimization for Machine Learning" (NIPS 2016 Workshop)

[Hossein Mobahi]

CALL FOR PAPERS

NIPS 2016 Workshop:

Nonconvex Optimization for Machine Learning: Theory and Practice

December 9, 2016

Barcelona, Spain

Website: https://sites.google.com/site/nonconvexnips2016

Facebook: https://www.facebook.com/nonconvex

An award (sponsored by Google) will be given to the best student paper. 

Submission Deadline:  September 20, 2016

Acceptance Notification:  October 3, 2016

Final Paper Due:  December 1, 2016

=======================================================

A large fraction of machine learning problems require nonconvex optimization. This includes deep learning, Bayesian inference, or clustering. The objective functions in all these instances are highly non-convex, and it is an open question if there are provable, polynomial time algorithms for these problems under realistic assumptions.

A diverse set of methods have been devised to solve nonconvex problems in different settings. They range from simple local search approaches such as gradient descent and alternating minimization to more involved frameworks such as simulated annealing, continuation methods, convex hierarchies, Bayesian optimization, or branch and bound. Moreover, special classes of nonconvex problems admit provable, efficient methods, such as quasi convex optimization, star convex optimization, submodular optimization, and matrix/tensor decomposition.

This workshop is motivated by the recent burst of research activity in nonconvex optimization, and serves as a platform to bring together and discuss recent progress in its many different domains. The workshop will span both theoretical frameworks and practical issues especially in the area of deep learning.

Invited Speakers:

Jean Lasserre  (LAAS-CNRS)

Ryan Adams  (Harvard)

Francis Bach  (Ecole Normale Superieure)

Nando de Freitas  (Google Deepmind)

Surya Ganguli  (Stanford)

Stefanie Jegelka  (MIT)

Suvrit Sra  (MIT)

Submissions:

The workshop covers a very broad range of topics that lie at the intersection of nonconvex optimization and machine learning. Submissions must be in PDF format using the NIPS style. The length is limited to at most 4 pages, excluding references. The submissions may include a supplementary appendix that is optional for reviewers to read. Work that has already appeared in a journal, workshop, or conference (including NIPS 2016) must be significantly extended to be eligible for workshop submission. Submissions will be accepted as poster presentations, and the papers will be published electronically on the workshop website. 

For consideration, submissions must be sent via Email to noncon...@gmail.com by September 20.

Workshop Organizers:

Anima Anandkumar  (UC Irvine)

Percy Liang  (Stanford)

Hossein Mobahi  (Google Research)

Stefanie Jegelka  (MIT)

Anna Choromanska  (NYU)