An Empirical Accuracy Law for Sequential Machine Translation: the Case of Google Translate
Computation and Language
2020-04-10 v2 Machine Learning
Machine Learning
Abstract
In this research, we have established, through empirical testing, a law that relates the number of translating hops to translation accuracy in sequential machine translation in Google Translate. Both accuracy and size decrease with the number of hops; the former displays a decrease closely following a power law. Such a law allows one to predict the behavior of translation chains that may be built as society increasingly depends on automated devices.
Keywords
Cite
@article{arxiv.2003.02817,
title = {An Empirical Accuracy Law for Sequential Machine Translation: the Case of Google Translate},
author = {Lucas Nunes Sequeira and Bruno Moreschi and Fabio Gagliardi Cozman and Bernardo Fontes},
journal= {arXiv preprint arXiv:2003.02817},
year = {2020}
}
Comments
11 pages, 8 figures (mostly graphs), a few mathematical functions and samples of the experiments