philip lelyveld The world of entertainment technology

Higher education needs to promote cross-disciplinary dialogue about AI

Google Translate's limitations spell out why we must re-visit old questions about artificial intelligence, says Lionel Tarassenko

September 14, 2021 Lionel Tarassenko

Advances in machine learning have been spectacular in the last five years. This form of artificial intelligence has led to very significant progress is areas such as autonomous driving, machine translation (to and from multiple languages) and automated text generation.

And yet… Useful as Google Translate is, it still makes some fairly basic mistakes. For example, it correctly renders “the window that I have shut” as “la fenêtre que j'ai fermée”, but incorrectly translates “the key that I have found” into French as “la clé que j'ai trouvé”.

Anyone with a French A-level will tell you that, with the avoir verb, the past participle must agree with the direct object when it precedes the verb. “Clé” is feminine, so the extra ‘e’ is also needed on the end of the participle. Testing with similar examples gives a phrase translation accuracy of around 50 per cent, which isn’t great.

To someone like me who has been working in machine learning (ML) for the past 30 years, this is not surprising. Translation is only as good as the data fed to the ML algorithm during the learning phase. Google Translate has no understanding of French grammar: it learns through brute repetition of exemplar sequences. Evidently there are not enough examples in Google’s training data of phrases with feminine nouns as objects preceding avoir for the correct translation to be given every time.

Lionel Tarassenko is president of Reuben College, University of Oxford.

See the full story here: https://www.timeshighereducation.com/blog/higher-education-needs-promote-cross-disciplinary-dialogue-about-aihttps://www.timeshighereducation.com/blog/higher-education-needs-promote-cross-disciplinary-dialogue-about-ai