Hartmann, Silvana ; Mújdricza-Maydt, Éva ; Kuznetsov, Ilia ; Gurevych, Iryna ; Frank, Anette (2017)
Assessing SRL Frameworks with Automatic Training Data Expansion.
Valencia, Spain
Konferenzveröffentlichung, Bibliographie
Kurzbeschreibung (Abstract)
We present the first experiment-based study that explicitly contrasts the three major semantic role labeling frameworks. As a prerequisite, we create a dataset labeled with parallel FrameNet-, PropBank-, and VerbNet-style labels for German. We train a state-of-the-art SRL tool for German for the different annotation styles and provide a comparative analysis across frameworks. We further explore the behavior of the frameworks with automatic training data generation. VerbNet provides larger semantic expressivity than PropBank, and we find that its generalization capacity approaches PropBank in SRL training, but it benefits less from training data expansion than the sparse-data affected FrameNet.
| Typ des Eintrags: | Konferenzveröffentlichung |
|---|---|
| Erschienen: | 2017 |
| Autor(en): | Hartmann, Silvana ; Mújdricza-Maydt, Éva ; Kuznetsov, Ilia ; Gurevych, Iryna ; Frank, Anette |
| Art des Eintrags: | Bibliographie |
| Titel: | Assessing SRL Frameworks with Automatic Training Data Expansion |
| Sprache: | Englisch |
| Publikationsjahr: | April 2017 |
| Verlag: | Association for Computational Linguistics |
| Buchtitel: | Proceedings of the 11th Linguistics Annotation Workshop (LAW XI) at EACL 2017 |
| Veranstaltungsort: | Valencia, Spain |
| URL / URN: | https://aclweb.org/anthology/W/W17/W17-0814.pdf |
| Kurzbeschreibung (Abstract): | We present the first experiment-based study that explicitly contrasts the three major semantic role labeling frameworks. As a prerequisite, we create a dataset labeled with parallel FrameNet-, PropBank-, and VerbNet-style labels for German. We train a state-of-the-art SRL tool for German for the different annotation styles and provide a comparative analysis across frameworks. We further explore the behavior of the frameworks with automatic training data generation. VerbNet provides larger semantic expressivity than PropBank, and we find that its generalization capacity approaches PropBank in SRL training, but it benefits less from training data expansion than the sparse-data affected FrameNet. |
| Freie Schlagworte: | UKP_reviewed;Natural Language Processing;UKP_p_QAEduInf;UKP_p_CLARIND |
| ID-Nummer: | TUD-CS-2017-0045 |
| Fachbereich(e)/-gebiet(e): | 20 Fachbereich Informatik 20 Fachbereich Informatik > Ubiquitäre Wissensverarbeitung DFG-Graduiertenkollegs DFG-Graduiertenkollegs > Graduiertenkolleg 1994 Adaptive Informationsaufbereitung aus heterogenen Quellen |
| Hinterlegungsdatum: | 23 Feb 2017 11:35 |
| Letzte Änderung: | 24 Jan 2020 12:03 |
| PPN: | |
| Export: | |
| Suche nach Titel in: | TUfind oder in Google |
 |
Frage zum Eintrag |
Optionen (nur für Redakteure)
 |
Redaktionelle Details anzeigen |
Drucken
Impressum