Interactive-Predictive Neural Machine Translation through Reinforcement and Imitation

TL;DR

This paper introduces an interactive-predictive neural machine translation framework that leverages reinforcement and imitation learning to personalize translations efficiently with minimal human feedback.

Contribution

It presents a novel interactive translation approach that uses weak feedback and expert demonstrations to improve translation quality with less human effort.

Findings

Achieves near-supervised performance with reduced human input

Utilizes weak feedback and imitation learning effectively

Demonstrates promising results on two language pairs

Abstract

We propose an interactive-predictive neural machine translation framework for easier model personalization using reinforcement and imitation learning. During the interactive translation process, the user is asked for feedback on uncertain locations identified by the system. Responses are weak feedback in the form of "keep" and "delete" edits, and expert demonstrations in the form of "substitute" edits. Conditioning on the collected feedback, the system creates alternative translations via constrained beam search. In simulation experiments on two language pairs our systems get close to the performance of supervised training with much less human effort.