Published in AI

Meta claims its AI has cracked translation

by on11 July 2022

Boffins say what they meant to say was that the claim is lost in translation

Facebook claims that its AI tools have cracked translation and can translate 204 different languages, twice as many as ever before attempted, at a higher quality than previously achieved.

Despite what companies believe, automatic translation is an incredibly low bar and soon one of them is going to be sued for using it.

However, a new paper from artificial intelligence researchers at Facebook's parent company Meta claims to be able to speak rare languages such as the Acehnese people of Indonesia and the Chokwe people of Central and Southern Africa, which have always been hard for computers to translate because they have very little presence online.

Facebook and Meta CEO Mark Zuckerberg hailed the feat, calling AI translation a "superpower", however the researchers and other language experts are not as enthusiastic.

"The paper presents impressive work to push production-level translation quality to 200 languages," said Professor Philipp Koehn of Johns Hopkins University, one of 38 academics and Meta researchers who collaborated on the work.

"There also will be a lot of resources released that allow everybody to use this model and retrain it on their own, fostering research in that area."

Yet although the paper claimed it was "laying the important groundwork towards realising a universal translation system," computer scientists who weren't involved in the project stressed that it was a small step on a long and winding path, with no obvious end in sight.

The paper's central machine learning technique, a model known by the baroque term Sparsely Gated Mixture of Experts, was not in itself new, said Dr Alexandra Birch-Mayne, Reader in Natural Language Processing at the University of Edinburgh.

Its biggest contribution, she said, was pulling together, cleaning and presenting new data on languages which did not appear widely on the internet, the main source of data for machine translation.

"It's an impressive engineering feat. It's not necessarily a breakthrough in terms of the fundamental science," Dr Birch-Mayne told Sky News.

Measuring machine learning is a challenging task and the boffins thought that the metric known as BLEU, the Meta paper improved the quality of translation over the previous state-of-the-art by 44 per cent. But anyone in the translation business will tell you that BLEU is rubbish, even if it is standard practice in natural language processing research to quote BLEU scores.

Other researchers criticised the fact that the paper had been released without peer review, accusing Meta of practising "peer review by media".

Professor Koehn defended the approach, saying it was "common practice in the field… for better or worse" and helped improve the speed of communication of research results.

While AI has shaken up the language translation business, it is really a court case waiting to happen.  The accuracy and ability of AI translations is so low and literal it is only a matter of time before a cost cutting company relies on a AI contract interpretation or a similar legal document which costs them millions.   Another AI translation problem is the inability to understand management bullshit, business or industrial jargon or slang.  

Last modified on 11 July 2022
Rate this item
(0 votes)

Read more about: