UPDATED 11:43 EDT / DECEMBER 07 2012

Nuance Exec Loves Hadoop: Discusses Data Science Behind Voice Recognition

Frederic Van Haren is the senior director of research for Nuance Communications, a technology provider that specializes in image recognition and speech technology. His company makes extensive use of HP solutions to power its 10-petabyte storage environment, a case study that he discussed during a session with John Furrier and Dave Vellante at HP Discover 2012 (full video below).

Van Haren started out by providing some background. Nuance is responsible for the embedded technology that allows scanners to convert an image to an editable format, he said, but his firm is known mainly for its flagship speech recognition solution: Dragon Dictation. The software is licensed by all the top GPS providers, and the end-user versions have become just as popular on both PCs and mobile devices. Van Haren explains that the handset edition, labeled Dragon Search, adds intelligence into the mix by allowing for natural language voice queries.

This naturally involves a lot of data. Due to its status, Nuance has to be able to cope with almost every accent and dialect, and the only way of doing that is constantly analyzing user communications.

The executive duly notes that it’s not very feasible to install petabytes of data on a laptop and process it locally, so the task is outsourced to Nuance’s servers. He then explains what goes in the company’s cluster, and why it needs 15,000 hard drives to power it.

Data from end-points turns up in what he calls the production environment, where it’s made workable and shipped off to the research environment – his turf. That’s where the information is analyzed, gets integrated into Nuance’s language model, and is then sent back to the production environment.  Because phone numbers are usually associated with only one user, the company can afford to create tailored models for individual mobile customers to offer even greater accurately.

Nuance has a very unique case study, which Van Halen said forced its engineers to come up with its own unique solutions. He sees a lot of promise in flash and Hadoop in particular, but says that the technology has not yet matured enough to address all of his company’s needs.


A message from John Furrier, co-founder of SiliconANGLE:

Your vote of support is important to us and it helps us keep the content FREE.

One click below supports our mission to provide free, deep, and relevant content.  

Join our community on YouTube

Join the community that includes more than 15,000 #CubeAlumni experts, including Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger, and many more luminaries and experts.

“TheCUBE is an important partner to the industry. You guys really are a part of our events and we really appreciate you coming and I know people appreciate the content you create as well” – Andy Jassy

THANK YOU