Robust data-driven accelerated mirror descent

Tan, Hong Ye; Mukherjee, Subhadip; Tang, Junqi; Hauptmann, Andreas; Schönlieb, Carola-Bibiane (2023-05-05)
 
Avaa tiedosto
Lataukset: 

URL:
https://doi.org/10.1109/ICASSP49357.2023.10096875

Tan, Hong Ye
Mukherjee, Subhadip
Tang, Junqi
Hauptmann, Andreas
Schönlieb, Carola-Bibiane
Institute of Electrical and Electronics Engineers
05.05.2023

H. Y. Tan, S. Mukherjee, J. Tang, A. Hauptmann and C. -B. Schönlieb, "Robust Data-Driven Accelerated Mirror Descent," ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Rhodes Island, Greece, 2023, pp. 1-5, doi: 10.1109/ICASSP49357.2023.10096875

https://rightsstatements.org/vocab/InC/1.0/
© 2023 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
https://rightsstatements.org/vocab/InC/1.0/
doi:https://doi.org/10.1109/icassp49357.2023.10096875
Näytä kaikki kuvailutiedot
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe2023071090466
Tiivistelmä

Abstract

Learning-to-optimize is an emerging framework that leverages training data to speed up the solution of certain optimization problems. One such approach is based on the classical mirror descent algorithm, where the mirror map is modelled using input-convex neural networks. In this work, we extend this functional parameterization approach by introducing momentum into the iterations, based on the classical accelerated mirror descent. Our approach combines short-time accelerated convergence with stable long-time behavior. We empirically demonstrate additional robustness with respect to multiple parameters on denoising and deconvolution experiments.

Kokoelmat