Active Learning for Interactive Neural Machine Translation of Data Streams

TL;DR

This paper introduces active learning methods for interactive neural machine translation of continuous data streams, reducing human effort and improving translation quality through novel sample selection strategies based on attention mechanisms.

Contribution

It presents two new active learning techniques leveraging attention information to select source sentences for human validation in streaming translation tasks.

Findings

Reduces human effort in translation streams

Improves translation quality significantly

Outperforms classical sample selection approaches

Abstract

We study the application of active learning techniques to the translation of unbounded data streams via interactive neural machine translation. The main idea is to select, from an unbounded stream of source sentences, those worth to be supervised by a human agent. The user will interactively translate those samples. Once validated, these data is useful for adapting the neural machine translation model. We propose two novel methods for selecting the samples to be validated. We exploit the information from the attention mechanism of a neural machine translation system. Our experiments show that the inclusion of active learning techniques into this pipeline allows to reduce the effort required during the process, while increasing the quality of the translation system. Moreover, it enables to balance the human effort required for achieving a certain translation quality. Moreover, our neural system outperforms classical approaches by a large margin.