Speech Technology Magazine

 

Star Performer: Apple Is More Siri-ous

By Leonard Klie - Posted Jul 29, 2014
Page1 of 1
Bookmark and Share

Apple ushered in the smartphone era with its first iPhone release in 2007, and then revolutionized the handset industry—and indeed the speech industry—when it added the Siri personal assistant app as a standard feature on the iPhone in 2011.

But Apple soon encountered a handful of other personal assistant applications all created to rival Siri. In response, Apple quietly acquired automatic speech recognition (ASR) technology company Novauris Technologies late last year to help it improve Siri.

Novauris, which spun off from the firm that developed Nuance Communications' Dragon software, had been developing large-vocabulary ASR technology for access to information stored locally on mobile devices or remotely on servers. Its voice recognition products support a wide range of languages, including English, German, French, Japanese, Korean, Spanish, Italian, and Mandarin Chinese.

Both of those elements became key components of a revamped and much-improved Siri that Apple launched in September 2013 on its newest iPhone 5c and 5s models. In the new version, Siri's speech recognition and natural language interpretation is done in the cloud, something that Bill Meisel, president of speech consulting firm TMA Associates, says will allow the voice capabilities to be "constantly improved."

Apple also redesigned Siri's interface, gave it a wider selection of male and female voices, and expanded its global reach with support for 19 languages and regional dialects. The company made the upgraded Siri faster at answering questions and more accurate through improved speech recognition. Users can even change the way Siri pronounces words.

Additionally, Siri searches more locations, including Bing, Wikipedia, and Twitter, for answers to users' queries, and now performs other tasks, such as controlling more device settings, checking for missed calls and returning calls, playing voicemail, posting to social media sites, and controlling iTunes Radio. Users can search Twitter to see what someone is saying; Siri can pull up individual accounts and present the most recent tweets from those accounts. Additionally, users can search for images with Siri, make dinner reservations with an OpenTable integration, and check product and service reviews with an integrated Yelp application.

Meisel also alluded to a new tap-to-edit feature that allows for text entry. This, he says, "would make Siri useful when one can't speak, something necessary if the personal assistant becomes increasingly perceived as a primary user interface."

Jeff Kagan, a wireless technology industry analyst, says the Siri upgrades "are very good," and calls them "a blend of both real improvement and cosmetic changes."

Apple also brought Siri to the car, signing a deal with Chevrolet to place Siri in most of its 2014 models. This deal allows drivers to perform Siri-related tasks using a voice activation button on the steering wheel. Drivers can access stored music and calendars, change radio stations, get directions, send voice-to-text messages, and make calls—all without ever having to touch or even look at their phones. Honda has already brought Siri technology to select Honda and Acura models, including the Honda Accord and Acura RDX and ILX, as a dealer-installed accessory. Other automakers rolling out Siri integrations are BMW, General Motors, Toyota, Mercedes-Benz, Ferrari, and Audi.

And the company's not finished there. It is reportedly assembling a team of A-list speech recognition experts to create a Siri engine based on neural networking, or machine learning algorithms that function much like the neurons in the human brain. This means Siri will be able to understand you, even if you've had a few too many cocktails.


Page1 of 1