Saved in:
Bibliographic Details
Main Authors: Thompson, Brian, Dhaliwal, Mehak Preet, Frisch, Peter, Domhan, Tobias, Federico, Marcello
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2401.05749
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929375706873856
author Thompson, Brian
Dhaliwal, Mehak Preet
Frisch, Peter
Domhan, Tobias
Federico, Marcello
author_facet Thompson, Brian
Dhaliwal, Mehak Preet
Frisch, Peter
Domhan, Tobias
Federico, Marcello
contents We show that content on the web is often translated into many languages, and the low quality of these multi-way translations indicates they were likely created using Machine Translation (MT). Multi-way parallel, machine generated content not only dominates the translations in lower resource languages; it also constitutes a large fraction of the total web content in those languages. We also find evidence of a selection bias in the type of content which is translated into many languages, consistent with low quality English content being translated en masse into many lower resource languages, via MT. Our work raises serious concerns about training models such as multilingual large language models on both monolingual and bilingual data scraped from the web.
format Preprint
id arxiv_https___arxiv_org_abs_2401_05749
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle A Shocking Amount of the Web is Machine Translated: Insights from Multi-Way Parallelism
Thompson, Brian
Dhaliwal, Mehak Preet
Frisch, Peter
Domhan, Tobias
Federico, Marcello
Computation and Language
Artificial Intelligence
We show that content on the web is often translated into many languages, and the low quality of these multi-way translations indicates they were likely created using Machine Translation (MT). Multi-way parallel, machine generated content not only dominates the translations in lower resource languages; it also constitutes a large fraction of the total web content in those languages. We also find evidence of a selection bias in the type of content which is translated into many languages, consistent with low quality English content being translated en masse into many lower resource languages, via MT. Our work raises serious concerns about training models such as multilingual large language models on both monolingual and bilingual data scraped from the web.
title A Shocking Amount of the Web is Machine Translated: Insights from Multi-Way Parallelism
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2401.05749