Bilingual-GAN: A Step Towards Parallel Text Generation

  • 2019-04-09 15:42:08
  • Ahmad Rashid, Alan Do-Omri, Md. Akmal Haidar, Qun Liu, Mehdi Rezagholizadeh
  • 5

Abstract

Latent space based GAN methods and attention based sequence to sequencemodels have achieved impressive results in text generation and unsupervisedmachine translation respectively. Leveraging the two domains, we propose anadversarial latent space based model capable of generating parallel sentencesin two languages concurrently and translating bidirectionally. The bilingualgeneration goal is achieved by sampling from the latent space that is sharedbetween both languages. First two denoising autoencoders are trained, withshared encoders and back-translation to enforce a shared latent state betweenthe two languages. The decoder is shared for the two translation directions.Next, a GAN is trained to generate synthetic "code" mimicking the languages'shared latent space. This code is then fed into the decoder to generate text ineither language. We perform our experiments on Europarl and Multi30k datasets,on the English-French language pair, and document our performance using bothsupervised and unsupervised machine translation.

 

Quick Read (beta)

loading the full paper ...