Multi-agent Learning for Neural Machine Translation

TL;DR

This paper introduces a multi-agent training framework for neural machine translation where diverse agents learn collaboratively, leading to improved translation quality across multiple language pairs and benchmarks.

Contribution

It extends traditional NMT training with multiple interacting agents, enhancing knowledge sharing and translation performance beyond existing single-agent methods.

Findings

Achieved significant improvements over strong baselines.

Demonstrated competitive performance on multiple translation tasks.

Effective multi-agent interaction enhances translation quality.

Abstract

Conventional Neural Machine Translation (NMT) models benefit from the training with an additional agent, e.g., dual learning, and bidirectional decoding with one agent decoding from left to right and the other decoding in the opposite direction. In this paper, we extend the training framework to the multi-agent scenario by introducing diverse agents in an interactive updating process. At training time, each agent learns advanced knowledge from others, and they work together to improve translation quality. Experimental results on NIST Chinese-English, IWSLT 2014 German-English, WMT 2014 English-German and large-scale Chinese-English translation tasks indicate that our approach achieves absolute improvements over the strong baseline systems and shows competitive performance on all tasks.