Docs | Installation | Cite
reg4opt is a Python implementation of operator regression and convex regression methods presented in the paper "OpReg-Boost: Learning to Accelerate Online Algorithms with Operator Regression" by Nicola Bastianello, Andrea Simonetto, Emiliano Dall'Anese.
The documentation is available here.
reg4opt works on Python 3.7 and depends on: tvopt, numpy, scipy. To run the examples, cvxpy may also be needed.
pip install reg4opt
@article{bastianello_opregboost_2021,
title = {{OpReg}-{Boost}: {Learning} to {Accelerate} {Online} {Algorithms} with {Operator} {Regression}},
url = {http://arxiv.org/abs/2105.13271},
journal = {arXiv:2105.13271 [cs, math]},
author = {Bastianello, Nicola and Simonetto, Andrea and Dall'Anese, Emiliano},
year = {2021}
}
reg4opt is developed by Nicola Bastianello under the supervision of Andrea Simonetto and Emiliano Dall'Anese