Impact of Corpora Quality on Neural Machine Translation
2018-10-19Code Available0· sign in to hype
Matīss Rikters
Code Available — Be the first to reproduce this paper.
ReproduceCode
- github.com/M4t1ss/parallel-corpora-toolsOfficialIn papernone★ 0
Abstract
Large parallel corpora that are automatically obtained from the web, documents or elsewhere often exhibit many corrupted parts that are bound to negatively affect the quality of the systems and models that learn from these corpora. This paper describes frequent problems found in data and such data affects neural machine translation systems, as well as how to identify and deal with them. The solutions are summarised in a set of scripts that remove problematic sentences from input corpora.
Tasks
Benchmark Results
| Dataset | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| WMT 2017 English-Latvian | Transformer trained on highly filtered data | BLEU | 22.89 | — | Unverified |
| WMT 2017 Latvian-English | Transformer trained on highly filtered data | BLEU | 24.37 | — | Unverified |
| WMT 2018 English-Finnish | Transformer trained on highly filtered data | BLEU | 17.4 | — | Unverified |
| WMT 2018 Finnish-English | Transformer trained on highly filtered data | BLEU | 24 | — | Unverified |