# Exemplar-Controllable Paraphrasing and Translation using Bitext

arXiv: https://arxiv.org/abs/2010.05856 · HTML: https://mlsanity.com/paper/2010.05856/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Oct 2020.

## Built on by (most cited first, 2 of 2)

- [Fighting Fire with Fire: The Dual Role of LLMs in Crafting and Detecting Elusive Disinformation](https://mlsanity.com/paper/2310.15515/) (2023)
- [Leveraging Natural Supervision for Language Representation Learning and Generation](https://mlsanity.com/paper/2207.10617/) (2022)

## Abstract (arXiv)

Most prior work on exemplar-based syntactically controlled paraphrase generation relies on automatically-constructed large-scale paraphrase datasets, which are costly to create. We sidestep this prerequisite by adapting models from prior work to be able to learn solely from bilingual text (bitext). Despite only using bitext for training, and in near zero-shot conditions, our single proposed model can perform four tasks: controlled paraphrase generation in both languages and controlled machine translation in both language directions. To evaluate these tasks quantitatively, we create three novel evaluation datasets. Our experimental results show that our models achieve competitive results on controlled paraphrase generation and strong performance on controlled machine translation. Analysis shows that our models learn to disentangle semantics and syntax in their latent representations, but still suffer from semantic drift.