Speech Recognition software has advanced considerably in recent years, even allowing users to simply seek information through Google's (News - Alert) Android platform just by speaking directly into their smartphones. However, without knowing what factors within human language are most important, speech recognition technology will essentially be useless.

Voice commands and speech recognition technology are crucial to Ford's SYNC system, as it was created as a way for people to complete tasks and receive information without putting any lives including their own in danger because there is no longer a need to take their hands off of the wheel.  Due to this fact, Ford has decided to team with Nuance (News - Alert) Communications  to develop software that can not only recognize specific words and phrases but can accurately pinpoint what the person speaking is looking for.

An article from MSNBC.com stated SYNC works off of "structured commands." The company prerecords phrases that drivers must say in order for the car to complete the action they want.

Two huge obstacles have arisen through use of this technique. First, there are an overwhelming amount of commands. The article stated, “Drivers today can order their cars to do everything from make phone calls and find directions to play music and adjust the cabin temperature.”

Second, Ford is only “guestimating” what it thinks people will say which isn’t always accurate. For example, Ford first programmed SYNC to recognize the words "Play Tracks." But unless you are a music producer, you probably don’t use the word track when speaking about music. A much better command would be either “Play Songs” or “Play Tunes.”

"You can't stop someone from saying something," Brigitte Richardson (News - Alert), Ford's lead engineer on its global voice control technology/speech systems said in a statement.

In all actuality, Ford can't make drivers change the words they use or the way they speak so they can successfully use the program.  Instead, the company must change their software. With this new collaborative effort, the company is hoping to create software based on advanced algorithms called "statistical language modeling" (SLM).

According to MSNBC, first developed in the 1980s, SLM estimates the probability of how people will group together words, phrases, and sentences according to their natural speech patterns. Nuance wants to take this idea a step further by organizing the words into "semantic classifications" of meanings. Ed Chrumka, senior product manager of connected car services, stated that Nuance is currently in the midst of creating an "inference engine" that can learn, understand, and interpret voice commands.

SLM is "a totally different way of doing things," Richardson said in a statement. "It wants to incorporate more natural ways of talking."

Although this innovative idea is great, human language is extremely intricate and includes countless phrases, words, and sentences. For a computer, recognizing words is the easy part. Figuring out the message behind the speaker's words is in another realm of its own.  

"Perhaps the most frustrating aspect of statistical language modeling is the contrast between our intuition as speakers of natural language and the over-simplistic nature of our most successful models," Ronald Rosenfeld, a computer science professor at Carnegie Mellon University, wrote in a research paper. "As native speakers, we feel strongly that language has a deep structure. Yet we are not sure how to articulate that structure, let alone encode it, in a probabilistic framework."

Currently, Google's speech recognition technology has access to an unlimited amount of resources stored in the cloud, but Ford wants its speech recognition software to be located on one, self-contained chip installed directly in the car.