Auto-Encoding Variational Neural Machine Translation

We present a deep generative model of bilingual sentence pairs. The model generates source and target sentences jointly from a shared latent representation and is parameterised by neural networks. Efficient training is done by 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 in all such scenarios.