Towards Reinforcement Learning for Pivot-based Neural Machine Translation with Non-autoregressive Transformer
This work addresses the inefficiency in pivot-based NMT for low-resource language pairs, but it appears incremental as it applies known RL methods to a specific domain.
The authors tackled the problem of pivot-based neural machine translation (NMT) in low-resource setups, where existing models lack optimization for the source-target task, by proposing a reinforcement learning approach with a non-autoregressive transformer to train an end-to-end integrated model, resulting in an unspecified improvement.
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.