Towards Reinforcement Learning for Pivot-based Neural Machine Translation with Non-autoregressive Transformer

TL;DR

This paper introduces a reinforcement learning approach to train a non-autoregressive transformer for pivot-based neural machine translation, enabling end-to-end source-target translation in low-resource settings.

Contribution

It proposes an integrated RL-based training method for pivot-based NMT using a non-autoregressive transformer, connecting sub-tasks into a unified model.

Findings

Improved translation quality in low-resource language pairs.

End-to-end training enhances source-target translation performance.

Demonstrates effectiveness of RL in pivot-based NMT.

Abstract

Pivot-based neural machine translation (NMT) is commonly used in low-resource setups, especially for translation between non-English language pairs. It benefits from using high resource source-pivot and pivot-target language pairs and an individual system is trained for both sub-tasks. However, these models have no connection during training, and the source-pivot model is not optimized to produce the best translation for the source-target task. In this work, we propose to train a pivot-based NMT system with the reinforcement learning (RL) approach, which has been investigated for various text generation tasks, including machine translation (MT). We utilize a non-autoregressive transformer and present an end-to-end pivot-based integrated model, enabling training on source-target data.