Can Synthetic Translations Improve Bitext Quality?
Not on the map. Why?
Abstract (arXiv)
Synthetic translations have been used for a wide range of NLP tasks primarily as a means of data augmentation. This work explores, instead, how synthetic translations can be used to revise potentially imperfect reference translations in mined bitext. We find that synthetic samples can improve bitext quality without any additional bilingual supervision when they replace the originals based on a semantic equivalence classifier that helps mitigate NMT noise. The improved quality of the revised bitext is confirmed intrinsically via human evaluation and extrinsically through bilingual induction and MT tasks.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 34 of its 52 references that are in the map. First posted on arXiv 15 Mar 2022.
Builds on
34 of 52 references.
References · 34
- 2017Attention Is All You Needcited by 59,932
- 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
- 2014Adam: A Method for Stochastic Optimizationcited by 36,960
- 2015Neural Machine Translation of Rare Words with Subword Unitscited by 2,898
- 2020mT5: A massively multilingual pre-trained text-to-text transformercited by 1,813
- 2018XNLI: Evaluating Cross-lingual Sentence Representationscited by 913
- 2015Improving Neural Machine Translation Models with Monolingual Datacited by 784
- 2016Sequence-Level Knowledge Distillationcited by 761
- 2017Word Translation Without Parallel Datacited by 480
- 2017Non-Autoregressive Neural Machine Translationcited by 457
- 2018Massively Multilingual Sentence Embeddings for Zero-Shot Cross-Lingual Transfer and Beyondcited by 434
- 2019PAWS-X: A Cross-lingual Adversarial Dataset for Paraphrase Identificationcited by 262
22 more
- 2019Evaluating Gender Bias in Machine Translationcited by 247
- 2021Quality at a Glance: An Audit of Web-Crawled Multilingual Datasetscited by 241
- 2019WikiMatrix: Mining 135M Parallel Sentences in 1620 Language Pairs from Wikipediacited by 221
- 2020SimAlign: High Quality Word Alignments without Parallel Training Data using Static and Contextualized Embeddingscited by 155
- 2018When and Why are Pre-trained Word Embeddings Useful for Neural Machine Translation?cited by 125
- 2019Understanding Knowledge Distillation in Non-autoregressive Machine Translationcited by 121
- 2016Modeling Coverage for Neural Machine Translationcited by 110
- 2020Reducing Gender Bias in Neural Machine Translation as a Domain Adaptation Problemcited by 86
- 2018On the Impact of Various Types of Noise on Neural Machine Translationcited by 76
- 2020Language ID in the Wild: Unexpected Challenges on the Path to a Thousand-Language Web Text Corpuscited by 76
- 2020Mitigating Gender Bias in Machine Translation with Target Gender Annotationscited by 42
- 2018Dual Conditional Cross-Entropy Filtering of Noisy Parallel Corporacited by 40
- 2018Back-Translation Sampling by Targeting Difficult Words in Neural Machine Translationcited by 19
- 2021Bilingual Lexicon Induction via Unsupervised Bitext Construction and Word Alignmentcited by 19
- 2019Controlling Neural Machine Translation Formality with Synthetic Supervisioncited by 18
- 2020Detecting Fine-Grained Cross-Lingual Semantic Divergences without Supervision by Learning to Rankcited by 18
- 2020Data Rejuvenation: Exploiting Inactive Training Examples for Neural Machine Translationcited by 17
- 2019Bilingual Lexicon Induction through Unsupervised Machine Translationcited by 16
- 2021How Does Distilled Data Complexity Impact the Quality and Confidence of Non-Autoregressive Machine Translation?cited by 12
- 2019Lost in Evaluation: Misleading Benchmarks for Bilingual Dictionary Inductioncited by 10
- 2021Beyond Noise: Mitigating the Impact of Fine-grained Semantic Divergences on Neural Machine Translationcited by 9
- 2021BitextEdit: Automatic Bitext Editing for Improved Low-Resource Machine Translationcited by 3
Built on by
Citing papers per year: 2023: 1.
1 paper.