From Speech Strategy News, May 2010
On April 28, Apple
agreed to buy Siri Inc. to gain
technology that lets users do voice searches using natural-language inquiries.
Terms were not disclosed. Siri has raised about $23.8 million in funding in the
past two years from venture capital firms including Menlo Ventures, Morgenthaler
Ventures, and the Li Ka-Shing Foundation.
Siri launched
a free “Do Engine” application for the iPhone, which takes speech or text input
in natural language and attempts to do what is asked (SSN, March 2010, p. 1).
Siri is building variations on a technology developed by SRI International, which several years ago led a $200 million
research project on artificial intelligence funded by the Defense Advanced Research Projects Agency (DARPA), CALO (Cognitive
Assistant that Learns and Organizes). The speech recognition technology
supporting Siri comes from Nuance.
Siri also planned versions for the iPod Touch, iPhone 3G, and possibly additional
mobile platforms, although the last possibility may be changed by the Apple
acquisition.
Adam Cheyer, Co-Founder and VP Engineering, Siri, spoke
at the Mobile Voice Conference before the announcement of the Apple acquisition
and characterized the Siri cloud-based service (with a client software download
for the mobile phone) as “a speech-enabled concierge at your service.” He said
it is a “Virtual
Personal Assistant” that, like a human assistant,
- You
tell it what you want to do;
- It
will help you get things done; and
- It
gets better over time.
The first version of Siri is characterized as an “Out
and About Mobile Entertainment Assistant,” which can understand and can perform
actions in the domains of restaurants, events, taxis, movies, local info (such
as the nearest gas station), and weather. Siri will improve by personalization
and by understanding a broader set of tasks. A future upgrade will allow
handling reminders, flights stats, and reference questions. Every month, Siri
plans to add new partner services and expand what it can do. A slide from
Cheyer’s Mobile Voice presentation summarizes the elements of the service:
