Randomized Optimal Transport on a Graph: framework and new distance measures

The recently developed bag-of-paths (BoP) framework consists in setting a Gibbs-Boltzmann distribution on all feasible paths of a graph. This probability distribution favors short paths over long ones, with a free parameter (the temperature $T$) controlling the entropic level of the distribution. Th...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Guex, Guillaume, Kivimäki, Ilkka, Saerens, Marco
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Guex, Guillaume
Kivimäki, Ilkka
Saerens, Marco
description The recently developed bag-of-paths (BoP) framework consists in setting a Gibbs-Boltzmann distribution on all feasible paths of a graph. This probability distribution favors short paths over long ones, with a free parameter (the temperature $T$) controlling the entropic level of the distribution. This formalism enables the computation of new distances or dissimilarities, interpolating between the shortest-path and the resistance distance, which have been shown to perform well in clustering and classification tasks. In this work, the bag-of-paths formalism is extended by adding two independent equality constraints fixing starting and ending nodes distributions of paths (margins). When the temperature is low, this formalism is shown to be equivalent to a relaxation of the optimal transport problem on a network where paths carry a flow between two discrete distributions on nodes. The randomization is achieved by considering free energy minimization instead of traditional cost minimization. Algorithms computing the optimal free energy solution are developed for two types of paths: hitting (or absorbing) paths and non-hitting, regular, paths, and require the inversion of an $n \times n$ matrix with $n$ being the number of nodes. Interestingly, for regular paths on an undirected graph, the resulting optimal policy interpolates between the deterministic optimal transport policy ($T \rightarrow 0^{+}$) and the solution to the corresponding electrical circuit ($T \rightarrow \infty$). Two distance measures between nodes and a dissimilarity between groups of nodes, both integrating weights on nodes, are derived from this framework.
doi_str_mv 10.48550/arxiv.1806.03232
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_1806_03232</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>1806_03232</sourcerecordid><originalsourceid>FETCH-LOGICAL-a672-e2acf54edaeaec200ae367ff5c61ff9af09a1a09a3053730f859fbe6766207bc3</originalsourceid><addsrcrecordid>eNotj8FOwzAQRH3hgAofwIn9gYSNXdsJN1RBQYpUCeUebZ21iGicyA4U-HpC6WXmNE_zhLgpMF-XWuMdxa_-My9KNDkqqeSlqF8pdOPQ_3AHu2nuBzpAEymkaYwzjAEItpGmt3vwkQY-jvEdlgUEPkLXp5mCYxiY0kfkdCUuPB0SX597JZqnx2bznNW77cvmoc7IWJmxJOf1mjtiYicRiZWx3mtnCu8r8lhRQUso1Moq9KWu_J6NNUai3Tu1Erf_2JNOO8Xldfxu_7Tak5b6BUy0SQI</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Randomized Optimal Transport on a Graph: framework and new distance measures</title><source>arXiv.org</source><creator>Guex, Guillaume ; Kivimäki, Ilkka ; Saerens, Marco</creator><creatorcontrib>Guex, Guillaume ; Kivimäki, Ilkka ; Saerens, Marco</creatorcontrib><description>The recently developed bag-of-paths (BoP) framework consists in setting a Gibbs-Boltzmann distribution on all feasible paths of a graph. This probability distribution favors short paths over long ones, with a free parameter (the temperature $T$) controlling the entropic level of the distribution. This formalism enables the computation of new distances or dissimilarities, interpolating between the shortest-path and the resistance distance, which have been shown to perform well in clustering and classification tasks. In this work, the bag-of-paths formalism is extended by adding two independent equality constraints fixing starting and ending nodes distributions of paths (margins). When the temperature is low, this formalism is shown to be equivalent to a relaxation of the optimal transport problem on a network where paths carry a flow between two discrete distributions on nodes. The randomization is achieved by considering free energy minimization instead of traditional cost minimization. Algorithms computing the optimal free energy solution are developed for two types of paths: hitting (or absorbing) paths and non-hitting, regular, paths, and require the inversion of an $n \times n$ matrix with $n$ being the number of nodes. Interestingly, for regular paths on an undirected graph, the resulting optimal policy interpolates between the deterministic optimal transport policy ($T \rightarrow 0^{+}$) and the solution to the corresponding electrical circuit ($T \rightarrow \infty$). Two distance measures between nodes and a dissimilarity between groups of nodes, both integrating weights on nodes, are derived from this framework.</description><identifier>DOI: 10.48550/arxiv.1806.03232</identifier><language>eng</language><subject>Computer Science - Learning ; Computer Science - Social and Information Networks ; Statistics - Machine Learning</subject><creationdate>2018-06</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/1806.03232$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.1806.03232$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Guex, Guillaume</creatorcontrib><creatorcontrib>Kivimäki, Ilkka</creatorcontrib><creatorcontrib>Saerens, Marco</creatorcontrib><title>Randomized Optimal Transport on a Graph: framework and new distance measures</title><description>The recently developed bag-of-paths (BoP) framework consists in setting a Gibbs-Boltzmann distribution on all feasible paths of a graph. This probability distribution favors short paths over long ones, with a free parameter (the temperature $T$) controlling the entropic level of the distribution. This formalism enables the computation of new distances or dissimilarities, interpolating between the shortest-path and the resistance distance, which have been shown to perform well in clustering and classification tasks. In this work, the bag-of-paths formalism is extended by adding two independent equality constraints fixing starting and ending nodes distributions of paths (margins). When the temperature is low, this formalism is shown to be equivalent to a relaxation of the optimal transport problem on a network where paths carry a flow between two discrete distributions on nodes. The randomization is achieved by considering free energy minimization instead of traditional cost minimization. Algorithms computing the optimal free energy solution are developed for two types of paths: hitting (or absorbing) paths and non-hitting, regular, paths, and require the inversion of an $n \times n$ matrix with $n$ being the number of nodes. Interestingly, for regular paths on an undirected graph, the resulting optimal policy interpolates between the deterministic optimal transport policy ($T \rightarrow 0^{+}$) and the solution to the corresponding electrical circuit ($T \rightarrow \infty$). Two distance measures between nodes and a dissimilarity between groups of nodes, both integrating weights on nodes, are derived from this framework.</description><subject>Computer Science - Learning</subject><subject>Computer Science - Social and Information Networks</subject><subject>Statistics - Machine Learning</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2018</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj8FOwzAQRH3hgAofwIn9gYSNXdsJN1RBQYpUCeUebZ21iGicyA4U-HpC6WXmNE_zhLgpMF-XWuMdxa_-My9KNDkqqeSlqF8pdOPQ_3AHu2nuBzpAEymkaYwzjAEItpGmt3vwkQY-jvEdlgUEPkLXp5mCYxiY0kfkdCUuPB0SX597JZqnx2bznNW77cvmoc7IWJmxJOf1mjtiYicRiZWx3mtnCu8r8lhRQUso1Moq9KWu_J6NNUai3Tu1Erf_2JNOO8Xldfxu_7Tak5b6BUy0SQI</recordid><startdate>20180607</startdate><enddate>20180607</enddate><creator>Guex, Guillaume</creator><creator>Kivimäki, Ilkka</creator><creator>Saerens, Marco</creator><scope>AKY</scope><scope>EPD</scope><scope>GOX</scope></search><sort><creationdate>20180607</creationdate><title>Randomized Optimal Transport on a Graph: framework and new distance measures</title><author>Guex, Guillaume ; Kivimäki, Ilkka ; Saerens, Marco</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a672-e2acf54edaeaec200ae367ff5c61ff9af09a1a09a3053730f859fbe6766207bc3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2018</creationdate><topic>Computer Science - Learning</topic><topic>Computer Science - Social and Information Networks</topic><topic>Statistics - Machine Learning</topic><toplevel>online_resources</toplevel><creatorcontrib>Guex, Guillaume</creatorcontrib><creatorcontrib>Kivimäki, Ilkka</creatorcontrib><creatorcontrib>Saerens, Marco</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv Statistics</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Guex, Guillaume</au><au>Kivimäki, Ilkka</au><au>Saerens, Marco</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Randomized Optimal Transport on a Graph: framework and new distance measures</atitle><date>2018-06-07</date><risdate>2018</risdate><abstract>The recently developed bag-of-paths (BoP) framework consists in setting a Gibbs-Boltzmann distribution on all feasible paths of a graph. This probability distribution favors short paths over long ones, with a free parameter (the temperature $T$) controlling the entropic level of the distribution. This formalism enables the computation of new distances or dissimilarities, interpolating between the shortest-path and the resistance distance, which have been shown to perform well in clustering and classification tasks. In this work, the bag-of-paths formalism is extended by adding two independent equality constraints fixing starting and ending nodes distributions of paths (margins). When the temperature is low, this formalism is shown to be equivalent to a relaxation of the optimal transport problem on a network where paths carry a flow between two discrete distributions on nodes. The randomization is achieved by considering free energy minimization instead of traditional cost minimization. Algorithms computing the optimal free energy solution are developed for two types of paths: hitting (or absorbing) paths and non-hitting, regular, paths, and require the inversion of an $n \times n$ matrix with $n$ being the number of nodes. Interestingly, for regular paths on an undirected graph, the resulting optimal policy interpolates between the deterministic optimal transport policy ($T \rightarrow 0^{+}$) and the solution to the corresponding electrical circuit ($T \rightarrow \infty$). Two distance measures between nodes and a dissimilarity between groups of nodes, both integrating weights on nodes, are derived from this framework.</abstract><doi>10.48550/arxiv.1806.03232</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.1806.03232
ispartof
issn
language eng
recordid cdi_arxiv_primary_1806_03232
source arXiv.org
subjects Computer Science - Learning
Computer Science - Social and Information Networks
Statistics - Machine Learning
title Randomized Optimal Transport on a Graph: framework and new distance measures
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-26T18%3A29%3A40IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Randomized%20Optimal%20Transport%20on%20a%20Graph:%20framework%20and%20new%20distance%20measures&rft.au=Guex,%20Guillaume&rft.date=2018-06-07&rft_id=info:doi/10.48550/arxiv.1806.03232&rft_dat=%3Carxiv_GOX%3E1806_03232%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true