mycoreobjectrosdok_document_0000027950rosdok_document_0000027950rosdokdocumentrosdok_derivate_00002298612026-10-13T13:21:18.643Z2026-10-10T09:31:08.340ZeditorKEeditorKEpublished1rosdok_derivate_0000229861truetruederivate_types:fulltextVolltextfulltextwf_edit_epub wf_register_epubstate:publishedstate:publishedveröffentlichtpublishedrosdok_id000058191986180778Oau2026-10-102026-10-13T13:21:16ZrdaConverted from PICA to MODS using Pica2MODS XSLT Transformer 2.14-SNAPSHOT [SCM: "" "" ""] with mode 'DEFAULT'.WorkingpaperAnew approach for cells in multidimensional recurrent neural networksA recent approach for offline handwriting recognition is to use multidimensional recurrent neural networks (MDRNN) with connectionist temporal classification which has shown to yield very good results on several datasets. MDRNNs contain special units – multidimensional Long Short-Term Memory (MDLSTM) cells. These cells suffer from instability especially for higher dimensionality. We analyze the reasons for this effect and introduce several cells with better stability. We present a method to design stable multidimensional cells using the theory of linear shift invariant systems. The new cells are compared to MDLSTMs on the Arabic and French ICDAR datasets, where they yield better results.GundramLeifertVerfasserInautTobiasStrauß1984 -VerfasserInaut11362945540000-0002-5344-4188RogerLabahn1959 -VerfasserInaut1709996450000-0003-1901-9644WelfWustlichVerfasserInauthttps://purl.uni-rostock.de/rosdok/id00005819urn:nbn:de:gbv:28-rosdok_id00005819-610.18453/rosdok_id00005819510 MathematikMathematisch-Naturwissenschaftliche Fakultätalle Rechte vorbehaltenRechteinhaber Universität Rostockfrei zugänglich (Open Access)Lizenz Metadaten: CC0en1 Online-Ressource (22 Seiten)2012University of Rostock, Institute of MathematicsRostockmonographic20122026Universitätsbibliothek RostockRostock2026Universitätsbibliothek Rostockhttps://purl.uni-rostock.de/rosdok/id00005819Gundram Leifert, Tobias Strauß, Roger Labahn, Welf Wustlichrosdok_id000058161986139832Odvhttps://purl.uni-rostock.de/rosdok/id00005816Preprints aus dem Institut für Mathematik3255797-812/40012-0004doctype:epubdoctype:epubDokumenttypDocument typedoctype:epub.workingpaperdoctype:epub.workingpaperWorkingpaperworking paperdiniPublType:workingPaper diniPublType2022:WorkingPaperinfo:eu-repo/semantics/workingPaperdocumentdiniPublType2022:WorkingPaperdiniPublType2022:WorkingPaperArbeitspapierWorking paperKDSF (Pu37)SDNB:510SDNB:510510 Mathematik510 Mathematicsinstitution:unirostockinstitution:unirostockUniversität RostockUniversity of RostockUniversität RostockUniversität RostockUni.Rostockhttp://d-nb.info/gnd/38329-6institution:unirostock.mnfinstitution:unirostock.mnfMathematisch-Naturwissenschaftliche FakultätFaculty of Mathematics and Natural SciencesUniversität Rostock. Mathematisch-Naturwissenschaftliche FakultätMathematisch-Natur-<br />wissenschaftliche FakultätUni.Rostock.Fakultaet.MNFhttp://d-nb.info/gnd/2147083-2licenseinfo:worklicenseinfo:workWerkworklicenseinfo:work.rightsreservedlicenseinfo:work.rightsreservedalle Rechte vorbehaltenall rights reserved/creativecommons/r/reserved/0.9/88x31.png[DE-28]Urheberrechtsschutz 1.0$gRights Statements$uhttp://rightsstatements.org/vocab/InC/1.0/http://rightsstatements.org/vocab/InC/1.0/http://rightsstatements.org/vocab/InC/1.0/licenseinfo:depositlicenseinfo:depositVeröffentlichungsgenehmigungpermission to storelicenseinfo:deposit.rightsunirostocklicenseinfo:deposit.rightsunirostockRechteinhaber Universität Rostockrightsholder Rostock Universityaccesscondition:openaccessaccesscondition:openaccessfrei zugänglich (Open Access)open accesshttp://purl.org/coar/access_right/c_abf2OAfreeinfo:eu-repo/semantics/openAccess[DE-28]Open Access$gControlled Vocabulary for Access Rights$uhttp://purl.org/coar/access_right/c_abf2licenseinfo:metadatalicenseinfo:metadataLizenzen für Metadatenlicenseinfo:metadata.cc0licenseinfo:metadata.cc0Lizenz Metadaten: CC0license metadata: CC0/creativecommons/p/zero/1.0/88x31.pnghttps://creativecommons.org/publicdomain/zero/1.0/rfc5646:enrfc5646:enEnglischEnglishengengAnew approach for cells in multidimensional recurrent neural networksA new approach for cells in multidimensional recurrent neural networksgnd:1136294554orcid:0000-0002-5344-4188gnd:170999645orcid:0000-0003-1901-9644gnd:1136294554orcid:0000-0002-5344-4188gnd:170999645orcid:0000-0003-1901-9644rosdok_document_0000027950-d3752750e50Gundram LeifertGundram Leifertrosdok_document_0000027950-d3752750e60gnd:1136294554orcid:0000-0002-5344-4188Tobias StraußTobias Straußrosdok_document_0000027950-d3752750e76gnd:170999645orcid:0000-0003-1901-9644Roger LabahnRoger Labahnrosdok_document_0000027950-d3752750e92Welf WustlichWelf WustlichGundram LeifertTobias StraußRoger LabahnWelf WustlichGundram LeifertTobias StraußRoger LabahnWelf WustlichGundram LeifertTobias StraußRoger LabahnWelf WustlichRostockUniversity of Rostock, Institute of Mathematicsepub.workingpaperhttps://purl.uni-rostock.de/rosdok/id00005819urn:nbn:de:gbv:28-rosdok_id00005819-610.18453/rosdok_id00005819A recent approach for offline handwriting recognition is to use multidimensional recurrent neural networks (MDRNN) with connectionist temporal classification which has shown to yield very good results on several datasets. MDRNNs contain special units – multidimensional Long Short-Term Memory (MDLSTM) cells. These cells suffer from instability especially for higher dimensionality. We analyze the reasons for this effect and introduce several cells with better stability. We present a method to design stable multidimensional cells using the theory of linear shift invariant systems. The new cells are compared to MDLSTMs on the Arabic and French ICDAR datasets, where they yield better results.20122012Preprints aus dem Institut für Mathematikrosdok_id000058161986139832Odvhttps://purl.uni-rostock.de/rosdok/id00005816Preprints aus dem Institut für Mathematik3255797-812/40012-0004https://purl.uni-rostock.de/rosdok/id000058163255797-8Gundram Leifert, Tobias Strauß, Roger Labahn, Welf Wustlichepub.workingpaper1 Leifert_approach_2012.pdf 248137 5587c650c1cc950ce106d5d7936bde94 rosdok_id000058191986180778Oau2026-10-102026-10-13T13:21:16ZrdaConverted from PICA to MODS using Pica2MODS XSLT Transformer 2.14-SNAPSHOT [SCM: "" "" ""] with mode 'DEFAULT'.WorkingpaperAnew approach for cells in multidimensional recurrent neural networksA recent approach for offline handwriting recognition is to use multidimensional recurrent neural networks (MDRNN) with connectionist temporal classification which has shown to yield very good results on several datasets. MDRNNs contain special units – multidimensional Long Short-Term Memory (MDLSTM) cells. These cells suffer from instability especially for higher dimensionality. We analyze the reasons for this effect and introduce several cells with better stability. We present a method to design stable multidimensional cells using the theory of linear shift invariant systems. The new cells are compared to MDLSTMs on the Arabic and French ICDAR datasets, where they yield better results.GundramLeifertVerfasserInautTobiasStrauß1984 -VerfasserInaut11362945540000-0002-5344-4188RogerLabahn1959 -VerfasserInaut1709996450000-0003-1901-9644WelfWustlichVerfasserInauthttps://purl.uni-rostock.de/rosdok/id00005819urn:nbn:de:gbv:28-rosdok_id00005819-610.18453/rosdok_id00005819510 MathematikMathematisch-Naturwissenschaftliche Fakultätalle Rechte vorbehaltenRechteinhaber Universität Rostockfrei zugänglich (Open Access)Lizenz Metadaten: CC0en1 Online-Ressource (22 Seiten)2012University of Rostock, Institute of MathematicsRostockmonographic20122026Universitätsbibliothek RostockRostock2026Universitätsbibliothek Rostockhttps://purl.uni-rostock.de/rosdok/id00005819Gundram Leifert, Tobias Strauß, Roger Labahn, Welf Wustlichrosdok_id000058161986139832Odvhttps://purl.uni-rostock.de/rosdok/id00005816Preprints aus dem Institut für Mathematik3255797-812/40012-0004 2026-10-10T09:31:08.340Z 2026-10-13T13:21:18.643Z 2026-10-23T13:21:18.653Z editorKE {"identifier":"rosdok/id00005819","type":"local_id","additional":"","service":"MCRLocalID","created":"2026-10-10T09:31:08.354Z"} {"identifier":"urn:nbn:de:gbv:28-rosdok_id00005819-6","type":"dnbUrn","additional":"","service":"RosDokURN","created":"2026-10-10T09:31:08.362Z"} {"identifier":"https://purl.uni-rostock.de/rosdok/id00005819","type":"purl","additional":"","service":"RosDokPURL","created":"2026-10-10T09:31:08.366Z","registrationStarted":"2026-10-10T09:41:31.333Z"} {"identifier":"10.18453/rosdok_id00005819","type":"doi","additional":"","service":"RosDokDOI","created":"2026-10-10T09:31:08.370Z","registrationStarted":"2026-10-10T09:41:31.097Z","registered":"2026-10-10T09:41:32.458Z"} editorKEfulltextfile/rosdok_document_0000027950/rosdok_derivate_0000229861/Leifert_approach_2012.pdfAnew approach for cells in multidimensional recurrent neural networksA new approach for cells in multidimensional recurrent neural networksgnd:1136294554orcid:0000-0002-5344-4188gnd:170999645orcid:0000-0003-1901-9644gnd:1136294554orcid:0000-0002-5344-4188gnd:170999645orcid:0000-0003-1901-9644rosdok_document_0000027950-d3752750e50Gundram LeifertGundram Leifertrosdok_document_0000027950-d3752750e60gnd:1136294554orcid:0000-0002-5344-4188Tobias StraußTobias Straußrosdok_document_0000027950-d3752750e76gnd:170999645orcid:0000-0003-1901-9644Roger LabahnRoger Labahnrosdok_document_0000027950-d3752750e92Welf WustlichWelf WustlichGundram LeifertTobias StraußRoger LabahnWelf WustlichGundram LeifertTobias StraußRoger LabahnWelf WustlichGundram LeifertTobias StraußRoger LabahnWelf WustlichRostockUniversity of Rostock, Institute of Mathematicsepub.workingpaperhttps://purl.uni-rostock.de/rosdok/id00005819urn:nbn:de:gbv:28-rosdok_id00005819-610.18453/rosdok_id00005819A recent approach for offline handwriting recognition is to use multidimensional recurrent neural networks (MDRNN) with connectionist temporal classification which has shown to yield very good results on several datasets. MDRNNs contain special units – multidimensional Long Short-Term Memory (MDLSTM) cells. These cells suffer from instability especially for higher dimensionality. We analyze the reasons for this effect and introduce several cells with better stability. We present a method to design stable multidimensional cells using the theory of linear shift invariant systems. The new cells are compared to MDLSTMs on the Arabic and French ICDAR datasets, where they yield better results.20122012Preprints aus dem Institut für Mathematikrosdok_id000058161986139832Odvhttps://purl.uni-rostock.de/rosdok/id00005816Preprints aus dem Institut für Mathematik3255797-812/40012-0004https://purl.uni-rostock.de/rosdok/id000058163255797-8Gundram Leifert, Tobias Strauß, Roger Labahn, Welf Wustlich[xslt]Saxonrosdok_id00005819https://purl.uni-rostock.de/rosdok/id00005819198618077810.18453/rosdok_id00005819urn:nbn:de:gbv:28-rosdok_id00005819-6Gundram Leifert, Tobias Strauß, Roger Labahn, Welf WustlichLeifert Gundram, Strauß Tobias, Labahn Roger, Wustlich WelfA new approach for cells in multidimensional recurrent neural networksWorkingpaperworking paperUniversity of Rostock, Institute of Mathematics, 2012A recent approach for offline handwriting recognition is to use multidimensional recurrent neural networks (MDRNN) with connectionist temporal classification which has shown to yield very good results on several datasets. MDRNNs contain special units – multidimensional Long Short-Term Memory…rosdok_id000058160012-000412/4Gundram LeifertTobias StraußRoger LabahnWelf WustlichAnew approach for cells in multidimensional recurrent neural networksPreprints aus dem Institut für MathematikUniversitätsbibliothek Rostockhttps://purl.uni-rostock.de/rosdok/id00005819GundramLeifertVerfasserInautTobiasStrauß1984 -VerfasserInaut11362945540000-0002-5344-4188RogerLabahn1959 -VerfasserInaut1709996450000-0003-1901-9644WelfWustlichVerfasserInaut[purl]https://purl.uni-rostock.de/rosdok/id00005819[urn]urn:nbn:de:gbv:28-rosdok_id00005819-6[doi]10.18453/rosdok_id00005819open_access20122012epoch:21th_centuryrfc5646:endoctype:epub.workingpaperaccesscondition:openaccessSDNB:510institution:unirostock.mnfstate:published