Exploiting Deep Representations for Neural Machine Translation

TL;DR

This paper introduces a method to utilize all layers of neural machine translation models through layer aggregation and multi-layer attention, improving translation quality by capturing diverse linguistic information.

Contribution

It proposes a novel approach to exploit information from all encoder and decoder layers using aggregation, multi-layer attention, and regularization, enhancing NMT performance.

Findings

Improved translation accuracy on WMT14 English-German data.

Effective across multiple language pairs.

Demonstrates universality of the approach.

Abstract

Advanced neural machine translation (NMT) models generally implement encoder and decoder as multiple layers, which allows systems to model complex functions and capture complicated linguistic structures. However, only the top layers of encoder and decoder are leveraged in the subsequent process, which misses the opportunity to exploit the useful information embedded in other layers. In this work, we propose to simultaneously expose all of these signals with layer aggregation and multi-layer attention mechanisms. In addition, we introduce an auxiliary regularization term to encourage different layers to capture diverse information. Experimental results on widely-used WMT14 English-German and WMT17 Chinese-English translation data demonstrate the effectiveness and universality of the proposed approach.