# Auto-Encoding Variational Neural Machine Translation

arXiv: https://arxiv.org/abs/1807.10564 · HTML: https://mlsanity.com/paper/1807.10564/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 27 Jul 2018.

## Built on by (most cited first, 11 of 11)

- [Survey of Low-Resource Machine Translation](https://mlsanity.com/paper/2109.00486/) (2021)
- [Uncertainty in Natural Language Generation: From Theory to Applications](https://mlsanity.com/paper/2307.15703/) (2023)
- [Amortized Variational Inference: A Systematic Review](https://mlsanity.com/paper/2209.10888/) (2022)
- [Mitigating Data Imbalance and Representation Degeneration in Multilingual Machine Translation](https://mlsanity.com/paper/2305.12786/) (2023)
- [Flow-Adapter Architecture for Unsupervised Machine Translation](https://mlsanity.com/paper/2204.12225/) (2022)
- [A Joint Learning Model with Variational Interaction for Multilingual Program Translation](https://mlsanity.com/paper/2408.14515/) (2024)
- [A Conditional Generative Matching Model for Multi-lingual Reply Suggestion](https://mlsanity.com/paper/2109.07046/) (2021)
- [Elastic Interaction Energy-Based Generative Model: Approximation in Feature Space](https://mlsanity.com/paper/2303.10553/) (2023)
- [Shared Latent Space by Both Languages in Non-Autoregressive Neural Machine Translation](https://mlsanity.com/paper/2305.03511/) (2023)
- [A Study of Posterior Stability for Time-Series Latent Diffusion](https://mlsanity.com/paper/2405.14021/) (2024)
- [Boosting Summarization with Normalizing Flows and Aggressive Training](https://mlsanity.com/paper/2311.00588/) (2023)

## Abstract (arXiv)

We present a deep generative model of bilingual sentence pairs for machine translation. The model generates source and target sentences jointly from a shared latent representation and is parameterised by neural networks. We perform efficient training using amortised variational inference and reparameterised gradients. Additionally, we discuss the statistical implications of joint modelling and propose an efficient approximation to maximum a posteriori decoding for fast test-time predictions. We demonstrate the effectiveness of our model in three machine translation scenarios: in-domain training, mixed-domain training, and learning from a mix of gold-standard and synthetic data. Our experiments show consistently that our joint formulation outperforms conditional modelling (i.e. standard neural machine translation) in all such scenarios.