Michal Ždanský Siri has been with us for almost three years now. For the first time, Apple introduced the voice assistant together with the iPhone 4S, where it represented one of the main unique functions of the new phone. Apple has come under fire for Siri, mainly due to inaccuracies and poor recognition. Since its introduction, the service has acquired many other functions and sources of information that Siri can work with, however, it is still a far from ideal technology, which also supports only a handful of languages, among which you will not find Czech.
The backend for Siri, namely the part that takes care of speech recognition and conversion to text, was provided by Nuance Communications, a market leader in its field. Despite the long-standing collaboration, Apple is likely planning to create its own team to develop similar technology that would be faster and more accurate than Nuance's current implementation.
Rumors of replacing Nuance with its own solution have been around since 2011, when Apple hired a number of key employees who could form a new speech recognition team. Already in 2012, he hired the co-founder of the Amazon V9 search engine, who is in charge of the entire Siri project. However, the biggest wave of recruitment came a year later. Among them was, for example, Alex Acero, a former Microsoft employee working on a speech recognition project that could quite possibly be the forerunner of Cortana, the new voice assistant in Windows Phone. Another personality is Lary Gillick, former VP of research at Nuance, who currently holds the title of Siri's Lead Speech Researcher.
Between 2012 and 2013, Apple was supposed to hire additional workers, some of whom are former Nuance employees. Apple is to concentrate these workers in its offices in the American state of Massachusetts, specifically in the cities of Boston and Cambridge, where the new voice recognition engine is to be created. The Boston team is reportedly led by Gunnar Evermann, a former Siri project manager.
We can't expect to see Apple's own engine when iOS 8 is released. Apple will likely replace Nunace technology quietly in future updates to the operating system. However, in iOS 8 we will see one pleasant new feature in speech recognition – support for multiple languages for dictation, including Czech. If Apple does indeed replace Naunce with its own solution, let's hope the transition goes better than when introducing its own maps. However, co-founder Sir Norman Winarsky sees any change positively, according to a quote from a 2011 interview: "In theory, if better voice recognition comes along (or Apple buys it), they'll probably be able to replace Nuance without too much trouble."
Chap, is it good that ios8 will bring a fully integrated path in Siri?
No. Siri will not know Czech. Text can be dictated in Czech where the keyboard pops up.
Aya-chspu-well, nothing much, but at least something. Better than a Tiantung in the eye…..
but he and it's great news. The most difficult for languages, including Czech, is the conversion of sound to text. The question of full support is just a question of a few (hundreds) of regular expressions (or what they do on the servers). Siri could theoretically be able to speak Czech for the next few months (like out of nowhere they announced Japanese or whatever) or a year... Depending on the laziness of the Apple developers.
otherwise, I don't know what will replace the backend... just so it doesn't end up like maps…
I'll just add .. and where the internet will be available :)
I wonder who provides the backend for dictation? Also Nuance, or is it somehow handled by Apple?