A Review of the Gumbel-max Trick and its Extensions for Discrete Stochasticity in Machine Learning
The Gumbel-max trick is a method to draw a sample from a categorical distribution, given by its unnormalized (log-)probabilities. Over the past years, the machine learning community has proposed several extensions of this trick to facilitate, e.g., drawing multiple samples, sampling from structured...
Gespeichert in:
| Veröffentlicht in: | IEEE transactions on pattern analysis and machine intelligence 2023-02, Vol.45 (2), p.1353-1371 |
|---|---|
| Hauptverfasser: | , , , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext bestellen |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 1371 |
|---|---|
| container_issue | 2 |
| container_start_page | 1353 |
| container_title | IEEE transactions on pattern analysis and machine intelligence |
| container_volume | 45 |
| creator | Huijben, Iris A. M. Kool, Wouter Paulus, Max B. van Sloun, Ruud J. G. |
| description | The Gumbel-max trick is a method to draw a sample from a categorical distribution, given by its unnormalized (log-)probabilities. Over the past years, the machine learning community has proposed several extensions of this trick to facilitate, e.g., drawing multiple samples, sampling from structured domains, or gradient estimation for error backpropagation in neural network optimization. The goal of this survey article is to present background about the Gumbel-max trick, and to provide a structured overview of its extensions to ease algorithm selection. Moreover, it presents a comprehensive outline of (machine learning) literature in which Gumbel-based algorithms have been leveraged, reviews commonly-made design choices, and sketches a future perspective. |
| doi_str_mv | 10.1109/TPAMI.2022.3157042 |
| format | Article |
| fullrecord | <record><control><sourceid>proquest_RIE</sourceid><recordid>TN_cdi_proquest_miscellaneous_2637320514</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>9729603</ieee_id><sourcerecordid>2637320514</sourcerecordid><originalsourceid>FETCH-LOGICAL-c395t-a87199616c251c6b1eab30f35501066d9e917ddd2cc823d631c0ce24d5878b5b3</originalsourceid><addsrcrecordid>eNpdkcFu1DAQhi1ERZfCC4CELHHhkq3Hju34uCqlVNqqCJaz5TgT1mXjFNuB9u1J2aWHnuYw3_9rNB8hb4AtAZg53XxZXV0uOeN8KUBqVvNnZAFGmEpIYZ6TBQPFq6bhzTF5mfMNY1BLJl6QYyG5rI2WC9Ku6Ff8HfAPHXtatkgvpqHFXTW4O7pJwf-kLnY0lEzP7wrGHMaYaT8m-jFkn7Ag_VZGv3W5BB_KPQ2RXjm_DRHpGl2KIf54RY56t8v4-jBPyPdP55uzz9X6-uLybLWuvDCyVK7RYIwC5bkEr1pA1wrWCykZMKU6gwZ013Xc-4aLTgnwzCOvO9noppWtOCEf9r23afw1YS52mG_E3c5FHKdsuRJacCahntH3T9CbcUpxvs5yrUDoGjSbKb6nfBpzTtjb2xQGl-4tMPtgwP4zYB8M2IOBOfTuUD21A3aPkf8vn4G3eyAg4uPaaG4UE-IvL4SItQ</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2761374170</pqid></control><display><type>article</type><title>A Review of the Gumbel-max Trick and its Extensions for Discrete Stochasticity in Machine Learning</title><source>IEEE Electronic Library (IEL)</source><creator>Huijben, Iris A. M. ; Kool, Wouter ; Paulus, Max B. ; van Sloun, Ruud J. G.</creator><creatorcontrib>Huijben, Iris A. M. ; Kool, Wouter ; Paulus, Max B. ; van Sloun, Ruud J. G.</creatorcontrib><description>The Gumbel-max trick is a method to draw a sample from a categorical distribution, given by its unnormalized (log-)probabilities. Over the past years, the machine learning community has proposed several extensions of this trick to facilitate, e.g., drawing multiple samples, sampling from structured domains, or gradient estimation for error backpropagation in neural network optimization. The goal of this survey article is to present background about the Gumbel-max trick, and to provide a structured overview of its extensions to ease algorithm selection. Moreover, it presents a comprehensive outline of (machine learning) literature in which Gumbel-based algorithms have been leveraged, reviews commonly-made design choices, and sketches a future perspective.</description><identifier>ISSN: 0162-8828</identifier><identifier>EISSN: 1939-3539</identifier><identifier>EISSN: 2160-9292</identifier><identifier>DOI: 10.1109/TPAMI.2022.3157042</identifier><identifier>PMID: 35254975</identifier><identifier>CODEN: ITPIDJ</identifier><language>eng</language><publisher>United States: IEEE</publisher><subject>Algorithms ; Back propagation ; Back propagation networks ; categorical distribution ; Computational modeling ; Data models ; Deep learning ; gradient estimation ; Gumbel-max trick ; gumbel-softmax ; Laplace equations ; Machine learning ; Neural networks ; Optimization ; Random variables ; sampling ; Sketches ; Stochastic processes ; structured models</subject><ispartof>IEEE transactions on pattern analysis and machine intelligence, 2023-02, Vol.45 (2), p.1353-1371</ispartof><rights>Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2023</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c395t-a87199616c251c6b1eab30f35501066d9e917ddd2cc823d631c0ce24d5878b5b3</citedby><cites>FETCH-LOGICAL-c395t-a87199616c251c6b1eab30f35501066d9e917ddd2cc823d631c0ce24d5878b5b3</cites><orcidid>0000-0002-1837-1454 ; 0000-0002-2629-3898 ; 0000-0003-2845-0495</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/9729603$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,780,784,796,27923,27924,54757</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/9729603$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/35254975$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Huijben, Iris A. M.</creatorcontrib><creatorcontrib>Kool, Wouter</creatorcontrib><creatorcontrib>Paulus, Max B.</creatorcontrib><creatorcontrib>van Sloun, Ruud J. G.</creatorcontrib><title>A Review of the Gumbel-max Trick and its Extensions for Discrete Stochasticity in Machine Learning</title><title>IEEE transactions on pattern analysis and machine intelligence</title><addtitle>TPAMI</addtitle><addtitle>IEEE Trans Pattern Anal Mach Intell</addtitle><description>The Gumbel-max trick is a method to draw a sample from a categorical distribution, given by its unnormalized (log-)probabilities. Over the past years, the machine learning community has proposed several extensions of this trick to facilitate, e.g., drawing multiple samples, sampling from structured domains, or gradient estimation for error backpropagation in neural network optimization. The goal of this survey article is to present background about the Gumbel-max trick, and to provide a structured overview of its extensions to ease algorithm selection. Moreover, it presents a comprehensive outline of (machine learning) literature in which Gumbel-based algorithms have been leveraged, reviews commonly-made design choices, and sketches a future perspective.</description><subject>Algorithms</subject><subject>Back propagation</subject><subject>Back propagation networks</subject><subject>categorical distribution</subject><subject>Computational modeling</subject><subject>Data models</subject><subject>Deep learning</subject><subject>gradient estimation</subject><subject>Gumbel-max trick</subject><subject>gumbel-softmax</subject><subject>Laplace equations</subject><subject>Machine learning</subject><subject>Neural networks</subject><subject>Optimization</subject><subject>Random variables</subject><subject>sampling</subject><subject>Sketches</subject><subject>Stochastic processes</subject><subject>structured models</subject><issn>0162-8828</issn><issn>1939-3539</issn><issn>2160-9292</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>RIE</sourceid><recordid>eNpdkcFu1DAQhi1ERZfCC4CELHHhkq3Hju34uCqlVNqqCJaz5TgT1mXjFNuB9u1J2aWHnuYw3_9rNB8hb4AtAZg53XxZXV0uOeN8KUBqVvNnZAFGmEpIYZ6TBQPFq6bhzTF5mfMNY1BLJl6QYyG5rI2WC9Ku6Ff8HfAPHXtatkgvpqHFXTW4O7pJwf-kLnY0lEzP7wrGHMaYaT8m-jFkn7Ag_VZGv3W5BB_KPQ2RXjm_DRHpGl2KIf54RY56t8v4-jBPyPdP55uzz9X6-uLybLWuvDCyVK7RYIwC5bkEr1pA1wrWCykZMKU6gwZ013Xc-4aLTgnwzCOvO9noppWtOCEf9r23afw1YS52mG_E3c5FHKdsuRJacCahntH3T9CbcUpxvs5yrUDoGjSbKb6nfBpzTtjb2xQGl-4tMPtgwP4zYB8M2IOBOfTuUD21A3aPkf8vn4G3eyAg4uPaaG4UE-IvL4SItQ</recordid><startdate>20230201</startdate><enddate>20230201</enddate><creator>Huijben, Iris A. M.</creator><creator>Kool, Wouter</creator><creator>Paulus, Max B.</creator><creator>van Sloun, Ruud J. G.</creator><general>IEEE</general><general>The Institute of Electrical and Electronics Engineers, Inc. (IEEE)</general><scope>97E</scope><scope>RIA</scope><scope>RIE</scope><scope>NPM</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7SC</scope><scope>7SP</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope><scope>7X8</scope><orcidid>https://orcid.org/0000-0002-1837-1454</orcidid><orcidid>https://orcid.org/0000-0002-2629-3898</orcidid><orcidid>https://orcid.org/0000-0003-2845-0495</orcidid></search><sort><creationdate>20230201</creationdate><title>A Review of the Gumbel-max Trick and its Extensions for Discrete Stochasticity in Machine Learning</title><author>Huijben, Iris A. M. ; Kool, Wouter ; Paulus, Max B. ; van Sloun, Ruud J. G.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c395t-a87199616c251c6b1eab30f35501066d9e917ddd2cc823d631c0ce24d5878b5b3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Algorithms</topic><topic>Back propagation</topic><topic>Back propagation networks</topic><topic>categorical distribution</topic><topic>Computational modeling</topic><topic>Data models</topic><topic>Deep learning</topic><topic>gradient estimation</topic><topic>Gumbel-max trick</topic><topic>gumbel-softmax</topic><topic>Laplace equations</topic><topic>Machine learning</topic><topic>Neural networks</topic><topic>Optimization</topic><topic>Random variables</topic><topic>sampling</topic><topic>Sketches</topic><topic>Stochastic processes</topic><topic>structured models</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Huijben, Iris A. M.</creatorcontrib><creatorcontrib>Kool, Wouter</creatorcontrib><creatorcontrib>Paulus, Max B.</creatorcontrib><creatorcontrib>van Sloun, Ruud J. G.</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005-present</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998-Present</collection><collection>IEEE Electronic Library (IEL)</collection><collection>PubMed</collection><collection>CrossRef</collection><collection>Computer and Information Systems Abstracts</collection><collection>Electronics & Communications Abstracts</collection><collection>Technology Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><collection>MEDLINE - Academic</collection><jtitle>IEEE transactions on pattern analysis and machine intelligence</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Huijben, Iris A. M.</au><au>Kool, Wouter</au><au>Paulus, Max B.</au><au>van Sloun, Ruud J. G.</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>A Review of the Gumbel-max Trick and its Extensions for Discrete Stochasticity in Machine Learning</atitle><jtitle>IEEE transactions on pattern analysis and machine intelligence</jtitle><stitle>TPAMI</stitle><addtitle>IEEE Trans Pattern Anal Mach Intell</addtitle><date>2023-02-01</date><risdate>2023</risdate><volume>45</volume><issue>2</issue><spage>1353</spage><epage>1371</epage><pages>1353-1371</pages><issn>0162-8828</issn><eissn>1939-3539</eissn><eissn>2160-9292</eissn><coden>ITPIDJ</coden><abstract>The Gumbel-max trick is a method to draw a sample from a categorical distribution, given by its unnormalized (log-)probabilities. Over the past years, the machine learning community has proposed several extensions of this trick to facilitate, e.g., drawing multiple samples, sampling from structured domains, or gradient estimation for error backpropagation in neural network optimization. The goal of this survey article is to present background about the Gumbel-max trick, and to provide a structured overview of its extensions to ease algorithm selection. Moreover, it presents a comprehensive outline of (machine learning) literature in which Gumbel-based algorithms have been leveraged, reviews commonly-made design choices, and sketches a future perspective.</abstract><cop>United States</cop><pub>IEEE</pub><pmid>35254975</pmid><doi>10.1109/TPAMI.2022.3157042</doi><tpages>19</tpages><orcidid>https://orcid.org/0000-0002-1837-1454</orcidid><orcidid>https://orcid.org/0000-0002-2629-3898</orcidid><orcidid>https://orcid.org/0000-0003-2845-0495</orcidid><oa>free_for_read</oa></addata></record> |
| fulltext | fulltext_linktorsrc |
| identifier | ISSN: 0162-8828 |
| ispartof | IEEE transactions on pattern analysis and machine intelligence, 2023-02, Vol.45 (2), p.1353-1371 |
| issn | 0162-8828 1939-3539 2160-9292 |
| language | eng |
| recordid | cdi_proquest_miscellaneous_2637320514 |
| source | IEEE Electronic Library (IEL) |
| subjects | Algorithms Back propagation Back propagation networks categorical distribution Computational modeling Data models Deep learning gradient estimation Gumbel-max trick gumbel-softmax Laplace equations Machine learning Neural networks Optimization Random variables sampling Sketches Stochastic processes structured models |
| title | A Review of the Gumbel-max Trick and its Extensions for Discrete Stochasticity in Machine Learning |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-08T21%3A09%3A16IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_RIE&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=A%20Review%20of%20the%20Gumbel-max%20Trick%20and%20its%20Extensions%20for%20Discrete%20Stochasticity%20in%20Machine%20Learning&rft.jtitle=IEEE%20transactions%20on%20pattern%20analysis%20and%20machine%20intelligence&rft.au=Huijben,%20Iris%20A.%20M.&rft.date=2023-02-01&rft.volume=45&rft.issue=2&rft.spage=1353&rft.epage=1371&rft.pages=1353-1371&rft.issn=0162-8828&rft.eissn=1939-3539&rft.coden=ITPIDJ&rft_id=info:doi/10.1109/TPAMI.2022.3157042&rft_dat=%3Cproquest_RIE%3E2637320514%3C/proquest_RIE%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2761374170&rft_id=info:pmid/35254975&rft_ieee_id=9729603&rfr_iscdi=true |