Undirected Training of Run Transferable Libraries
This paper investigates the robustness of Run Transferable Libraries(RTLs) on scaled problems. RTLs provide GP with a library of functions which replace the usual primitive functions provided when approaching a problem. The RTL evolves from run to run using feedback based on function usage, and has...
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Tagungsbericht |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 370 |
|---|---|
| container_issue | |
| container_start_page | 361 |
| container_title | |
| container_volume | |
| creator | Keijzer, Maarten Ryan, Conor Murphy, Gearoid Cattolico, Mike |
| description | This paper investigates the robustness of Run Transferable Libraries(RTLs) on scaled problems. RTLs provide GP with a library of functions which replace the usual primitive functions provided when approaching a problem. The RTL evolves from run to run using feedback based on function usage, and has been shown to outperform GP by an order of magnitude on a variety of scalable problems.
RTLs can, however, also be applied across a domain of related problems, as well as across a range of scaled instances of a single problem. To do this successfully, it will need to balance a range of functions. We introduce a problem that can deceive the system into converging to a sub-optimal set of functions, and demonstrate that this is a consequence of the greediness of the library update algorithm.
We demonstrate that a much simpler, truly evolutionary, update strategy doesn’t suffer from this problem, and exhibits far better optimization properties than the original strategy. |
| doi_str_mv | 10.1007/978-3-540-31989-4_33 |
| format | Conference Proceeding |
| fullrecord | <record><control><sourceid>pascalfrancis_sprin</sourceid><recordid>TN_cdi_pascalfrancis_primary_17026811</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>17026811</sourcerecordid><originalsourceid>FETCH-LOGICAL-p228t-835a0345d193c2975d0a94db8d0d23a6f1cadae05310839e67d6e4c866bdd5663</originalsourceid><addsrcrecordid>eNotkEtLQzEQheMLrLX_wMXduIzOZPJciviCgiDtOuTe5JZovS1JXfjvvW2dzTDnHIbDx9gNwh0CmHtnLCeuJHBCZx2XnuiEXdGoHAQ8ZRPUiJxIurOjIZQkrc_ZBAgEd0bSJZvV-gnjEBqHdsJwOcRcUrdLsVmUkIc8rJpN33z8DPt7qH0qoV2nZp7bEkpO9Zpd9GFd0-x_T9ny-Wnx-Mrn7y9vjw9zvhXC7rglFYCkiuioE86oCMHJ2NoIUVDQPXYhhgSKECy5pE3USXZW6zZGpTVN2e3x7zbULqz7sUyXq9-W_B3Kr0cDQlvEMSeOuTpawyoV3242X9Uj-D05P5Lz5Ecc_gDK78nRH7fgWww</addsrcrecordid><sourcetype>Index Database</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype></control><display><type>conference_proceeding</type><title>Undirected Training of Run Transferable Libraries</title><source>Springer Books</source><creator>Keijzer, Maarten ; Ryan, Conor ; Murphy, Gearoid ; Cattolico, Mike</creator><contributor>Collet, Pierre ; Tettamanzi, Andrea ; van Hemert, Jano ; Keijzer, Maarten ; Tomassini, Marco</contributor><creatorcontrib>Keijzer, Maarten ; Ryan, Conor ; Murphy, Gearoid ; Cattolico, Mike ; Collet, Pierre ; Tettamanzi, Andrea ; van Hemert, Jano ; Keijzer, Maarten ; Tomassini, Marco</creatorcontrib><description>This paper investigates the robustness of Run Transferable Libraries(RTLs) on scaled problems. RTLs provide GP with a library of functions which replace the usual primitive functions provided when approaching a problem. The RTL evolves from run to run using feedback based on function usage, and has been shown to outperform GP by an order of magnitude on a variety of scalable problems.
RTLs can, however, also be applied across a domain of related problems, as well as across a range of scaled instances of a single problem. To do this successfully, it will need to balance a range of functions. We introduce a problem that can deceive the system into converging to a sub-optimal set of functions, and demonstrate that this is a consequence of the greediness of the library update algorithm.
We demonstrate that a much simpler, truly evolutionary, update strategy doesn’t suffer from this problem, and exhibits far better optimization properties than the original strategy.</description><identifier>ISSN: 0302-9743</identifier><identifier>ISBN: 3540254366</identifier><identifier>ISBN: 9783540254362</identifier><identifier>EISSN: 1611-3349</identifier><identifier>EISBN: 3540319891</identifier><identifier>EISBN: 9783540319894</identifier><identifier>DOI: 10.1007/978-3-540-31989-4_33</identifier><language>eng</language><publisher>Berlin, Heidelberg: Springer Berlin Heidelberg</publisher><subject>Algorithmics. Computability. Computer arithmetics ; Applied sciences ; Computer science; control theory; systems ; Exact sciences and technology ; Genetic Program ; Genetic Program System ; Library Content ; Multiplexer Problem ; Problem Instance ; Theoretical computing</subject><ispartof>Genetic Programming, 2005, p.361-370</ispartof><rights>Springer-Verlag Berlin Heidelberg 2005</rights><rights>2005 INIST-CNRS</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://link.springer.com/content/pdf/10.1007/978-3-540-31989-4_33$$EPDF$$P50$$Gspringer$$H</linktopdf><linktohtml>$$Uhttps://link.springer.com/10.1007/978-3-540-31989-4_33$$EHTML$$P50$$Gspringer$$H</linktohtml><link.rule.ids>309,310,779,780,784,789,790,793,4050,4051,27925,38255,41442,42511</link.rule.ids><backlink>$$Uhttp://pascal-francis.inist.fr/vibad/index.php?action=getRecordDetail&idt=17026811$$DView record in Pascal Francis$$Hfree_for_read</backlink></links><search><contributor>Collet, Pierre</contributor><contributor>Tettamanzi, Andrea</contributor><contributor>van Hemert, Jano</contributor><contributor>Keijzer, Maarten</contributor><contributor>Tomassini, Marco</contributor><creatorcontrib>Keijzer, Maarten</creatorcontrib><creatorcontrib>Ryan, Conor</creatorcontrib><creatorcontrib>Murphy, Gearoid</creatorcontrib><creatorcontrib>Cattolico, Mike</creatorcontrib><title>Undirected Training of Run Transferable Libraries</title><title>Genetic Programming</title><description>This paper investigates the robustness of Run Transferable Libraries(RTLs) on scaled problems. RTLs provide GP with a library of functions which replace the usual primitive functions provided when approaching a problem. The RTL evolves from run to run using feedback based on function usage, and has been shown to outperform GP by an order of magnitude on a variety of scalable problems.
RTLs can, however, also be applied across a domain of related problems, as well as across a range of scaled instances of a single problem. To do this successfully, it will need to balance a range of functions. We introduce a problem that can deceive the system into converging to a sub-optimal set of functions, and demonstrate that this is a consequence of the greediness of the library update algorithm.
We demonstrate that a much simpler, truly evolutionary, update strategy doesn’t suffer from this problem, and exhibits far better optimization properties than the original strategy.</description><subject>Algorithmics. Computability. Computer arithmetics</subject><subject>Applied sciences</subject><subject>Computer science; control theory; systems</subject><subject>Exact sciences and technology</subject><subject>Genetic Program</subject><subject>Genetic Program System</subject><subject>Library Content</subject><subject>Multiplexer Problem</subject><subject>Problem Instance</subject><subject>Theoretical computing</subject><issn>0302-9743</issn><issn>1611-3349</issn><isbn>3540254366</isbn><isbn>9783540254362</isbn><isbn>3540319891</isbn><isbn>9783540319894</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>2005</creationdate><recordtype>conference_proceeding</recordtype><recordid>eNotkEtLQzEQheMLrLX_wMXduIzOZPJciviCgiDtOuTe5JZovS1JXfjvvW2dzTDnHIbDx9gNwh0CmHtnLCeuJHBCZx2XnuiEXdGoHAQ8ZRPUiJxIurOjIZQkrc_ZBAgEd0bSJZvV-gnjEBqHdsJwOcRcUrdLsVmUkIc8rJpN33z8DPt7qH0qoV2nZp7bEkpO9Zpd9GFd0-x_T9ny-Wnx-Mrn7y9vjw9zvhXC7rglFYCkiuioE86oCMHJ2NoIUVDQPXYhhgSKECy5pE3USXZW6zZGpTVN2e3x7zbULqz7sUyXq9-W_B3Kr0cDQlvEMSeOuTpawyoV3242X9Uj-D05P5Lz5Ecc_gDK78nRH7fgWww</recordid><startdate>2005</startdate><enddate>2005</enddate><creator>Keijzer, Maarten</creator><creator>Ryan, Conor</creator><creator>Murphy, Gearoid</creator><creator>Cattolico, Mike</creator><general>Springer Berlin Heidelberg</general><general>Springer</general><scope>IQODW</scope></search><sort><creationdate>2005</creationdate><title>Undirected Training of Run Transferable Libraries</title><author>Keijzer, Maarten ; Ryan, Conor ; Murphy, Gearoid ; Cattolico, Mike</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-p228t-835a0345d193c2975d0a94db8d0d23a6f1cadae05310839e67d6e4c866bdd5663</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>2005</creationdate><topic>Algorithmics. Computability. Computer arithmetics</topic><topic>Applied sciences</topic><topic>Computer science; control theory; systems</topic><topic>Exact sciences and technology</topic><topic>Genetic Program</topic><topic>Genetic Program System</topic><topic>Library Content</topic><topic>Multiplexer Problem</topic><topic>Problem Instance</topic><topic>Theoretical computing</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Keijzer, Maarten</creatorcontrib><creatorcontrib>Ryan, Conor</creatorcontrib><creatorcontrib>Murphy, Gearoid</creatorcontrib><creatorcontrib>Cattolico, Mike</creatorcontrib><collection>Pascal-Francis</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Keijzer, Maarten</au><au>Ryan, Conor</au><au>Murphy, Gearoid</au><au>Cattolico, Mike</au><au>Collet, Pierre</au><au>Tettamanzi, Andrea</au><au>van Hemert, Jano</au><au>Keijzer, Maarten</au><au>Tomassini, Marco</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>Undirected Training of Run Transferable Libraries</atitle><btitle>Genetic Programming</btitle><date>2005</date><risdate>2005</risdate><spage>361</spage><epage>370</epage><pages>361-370</pages><issn>0302-9743</issn><eissn>1611-3349</eissn><isbn>3540254366</isbn><isbn>9783540254362</isbn><eisbn>3540319891</eisbn><eisbn>9783540319894</eisbn><abstract>This paper investigates the robustness of Run Transferable Libraries(RTLs) on scaled problems. RTLs provide GP with a library of functions which replace the usual primitive functions provided when approaching a problem. The RTL evolves from run to run using feedback based on function usage, and has been shown to outperform GP by an order of magnitude on a variety of scalable problems.
RTLs can, however, also be applied across a domain of related problems, as well as across a range of scaled instances of a single problem. To do this successfully, it will need to balance a range of functions. We introduce a problem that can deceive the system into converging to a sub-optimal set of functions, and demonstrate that this is a consequence of the greediness of the library update algorithm.
We demonstrate that a much simpler, truly evolutionary, update strategy doesn’t suffer from this problem, and exhibits far better optimization properties than the original strategy.</abstract><cop>Berlin, Heidelberg</cop><pub>Springer Berlin Heidelberg</pub><doi>10.1007/978-3-540-31989-4_33</doi><tpages>10</tpages></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 0302-9743 |
| ispartof | Genetic Programming, 2005, p.361-370 |
| issn | 0302-9743 1611-3349 |
| language | eng |
| recordid | cdi_pascalfrancis_primary_17026811 |
| source | Springer Books |
| subjects | Algorithmics. Computability. Computer arithmetics Applied sciences Computer science control theory systems Exact sciences and technology Genetic Program Genetic Program System Library Content Multiplexer Problem Problem Instance Theoretical computing |
| title | Undirected Training of Run Transferable Libraries |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-05T16%3A56%3A47IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-pascalfrancis_sprin&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=Undirected%20Training%20of%20Run%20Transferable%20Libraries&rft.btitle=Genetic%20Programming&rft.au=Keijzer,%20Maarten&rft.date=2005&rft.spage=361&rft.epage=370&rft.pages=361-370&rft.issn=0302-9743&rft.eissn=1611-3349&rft.isbn=3540254366&rft.isbn_list=9783540254362&rft_id=info:doi/10.1007/978-3-540-31989-4_33&rft_dat=%3Cpascalfrancis_sprin%3E17026811%3C/pascalfrancis_sprin%3E%3Curl%3E%3C/url%3E&rft.eisbn=3540319891&rft.eisbn_list=9783540319894&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |