Projects per year
Abstract
The limited availability of in-domain training data is a major issue in the training of application-specific neural machine translation models. Professional outsourcing of bilingual data collections is costly and often not feasible. In this paper we analyze the influence of using crowdsourcing as a scalable way to obtain translations of target in-domain data having in mind that the translations can be of a lower quality. We apply crowdsourcing with carefully designed quality controls to create parallel corpora for the educational domain by collecting translations of texts from MOOCs from English to eleven languages, which we then use to fine-tune neural machine translation models previously trained on general-domain data. The results from our research indicate that crowdsourced data collected with proper quality controls consistently yields performance gains over general-domain baseline systems, and systems fine-tuned with pre-existing in-domain corpora.
Original language | English |
---|---|
Title of host publication | 11th Edition of the Language Resources and Evaluation Conference |
Place of Publication | Miyazaki, Japan |
Publisher | European Language Resources Association (ELRA) |
Pages | 3343-3347 |
Number of pages | 5 |
ISBN (Electronic) | 979-10-95546-00-9 |
Publication status | E-pub ahead of print - 12 May 2018 |
Event | 11th Edition of the Language Resources and Evaluation Conference - Miyazaki, Japan Duration: 7 May 2018 → 12 May 2018 http://lrec2018.lrec-conf.org/en/ |
Conference
Conference | 11th Edition of the Language Resources and Evaluation Conference |
---|---|
Abbreviated title | LREC 2018 |
Country/Territory | Japan |
City | Miyazaki |
Period | 7/05/18 → 12/05/18 |
Internet address |
Fingerprint
Dive into the research topics of 'Improving Machine Translation of Educational Content via Crowdsourcing'. Together they form a unique fingerprint.Projects
- 1 Finished
-
Translation for Massive Open Online Courses- TraMooc
Koehn, P. & Birch-Mayne, A.
1/02/15 → 31/01/18
Project: Research