# APE-then-QE: Correcting then Filtering Pseudo Parallel Corpora for MT Training Data Creation

arXiv: https://arxiv.org/abs/2312.11312 · HTML: https://mlsanity.com/paper/2312.11312/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 27 references that are in the map.
- First posted on arXiv 18 Dec 2023.

## Builds on (most cited first, 9 of 9)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Neural Machine Translation by Jointly Learning to Align and Translate](https://mlsanity.com/paper/1409.0473/) (2014)
- [Sequence to Sequence Learning with Neural Networks](https://mlsanity.com/paper/1409.3215/) (2014)
- [A Call for Clarity in Reporting BLEU Scores](https://mlsanity.com/paper/1804.08771/) (2018)
- [Language-agnostic BERT Sentence Embedding](https://mlsanity.com/paper/2007.01852/) (2020)
- [PMIndia -- A Collection of Parallel Corpora of Languages of India](https://mlsanity.com/paper/2001.09907/) (2020)
- [Unbabel's Submission to the WMT2019 APE Shared Task: BERT-based Encoder-Decoder for Automatic Post-Editing](https://mlsanity.com/paper/1905.13068/) (2019)
- [Improving Machine Translation with Phrase Pair Injection and Corpus Filtering](https://mlsanity.com/paper/2301.08008/) (2023)
- [Netmarble AI Center's WMT21 Automatic Post-Editing Shared Task Submission](https://mlsanity.com/paper/2109.06515/) (2021)

## Abstract (arXiv)

Automatic Post-Editing (APE) is the task of automatically identifying and correcting errors in the Machine Translation (MT) outputs. We propose a repair-filter-use methodology that uses an APE system to correct errors on the target side of the MT training data. We select the sentence pairs from the original and corrected sentence pairs based on the quality scores computed using a Quality Estimation (QE) model. To the best of our knowledge, this is a novel adaptation of APE and QE to extract quality parallel corpus from the pseudo-parallel corpus. By training with this filtered corpus, we observe an improvement in the Machine Translation system's performance by 5.64 and 9.91 BLEU points, for English-Marathi and Marathi-English, over the baseline model. The baseline model is the one that is trained on the whole pseudo-parallel corpus. Our work is not limited by the characteristics of English or Marathi languages; and is language pair-agnostic, given the necessary QE and APE data.