Over the summer, Google revolutionized the way that humans interact with intelligent software to recognize images and a variety of other things simply by watching YouTube (News - Alert) videos. This “self-taught software” was developed to help further advance the approach made to image recognition technology so that individuals could extend their identification to encompass a much broader range of objects. This, in turn, would make image search engines more powerful to help make robots better at interpreting their surroundings. Now, this technology is being applied to Google’s (News - Alert) products to make them “smarter,” with speech recognition being the first to benefit, as reported by Telepresence Options.

The way this technology works is that Google’s learning software is based on simulating groups of connected brain cells that communicate and influence one another. When these “neural networks” are exposed to data, the relationships between different neurons can change, which then causes the network to develop the ability to react specifically to specific incoming data.

Neural networks are technically nothing new, as they have been used for decades for a variety of reasons, but Google’s engineers have now found a way to put this innovative technology to better use – by strengthening the computing power behind this process to allow these neural networks to learn without human assistance. In other words, so that they are “robust enough to be used commercially, not just as research demonstrations,” Telepresence (News - Alert) explains.

“Google is now using these neural networks to recognize speech more accurately, a technology increasingly important to Google's smartphone operating system, Android (News - Alert), as well as the search app it makes available for Apple devices,” the site adds.

Big strides have been made with this advancement, where Vincent Vanhoucke, a leader in Google’s speech-recognition efforts, claims that they’ve seen anywhere between 20 to 25 percent more improved speech-recognition efforts as a result, which means that more people will be able to enjoy a perfect and flawless user experience.

One thing remains sure – speech recognition, a crucial aspect to interactive voice response (IVR) technology, will continue to see major improvements made as technology further grows, evolves and betters itself.

To learn more about Google’s neural networks, click here.

Want to learn more about the latest in communications and technology? Then be sure to attend ITEXPO Austin 2012, happening now in Austin, TX.  Stay in touch with everything happening at ITEXPO (News - Alert). Follow us on Twitter.