Few-to-few Cross-domain Object Matching

Aditya Jitta; Arto Klami

Few-to-few Cross-domain Object Matching

Aditya Jitta, Arto Klami ;

Proceedings of The 3rd International Workshop on Advanced Methodologies for Bayesian Networks, PMLR 73:176-187, 2017.

Abstract

Cross-domain object matching refers to the task of inferring unknown alignment between objects in two data collections that do not have a shared data representation. In recent years several methods have been proposed for solving the special case that assumes each object is to be paired with exactly one object, resulting in a constrained optimization problem over permutations. A related problem formulation of cluster matching seeks to match a cluster of objects in one data set to a cluster of objects in the other set, which can be considered as many-to-many extension of cross-domain object matching and can be solved without explicit constraints. In this work we study the intermediate region between these two special cases, presenting a range of Bayesian inference algorithms that work also for few-to-few cross-domain object matching problems where constrained optimization is necessary but the optimization domain is broader than just permutations.

Related Material

Download PDF

@InProceedings{pmlr-v73-jitta17a, title = {Few-to-few Cross-domain Object Matching}, author = {Aditya Jitta and Arto Klami}, booktitle = {Proceedings of The 3rd International Workshop on Advanced Methodologies for Bayesian Networks}, pages = {176--187}, year = {2017}, editor = {Antti Hyttinen and Joe Suzuki and Brandon Malone}, volume = {73}, series = {Proceedings of Machine Learning Research}, address = {}, month = {20--22 Sep}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v73/jitta17a/jitta17a.pdf}, url = {http://proceedings.mlr.press/v73/jitta17a.html}, abstract = { Cross-domain object matching refers to the task of inferring unknown alignment between objects in two data collections that do not have a shared data representation. In recent years several methods have been proposed for solving the special case that assumes each object is to be paired with exactly one object, resulting in a constrained optimization problem over permutations. A related problem formulation of cluster matching seeks to match a cluster of objects in one data set to a cluster of objects in the other set, which can be considered as many-to-many extension of cross-domain object matching and can be solved without explicit constraints. In this work we study the intermediate region between these two special cases, presenting a range of Bayesian inference algorithms that work also for few-to-few cross-domain object matching problems where constrained optimization is necessary but the optimization domain is broader than just permutations. } }

%0 Conference Paper %T Few-to-few Cross-domain Object Matching %A Aditya Jitta %A Arto Klami %B Proceedings of The 3rd International Workshop on Advanced Methodologies for Bayesian Networks %C Proceedings of Machine Learning Research %D 2017 %E Antti Hyttinen %E Joe Suzuki %E Brandon Malone %F pmlr-v73-jitta17a %I PMLR %J Proceedings of Machine Learning Research %P 176--187 %U http://proceedings.mlr.press %V 73 %W PMLR %X Cross-domain object matching refers to the task of inferring unknown alignment between objects in two data collections that do not have a shared data representation. In recent years several methods have been proposed for solving the special case that assumes each object is to be paired with exactly one object, resulting in a constrained optimization problem over permutations. A related problem formulation of cluster matching seeks to match a cluster of objects in one data set to a cluster of objects in the other set, which can be considered as many-to-many extension of cross-domain object matching and can be solved without explicit constraints. In this work we study the intermediate region between these two special cases, presenting a range of Bayesian inference algorithms that work also for few-to-few cross-domain object matching problems where constrained optimization is necessary but the optimization domain is broader than just permutations.