The Edit Distance Transducer in Action: The University of Cambridge English-German System at WMT16

TL;DR

This paper explores combining Hiero and neural machine translation using a modified edit distance transducer, demonstrating improved translation quality over traditional lattice rescoring, especially with NMT ensembles.

Contribution

It introduces a novel loose coupling method of Hiero and NMT via an edit distance transducer, enhancing translation performance.

Findings

Loose coupling outperforms lattice rescoring.

Multiple NMT systems improve results.

Modified edit distance transducer effectively combines systems.

Abstract

This paper presents the University of Cambridge submission to WMT16. Motivated by the complementary nature of syntactical machine translation and neural machine translation (NMT), we exploit the synergies of Hiero and NMT in different combination schemes. Starting out with a simple neural lattice rescoring approach, we show that the Hiero lattices are often too narrow for NMT ensembles. Therefore, instead of a hard restriction of the NMT search space to the lattice, we propose to loosely couple NMT and Hiero by composition with a modified version of the edit distance transducer. The loose combination outperforms lattice rescoring, especially when using multiple NMT systems in an ensemble.