In Silico Vox: Voice Recognition Moves Forward

Let us sum up the problems with voice recognition in this day and age with a quick quote. Chris Tucker said it best in the movie Rush Hour when he asked Jackie Chan, “DO YOU SPEAK-A ANY ENGLISH?”

Voice recognition technology has always been hampered by the chips that use them. In this case, there still hasn’t been a hardware development that allows for all the necessary software to take the technology forward. This is especially true for mobile phones and PDAs, which have a lot less physical space to put voice recognition software on as opposed to computers.

Luckily for us, there’s a team at Carnegie Mellon University that’s trying to solve these problems. Their solution: putting all the necessary software on a new chip, in much the same way we have graphics cards on computers. At the Hot Chips conference in Stanford, they presented the results of their In Silico Vox project: a working prototype chip that can already recognize 1000 words. They expect to put a lot more in by next year.

What does this mean for the future? A lot more accessibility for people who don’t have the means to type, for one thing, as well as greater ease of use for people on cellphones. Instead of the distracting and potentially deadly habit of scrolling through your phone while driving, this should allow for people to tell their phones what to do. 

More importantly for the world, once we find a way to transform normal people’s clothing into spandex suits that increase its wearer’s natural abilities and serve as protective armor, we can combine the technology with the result of the In Silico Vox project to let this new technology be accessed by some kind of special morphing technology. This mighty technology will allow those men and women to morph into people of great power, and it would be best if we got some Army Rangers to use it for the good of mankind! Go Go Army Rangers!

Via Ars Technica

Let us sum up the problems with voice recognition in this day and age with a quick quote. Chris Tucker said it best in the movie Rush Hour when he asked Jackie Chan, “DO YOU SPEAK-A ANY ENGLISH?”

Voice recognition technology has always been hampered by the chips that use them. In this case, there still hasn’t been a hardware development that allows for all the necessary software to take the technology forward. This is especially true for mobile phones and PDAs, which have a lot less physical space to put voice recognition software on as opposed to computers.

Luckily for us, there’s a team at Carnegie Mellon University that’s trying to solve these problems. Their solution: putting all the necessary software on a new chip, in much the same way we have graphics cards on computers. At the Hot Chips conference in Stanford, they presented the results of their In Silico Vox project: a working prototype chip that can already recognize 1000 words. They expect to put a lot more in by next year.

What does this mean for the future? A lot more accessibility for people who don’t have the means to type, for one thing, as well as greater ease of use for people on cellphones. Instead of the distracting and potentially deadly habit of scrolling through your phone while driving, this should allow for people to tell their phones what to do. 

More importantly for the world, once we find a way to transform normal people’s clothing into spandex suits that increase its wearer’s natural abilities and serve as protective armor, we can combine the technology with the result of the In Silico Vox project to let this new technology be accessed by some kind of special morphing technology. This mighty technology will allow those men and women to morph into people of great power, and it would be best if we got some Army Rangers to use it for the good of mankind! Go Go Army Rangers!

Via Ars Technica