When the iPhone 4S came out, the big deal was the inclusion of Siri, the voice controlled “intelligent personal assistant” which answers your questions, refers people to businesses and places phone calls for you. It was an offshoot of a military project funded by DARPA (Defense Advanced Research Projects Agency, the people who brought us The Internet as we know it, AKA the military equivalent to James Bond’s gadgetmaster Q) to create a system that could control a computer with “natural language”, or regular speech without needing to use specialized commands. When the 4S hit the market, Siri helped give Apple a boost on sales. Google’s Android platform didn’t have a system which gave the flexibility like Siri, even through it could recognize and convert speech directly to text. Siri works with a search engine called Wolfram|Alphawhich takes questions in natural language and then answers them directly, unlike places like Google which return full websites.
Meet Iris, an up-and-coming program designed to give Android the same capabilities as the iPhone 4S.
Iris was made in October and released after only 8 hours of programming. Dexetra, the group that released it, was inspired by the thousands of blog posts and tweets about how wonderful Siri was. From Iris’s creator Narayan Babu: “Suddenly, I got the urge to do something similar for Android. Since we have been working on NLP and Machine learning for over an year now, I had a crazy belief that I could pull this off. Somehow I managed to write a tiny engine that could answer your questions, digging the results from the web.” (link) Even the name was inspired by Siri (Hint: read the name backwards). Tests conducted in our highly secret, hi-tech laboratory high in the Swiss Alps (actually just at my desk here in windy Northampton) have shown that just using very serious questions like “Where can I get my computer fixed in Northampton, MA” and “How many licks does it take to get to the Tootsie Roll center of a Tootsie Pop” produce proper answers like “Northampton Computer Repair” and “It all depends on a couple of factors, like type of Tootsie Pop, amount of Saliva and other factors. The world may never know!” So far Dexetra has been able to produce a great product which provides great answers and is able to control a few other applications on the phone. This allows you to tell the program to play a certain song stored on the phone, or view your contact list just by saying it out loud. Unfortunately it doesn’t seem Iris has the range of funny responses or “self-aware” humor yet.
Another program with the same goal is Evi. It was released on January 23rd by True Knowledge, a search engine that specializes in natural language searches. It came out to such an enthusiastic response that in the past week the company was unable to operate their server which handles the questions. At the time of writing I was unable to test out the functionality of this app since the servers were sill overloaded.
All of these applications work on the same basic principles. The voice recognition function of the smartphone is used to convert the spoken word into text. That text result is sent to the search engine, and the engine returns the answer in plain text for the app to read out loud. It seems the only limitation on this technology is the search engines themselves; as they advance in being able to understand natural language and returning relevant results, the more intelligent and useful these applications become. This technology is akin to the computers in Star Trek where people can talk directly to the machine without using any specific syntax or phrases.
Speaking of Star Trek, Google is looking to make their own version of Siri, named Majel after the voice of the computer on the TV show Star Trek, also the wife of Star Trek creator Gene Roddenberry. Using the full array of Google Apps, the Google X (the top secret lab where all Google’s big project development magic happens (sadly also cooler than my non-secret, non-lab)) aims to produce a stronger, more versatile personal assistant than what Apple’s Siri can provide. Whether they can produce this remains to be seen, but Android users are hopeful.
Used any useful Personal Assisant apps lately? Let us know!
