WhatsApp and artificial intelligence
Language detection
Language detection is the task of working out automatically which language a text is written in. In a WhatsApp assistant it decides which language to reply in, and what it finds hardest is telling closely related languages apart or handling sentences that mix two languages.
Also known aslanguage identification · language recognition · language guessing · detección de idioma
How it is used in a residence
In a residence with international students the language changes from one conversation to the next, and sometimes within the same one. The classic technique compares the text with a profile of each language, made of the letter sequences and short words that occur most often in it; today a language model handles it on the fly. The hard part is the borderline cases:
- Closely related languages: Spanish, Portuguese and Italian share many words, and telling similar languages apart is one of the great bottlenecks of these systems.
- Mixed messages: "hello, ¿a qué hora es la cena?" carries two languages in a single line.
- Messages with no signal: an "ok", an emoji or a room number say nothing about the writer's language.
Worth knowing
What people take for granted: that it is enough to set each resident's language when they are registered. A German student may write in Spanish to practise and switch to German when something is urgent. Replying in the language of the last message follows the person; fixing it in advance pigeonholes them. And if the reply comes out in another language by mistake, a safety net is needed: machine translation.
Sources
Related terms
- Speech-to-textSpeech-to-text, or audio transcription, is the automatic conversion of spoken words into written text, also known as speech recognition. On WhatsApp it lets an assistant understand a voice note as if the person had typed it, with the risk of mishearing a number or a name.
- Machine translationMachine translation is the use of software to turn text or speech in one language into another without a human translator. Today it is mostly done by neural systems and language models, which translate fluently but can go wrong with proper names or with words that have several meanings.