Self-Induced Curriculum Learning in Self-Supervised Neural Machine Translation

Self-supervised neural machine translation (SSNMT) jointly learns to identify and select suitable training data from comparable (rather than parallel) corpora and to translate, in a way that the two tasks support each other in a virtuous circle. In this study, we provide an in-depth analysis of the...

Full description

Saved in:
Bibliographic Details
Main Authors Ruiter, Dana, van Genabith, Josef, España-Bonet, Cristina
Format Journal Article
LanguageEnglish
Published 07.04.2020
Subjects
Computer Science - Computation and Language
Online AccessGet full text

Cover

More Information
Summary:Self-supervised neural machine translation (SSNMT) jointly learns to identify and select suitable training data from comparable (rather than parallel) corpora and to translate, in a way that the two tasks support each other in a virtuous circle. In this study, we provide an in-depth analysis of the sampling choices the SSNMT model makes during training. We show how, without it having been told to do so, the model self-selects samples of increasing (i) complexity and (ii) task-relevance in combination with (iii) performing a denoising curriculum. We observe that the dynamics of the mutual-supervision signals of both system internal representation types are vital for the extraction and translation performance. We show that in terms of the Gunning-Fog Readability index, SSNMT starts extracting and learning from Wikipedia data suitable for high school students and quickly moves towards content suitable for first year undergraduate students.
DOI:10.48550/arxiv.2004.03151