Welcome to the homepage of the

Mathematical Optimization Group

Chair of Applied Mathematics, Saarland University, Germany

Mathematical Optimization Group

Prof. Dr. Peter Ochs


Peter Ochs

Mahesh Chandra Mukkamala

Sheheryar Mehmood

News:
30.09.2020: Our webpage has moved to the University of Tuebingen.
13.08.2020: In September 2020, our group is going to move to the University of Tübingen.
Our paper "Convex-Concave Backtracking for Inertial Bregman Proximal Gradient Algorithms in Non-Convex Optimization" by Mahesh Chandra Mukkamala, Peter Ochs, Thomas Pock and Shoham Sabach is now published in SIAM Journal on Mathematics of Data Science.
01.01.2020: Sheheryar Mehmood joins the group.
Our paper Automatic Differentiation of Some First-Order Methods in Parametric Optimization is accepted at AISTATS 2020 in Palermo, Sicily, Italy.
15.10.2019: The new technical report "Automatic Differentiation of Some First-Order Methods in Parametric Optimization" by Sheheryar Mehmood and Peter Ochs is now available on arXiv.
10.10.2019: The new technical report "Bregman Proximal Framework for Deep Linear Neural Networks" by Mahesh Chandra Mukkamala, Felix Westerkamp, Emanuel Laude, Daniel Cremers and Peter Ochs is now available on arXiv.
Our paper Beyond Alternating Updates for Matrix Factorization with Inertial Bregman Proximal Gradient Algorithms is accepted at NeuRIPS 2019 in Vancouver, Canada.
Our paper Bregman Proximal Mappings and Bregman-Moreau Envelopes under Relative Prox-Regularity is accepted for publication in Journal of Optimization Theory and Applications (JOTA).
10.07.2019: The new technical report "Bregman Proximal Mappings and Bregman-Moreau Envelopes under Relative Prox-Regularity" by Emanuel Laude, Peter Ochs and Daniel Cremers is now available on arXiv.
Our paper Adaptive Fista is accepted for publication in SIAM Journal of Optimization (SIOPT).
Our paper On Quasi-Newton Forward--Backward Splitting: Proximal Calculus and Convergence is accepted for publication in SIAM Journal of Optimization (SIOPT).
23.05.2019: The new technical report "Beyond Alternating Updates for Matrix Factorization with Inertial Bregman Proximal Gradient Algorithms" by Mahesh Chandra Mukkamala and Peter Ochs is now available on arXiv.
Our paper Model Function Based Conditional Gradient Method with Armijo-like Line Search is accepted at ICML 2019 in Long Beach, California.
09.04.2019: The new technical report "Convex-Concave Backtracking for Inertial Bregman Proximal Gradient Algorithms in Non-Convex Optimization" by Mahesh Chandra Mukkamala, Peter Ochs, Thomas Pock and Shoham Sabach is now available on arXiv.
29.01.2019: The new technical report "Model Function Based Conditional Gradient Method with Armijo-like Line Search" by Yura Malitsky and Peter Ochs is now available on arXiv.
Our paper Non-smooth Non-convex Bregman Minimization: Unification and new Algorithms is accepted for publication in Journal of Optimization Theory and Applications (JOTA).
Our paper Unifying abstract inexact convergence theorems and block coordinate variable metric iPiano is accepted for publication in SIAM Journal of Optimization (SIOPT).
01.06.2018: Mahesh Chandra Mukkamala joins the group.
Our paper Lifting Layers: Analysis and Applications is accepted for oral presentation at ECCV 2018 in Munich, Germany.
Invited talk at the International Symposium on Mathematical Programming (ISMP), July 1-6, 2018, in Bordeaux, France: Peter Ochs will give a talk in the Minisymposium "Recent advances in first-order algorithms for non-smooth optimization", organized by Thomas Pock.
Invited Short Course at the INDAM intensive period on "Computational Methods for Inverse Problems in Imaging", June 11-15, 2018, University of Insubria DISAT, Como, Italy: Peter Ochs will give a Short Course on "Accelerations of Forward-Backward Splitting".
29.01.2018: The new technical report "On Quasi-Newton Forward--Backward Splitting: Proximal Calculus and Convergence" by Stephen Becker, Jalal Fadili and Peter Ochs is now available on arXiv.
29.01.2018: A revised version of "Local Convergence of the Heavy-ball Method and iPiano for Non-convex Optimization" is now available on arXiv and OptimizationOnline.
22.11.2017: A revised version of "Unifying abstract inexact convergence theorems and block coordinate variable metric iPiano" is now available on arXiv and OptimizationOnline.
15.11.2017: The new technical report "Adaptive Fista" by Thomas Pock and Peter Ochs is now available on arXiv and OptimizationOnline.
Invited talks at the SIAM Conference on Imaging Sciences, June 5-8, 2018, in Bologna, Italy: Peter Ochs will give a talk in the Minisymposium "Advanced optimization methods for image processing", organized by Marco Prato and Ignace Loris, and in the Minisymposium "Approaches for fast optimisation in imaging and inverse problems", organized by Jingwei Liang, Carola-Bibiane Schönlieb, Mila Nikolova.
Peter Ochs will be an invited speaker at the Workshop: Imaging and Vision from Theory to Applications, March 14-16, 2018, in Siegen, Germany.
Peter Ochs will be an invited speaker at the Mathematics and Image Analysis (MIA) Conference, January 15-17, 2018, in Berlin, Germany.
01.11.2017: Webpage is online.


MOP Group
©2017-2020
The author is not
responsible for
the content of
external pages.