Edinburgh Research Explorer

A Latent Morphology Model for Open-Vocabulary Neural Machine Translation

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Related Edinburgh Organisations

Open Access permissions



Original languageEnglish
Title of host publicationEighth International Conference on Learning Representations
Number of pages15
Publication statusAccepted/In press - 20 Dec 2020
EventEighth International Conference on Learning Representations - Millennium Hall, Addis Ababa, Ethiopia
Duration: 26 Apr 202030 Apr 2020


ConferenceEighth International Conference on Learning Representations
Abbreviated titleICLR 2020
CityAddis Ababa
Internet address


Translation into morphologically-rich languages challenges neural machinetranslation (NMT) models with extremely sparse vocabularies where atomic treatment of surface forms is unrealistic. This problem is typically addressed by either pre-processing words into subword units or performing translation directly at the level of characters. The former is based on word segmentation algorithms optimized using corpus-level statistics with no regard to the translation task. The latter learns directly from translation data but requires rather deep architectures. In this paper, we propose to translate words by modeling word formation through a hierarchical latent variable model which mimics the process of morphological inflection. Our model generates words one character at a time by composing two latent representations: a continuous one, aimed at capturing the lexical semantics, and a set of (approximately) discrete features, aimed at capturing the morphosyn-tactic function, which are shared among different surface forms. Our model achieves better accuracy in translation into three morphologically-rich languages than conventional open-vocabulary NMT methods, while also demonstrating a better generalization capacity under low to mid-resource settings.


Eighth International Conference on Learning Representations


Addis Ababa, Ethiopia

Event: Conference

Download statistics

No data available

ID: 138096094