# An Empirical Analysis of NMT-Derived Interlingual Embeddings and their Use in Parallel Sentence Identification

arXiv: https://arxiv.org/abs/1704.05415 · HTML: https://mlsanity.com/paper/1704.05415/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 18 Apr 2017.

## Built on by (most cited first, 9 of 9)

- [SpeechMatrix: A Large-Scale Mined Corpus of Multilingual Speech-to-Speech Translations](https://mlsanity.com/paper/2211.04508/) (2022)
- [Paraphrastic Representations at Scale](https://mlsanity.com/paper/2104.15114/) (2021)
- [Bilingual Lexicon Induction via Unsupervised Bitext Construction and Word Alignment](https://mlsanity.com/paper/2101.00148/) (2021)
- [Unsupervised Multilingual Sentence Embeddings for Parallel Corpus Mining](https://mlsanity.com/paper/2105.10419/) (2021)
- [Lightweight Cross-Lingual Sentence Representation Learning](https://mlsanity.com/paper/2105.13856/) (2021)
- [Beyond Contrastive Learning: A Variational Generative Model for Multilingual Retrieval](https://mlsanity.com/paper/2212.10726/) (2022)
- [Parallel-SFT: Improving Zero-Shot Cross-Programming-Language Transfer for Code RL](https://mlsanity.com/paper/2604.20835/) (2026)
- [EMS: Efficient and Effective Massively Multilingual Sentence Embedding Learning](https://mlsanity.com/paper/2205.15744/) (2022)
- [Translating Under Pressure: Domain-Aware LLMs for Crisis Communication](https://mlsanity.com/paper/2604.26597/) (2026)

## Abstract (arXiv)

End-to-end neural machine translation has overtaken statistical machine translation in terms of translation quality for some language pairs, specially those with large amounts of parallel data. Besides this palpable improvement, neural networks provide several new properties. A single system can be trained to translate between many languages at almost no additional cost other than training time. Furthermore, internal representations learned by the network serve as a new semantic representation of words -or sentences- which, unlike standard word embeddings, are learned in an essentially bilingual or even multilingual context. In view of these properties, the contribution of the present work is two-fold. First, we systematically study the NMT context vectors, i.e. output of the encoder, and their power as an interlingua representation of a sentence. We assess their quality and effectiveness by measuring similarities across translations, as well as semantically related and semantically unrelated sentence pairs. Second, as extrinsic evaluation of the first point, we identify parallel sentences in comparable corpora, obtaining an F1=98.2% on data from a shared task when using only NMT context vectors. Using context vectors jointly with similarity measures F1 reaches 98.9%.