Bitte benutzen Sie diese Referenz, um auf diese Ressource zu verweisen:
Volltext verfügbar? / Dokumentlieferung
doi:10.22028/D291-38654
Titel: | Does the Order of Training Samples Matter? Improving Neural Data-to-Text Generation with Curriculum Learning |
VerfasserIn: | Chang, Ernie Yeh, Hui-Syuan Demberg, Vera |
Sprache: | Englisch |
Verlag/Plattform: | arXiv |
Erscheinungsjahr: | 2021 |
DDC-Sachgruppe: | 400 Sprache, Linguistik |
Dokumenttyp: | Sonstiges |
Abstract: | Recent advancements in data-to-text generation largely take on the form of neural end-to-end systems. Efforts have been dedicated to improving text generation systems by changing the order of training samples in a process known as curriculum learning. Past research on sequence-to-sequence learning showed that curriculum learning helps to improve both the performance and convergence speed. In this work, we delve into the same idea surrounding the training samples consisting of structured data and text pairs, where at each update, the curriculum framework selects training samples based on the model's competence. Specifically, we experiment with various difficulty metrics and put forward a soft edit distance metric for ranking training samples. Our benchmarks show faster convergence speed where training time is reduced by 38.7% and performance is boosted by 4.84 BLEU. |
URL der Erstveröffentlichung: | https://arxiv.org/abs/2102.03554 |
Link zu diesem Datensatz: | urn:nbn:de:bsz:291--ds-386544 hdl:20.500.11880/34846 http://dx.doi.org/10.22028/D291-38654 |
Datum des Eintrags: | 4-Jan-2023 |
Bemerkung/Hinweis: | Preprint |
Fakultät: | MI - Fakultät für Mathematik und Informatik |
Fachrichtung: | MI - Informatik |
Professur: | MI - Prof. Dr. Vera Demberg |
Sammlung: | SciDok - Der Wissenschaftsserver der Universität des Saarlandes |
Dateien zu diesem Datensatz:
Es gibt keine Dateien zu dieser Ressource.
Alle Ressourcen in diesem Repository sind urheberrechtlich geschützt.