# Learning Paraphrastic Sentence Embeddings from Back-Translated Bitext

arXiv: https://arxiv.org/abs/1706.01847 · HTML: https://mlsanity.com/paper/1706.01847/

## Key facts

- Cited by 27 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 6 Jun 2017.

## Built on by (most cited first, 27 of 27)

- [Demystifying Prompts in Language Models via Perplexity Estimation](https://mlsanity.com/paper/2212.04037/) (2022)
- [NL-Augmenter: A Framework for Task-Sensitive Natural Language Augmentation](https://mlsanity.com/paper/2112.02721/) (2021)
- [Bad Characters: Imperceptible NLP Attacks](https://mlsanity.com/paper/2106.09898/) (2021)
- [Generating Syntactically Controlled Paraphrases without Using Annotated Parallel Pairs](https://mlsanity.com/paper/2101.10579/) (2021)
- [A Novel Estimator of Mutual Information for Learning to Disentangle Textual Representations](https://mlsanity.com/paper/2105.02685/) (2021)
- [Source and Target Bidirectional Knowledge Distillation for End-to-end Speech Translation](https://mlsanity.com/paper/2104.06457/) (2021)
- [Paraphrastic Representations at Scale](https://mlsanity.com/paper/2104.15114/) (2021)
- [Data Augmentation with Paraphrase Generation and Entity Extraction for Multimodal Dialogue System](https://mlsanity.com/paper/2205.04006/) (2022)
- [An Unsupervised Method for Building Sentence Simplification Corpora in Multiple Languages](https://mlsanity.com/paper/2109.00165/) (2021)
- [ConRPG: Paraphrase Generation using Contexts as Regularizer](https://mlsanity.com/paper/2109.00363/) (2021)
- [MOVER: Mask, Over-generate and Rank for Hyperbole Generation](https://mlsanity.com/paper/2109.07726/) (2021)
- [Paraphrase Identification with Deep Learning: A Review of Datasets and Methods](https://mlsanity.com/paper/2212.06933/) (2022)
- [Constraint Back-translation Improves Complex Instruction Following of Large Language Models](https://mlsanity.com/paper/2410.24175/) (2024)
- [A Survey of Data Synthesis Approaches](https://mlsanity.com/paper/2407.03672/) (2024)
- [NMTScore: A Multilingual Analysis of Translation-based Text Similarity Measures](https://mlsanity.com/paper/2204.13692/) (2022)
- [Semi-supervised Interactive Intent Labeling](https://mlsanity.com/paper/2104.13406/) (2021)
- [Monotonic Simultaneous Translation with Chunk-wise Reordering and Refinement](https://mlsanity.com/paper/2110.09646/) (2021)
- [Learning to Selectively Learn for Weakly-supervised Paraphrase Generation](https://mlsanity.com/paper/2109.12457/) (2021)
- [What Teaches Robots to Walk, Teaches Them to Trade too -- Regime Adaptive Execution using Informed Data and LLMs](https://mlsanity.com/paper/2406.15508/) (2024)
- [Improving the Diversity of Unsupervised Paraphrasing with Embedding Outputs](https://mlsanity.com/paper/2110.13231/) (2021)
- [Investigating the use of Paraphrase Generation for Question Reformulation in the FRANK QA system](https://mlsanity.com/paper/2206.02737/) (2022)
- [TextAug: Test time Text Augmentation for Multimodal Person Re-identification](https://mlsanity.com/paper/2312.01605/) (2023)
- [NIFTY Financial News Headlines Dataset](https://mlsanity.com/paper/2405.09747/) (2024)
- [Paraphrase Generation as Unsupervised Machine Translation](https://mlsanity.com/paper/2109.02950/) (2021)
- [NL2GDPR: Automatically Develop GDPR Compliant Android Application Features from Natural Language](https://mlsanity.com/paper/2208.13361/) (2022)
- [Sentence Simplification Using Paraphrase Corpus for Initialization](https://mlsanity.com/paper/2305.19754/) (2023)
- [SMCLM: Semantically Meaningful Causal Language Modeling for Autoregressive Paraphrase Generation](https://mlsanity.com/paper/2507.03415/) (2025)

## Abstract (arXiv)

We consider the problem of learning general-purpose, paraphrastic sentence embeddings in the setting of Wieting et al. (2016b). We use neural machine translation to generate sentential paraphrases via back-translation of bilingual sentence pairs. We evaluate the paraphrase pairs by their ability to serve as training data for learning paraphrastic sentence embeddings. We find that the data quality is stronger than prior work based on bitext and on par with manually-written English paraphrase pairs, with the advantage that our approach can scale up to generate large training sets for many languages and domains. We experiment with several language pairs and data sources, and develop a variety of data filtering techniques. In the process, we explore how neural machine translation output differs from human-written sentences, finding clear differences in length, the amount of repetition, and the use of rare words.