How/Why does Google return a result when Gibberish input is given using another language's keyboard?

by Kyle   Last Updated January 09, 2018 01:04 AM

I'm studying Russian. As such, my laptop (Lenovo T460 thinkpad if it matters) is set up to switch between English and Russian when I hit alt shift.

Well, being the human I am, I occasionally don't think to switch back to English. So, in the searchbar, I'll hit the keys for "Y, O, U, T, U, B, E". Only, my Cyrillic keyboard is still enabled. So, instead of English, the real search is "нщгегиу". (н shares a key with y, щ shares a key with o, and so on.)

Instead of returning gibberish like you'd think, Google (Chrome) still returns YouTube. Screenshot

This isn't limited to Youtube, I've found it working for all of my searches. Does google search using your default computer language? Does it do some weird fuzzy logic to figure out what's more likely? It seems silly to set up dozens of Hidden Markov Models to check the language each and every search...

Answers 1

You'd have to work at Google to get a real understanding as to how they do this, but the why? Well...

Because every letter in another language is mapped to the keyboard, and the biggest keyboard standard layout in existence is QWERTY. So it doesn't matter which language you input, somewhere along the line it's defaulting back to common searches based on those keystrokes. There's most likely a layer that interprets keystrokes rather than actual character data.

I do the same with Korean, especially if I'm too lazy to change virtual keyboard settings.

January 09, 2018 00:49 AM

Related Questions

Ranking is USA and UK

Updated July 29, 2017 20:04 PM

hreflang with session based language detection

Updated October 25, 2017 11:04 AM