Saved in:
Bibliographic Details
Main Authors: Rolínek, Michal, Swoboda, Paul, Zietlow, Dominik, Paulus, Anselm, Musil, Vít, Martius, Georg
Format: Preprint
Published: 2020
Subjects:
Online Access:https://arxiv.org/abs/2003.11657
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910743567269888
author Rolínek, Michal
Swoboda, Paul
Zietlow, Dominik
Paulus, Anselm
Musil, Vít
Martius, Georg
author_facet Rolínek, Michal
Swoboda, Paul
Zietlow, Dominik
Paulus, Anselm
Musil, Vít
Martius, Georg
contents Building on recent progress at the intersection of combinatorial optimization and deep learning, we propose an end-to-end trainable architecture for deep graph matching that contains unmodified combinatorial solvers. Using the presence of heavily optimized combinatorial solvers together with some improvements in architecture design, we advance state-of-the-art on deep graph matching benchmarks for keypoint correspondence. In addition, we highlight the conceptual advantages of incorporating solvers into deep learning architectures, such as the possibility of post-processing with a strong multi-graph matching solver or the indifference to changes in the training setting. Finally, we propose two new challenging experimental setups. The code is available at https://github.com/martius-lab/blackbox-deep-graph-matching
format Preprint
id arxiv_https___arxiv_org_abs_2003_11657
institution arXiv
publishDate 2020
record_format arxiv
spellingShingle Deep Graph Matching via Blackbox Differentiation of Combinatorial Solvers
Rolínek, Michal
Swoboda, Paul
Zietlow, Dominik
Paulus, Anselm
Musil, Vít
Martius, Georg
Machine Learning
Building on recent progress at the intersection of combinatorial optimization and deep learning, we propose an end-to-end trainable architecture for deep graph matching that contains unmodified combinatorial solvers. Using the presence of heavily optimized combinatorial solvers together with some improvements in architecture design, we advance state-of-the-art on deep graph matching benchmarks for keypoint correspondence. In addition, we highlight the conceptual advantages of incorporating solvers into deep learning architectures, such as the possibility of post-processing with a strong multi-graph matching solver or the indifference to changes in the training setting. Finally, we propose two new challenging experimental setups. The code is available at https://github.com/martius-lab/blackbox-deep-graph-matching
title Deep Graph Matching via Blackbox Differentiation of Combinatorial Solvers
topic Machine Learning
url https://arxiv.org/abs/2003.11657