Operator Splitting Methods in Data Analysis

**This workshop is by invitation only**

Location

This workshop will take place at SAMSI in Research Triangle Park, NC.

Description

Monotone operator splitting methods, which originated in the late 1970’s in the context of partial differential equations, have started to be highly effective for modeling and solving a wide range of data analysis and processing problems, in particular high-dimensional statistical data analysis.

This SAMSI workshop brings together experts from nonlinear analysis, optimization, data analysis and statistics.
to advance the field of operator splitting methods. Specific topics include:
(i) general forms of equilibria that are captured by monotone inclusions,
(ii) nonconvex/nonmonotone splitting methods,
(iii) scaling up of splitting strategies as the numbers of observations and features grow, and
(iv) robustness of splitting methods under perturbations and modeling errors.
A unique aspect will be the strong focus on presentations that illuminate technical difficulties and highlight open problems.


Supporting Media

**Planning for this workshop is ongoing. Please continue to monitor this page and as more information comes available, it will be posted here.