A neural network architecture for detecting grammatical errors in statistical machine translation
- Publication type
- A2
- Publication status
- Published
- Authors
- Tezcan, A., Hoste, V., & Macken, L.
- Journal
- PRAGUE BULLETIN OF MATHEMATICAL LINGUISTICS
- Issue
- 108
- Pagination
- 133-145
- Conference
- The 20th Annual Conference of the European Association for Machine Translation (EAMT) (Prague)
- Download
- (.pdf)
- View in Biblio
- (externe link)
Abstract
In this paper we present a Neural Network (NN) architecture for detecting grammatical er- rors in Statistical Machine Translation (SMT) using monolingual morpho-syntactic word rep- resentations in combination with surface and syntactic context windows. We test our approach on two language pairs and two tasks, namely detecting grammatical errors and predicting over- all post-editing e ort. Our results show that this approach is not only able to accurately detect grammatical errors but it also performs well as a quality estimation system for predicting over- all post-editing e ort, which is characterised by all types of MT errors. Furthermore, we show that this approach is portable to other languages.