Opening Workshop: August 29 – September 2, 2016

Application deadline is June 14, 2016.

Location

This workshop will be held at the Hamner Conference Center in Research Triangle Park.

The Opening Workshop will start with a day of tutorials on mathematical and statistical approaches towards Optimization, and their applications. The remaining days will feature presentations, discussions, formation of research working groups, and a poster reception.

Topics under consideration include: optimization under uncertainty; convex and semidefinite optimization; robust optimization sparse regression, and stochastic gradient descent; E/M & M/M algorithms; mixed integer, linear & nonlinear optimization; PDE-constrained optimization with uncertainties; statistical inverse problems; computation of high-dimensional covariance functions; and Bayesian optimization & computational decision analysis. Applications to be considered include: machine learning; image & signal processing, and compressed sensing; energy & power grids; finance; and radio therapy optimization.

Questions: email opt@samsi.info

Schedule and Supporting Media

Participant List

Posters

Printable Schedule

Speaker Titles/Abstracts

Monday, August 29, 2016
Hamner Conference Center RTP

Time Description Speaker Slides Videos
9:15-10:15 Optimization Methods for Computational Statistics and Data Analysis Steve Wright, Wisconsin PDF
10:30-11:30 Variational Functions of Gram Matrices: Properties and Optimization Algorithms Maryam Fazel, Washington PDF
1:00-2:00 PDE Constrained Optimization with Uncertain Data Matthias Heinkenschloss, Rice PDF
2:00-2:15 Working Groups at SAMSI Ilse Ipsen, SAMSI
2:30-3:30 A Selective Overview of High-dimensional Statistics and Optimization Hui Zou, Minnesota PDF
4:00-5:00 Sparse Multivariate Statistics with Discrete Optimization Rahul Mazumder, MIT PDF
5:00-7:00 Poster Session and Reception

Tuesday, August 30, 2016
Hamner Conference Center RTP

Time Description Speaker Slides Videos
9:00-9:45 Mathematical Challenges of Energy Systems Mihai Anitescu, Argonne PDF
9:45-10:30 The Evolving Optimization Needs and Challenges in the Oil and Gas Industry Amr El-Bakry, Exxon
11:00-11:45 # Numerical Optimization as Probabilistic Inference Philipp Hennig, Tuebingen, Germany PDF
1:00-1:45 Bayesian Inversion Applied to an Ice Sheet Flow Problem and to Power Grid Noemi Petra, UC Merced PDF
1:45-2:30 Toward more Faithful Simulations of Cascade Failures in Power Systems Jonathan Weare, Chicago PDF
3:00-5:00 Working Group Presentations

Wednesday, August 31, 2016
Hamner Conference Center RTP

Time Description Speaker Slides Videos
9:00-9:45 Mixed-integer PDE-constrained Optimization Sven Leyffer, Argonne PDF
9:45-10:30 Highly Scalable Parallel Branch and Bound Cynthia Phillips, Sandia PDF
11:00-11:45 New Results on Optimization in Shape Manifolds Volker Schulz, Trier, Germany PDF
1:00-1:45 Coordinate Update Algorithms for Optimization Problems in Machine Learning and Signal Processing Wotao Yin, UCLA PDF
1:45-2:30 Sub-sampled Newton Methods: Uniform and Non-Uniform Sampling Fred Roosta, UC Berkeley PDF
3:00-5:00
Optimization Challenges at the National Labs
Sven Leyffer, Argonne; Bart Van Bloemen Waanders, Sandia; Cynthia Phillips, Sandia; Habib Najm, Sandia; Jean-Paul Watson, Sandia Leyffer Phillips Watson

Thursday, September 1, 2016
Hamner Conference Center RTP

Time Description Speaker Slides Videos
9:00-9:45
Measure Transport and Optimization-based Samplers for Bayesian Computation
Youssef Marzouk, MIT PDF
9:45-10:30 Modeling and Optimization within Interacting Systems Michael Ferris, Wisconsin PDF
11:00-11:45 Distributed Algorithms for Optimization and Nash Games on Graphs Angelia Nedich, Arizona State PDF
1:00-1:45 Numerical Methods for Optimization-based Model Validation Ekaterina Kostina, Heidelberg, Germany PDF
1:45-2:30 Scalability and Data Synthesis in Computational Advertising and Marketing Jewell Thomas, MaxPoint
3:00-5:00 Working Group Formation

Friday, September 2, 2016
SAMSI

Time Description Speaker Slides Videos
9:00-9:45 Probabilistic Numerics: A New Concept? Mark Girolami, Warwick, UK
9:45-10:30 Algorithmic Regularization Paths: a new approach to variable selection Genevera Allen, Rice
11:00-11:45 Reduced Order Models in Optimization with PDEs/SDEs Ekkehard Sachs, U Trier
1:00 Adjourn