Although the company has many research irons in the fire at its Watson Research Center in Yorktown Heights, N.Y., and at other facilities, Big Blue is reaching the breakthrough point in enabling cell phone users to use speech input to interact with Web-based resources, wedding Web services with speech technology.
"Were putting speech technology into existing applications to give customers and cell phone users interaction with the Web," said Mike Nelson, director of on-demand business at IBM.
At Wake Forest University, students are using a new cell-phone-based wireless information system that uses IBM technology, nicknamed "Moby Deke," after the schools nickname, the Demon Deacons.
"Ninety-five percent of students use cell phones," said Anne Bishop, director of research and development for information systems at Wake Forest, explaining the reason to focus on cell phones as a user interface. Bishop had implemented a previous system that worked with Pocket PCs, but not all students at the campus gravitated toward the devices. "We found a small group that used it all the time; others never used it. We wanted to combine the Pocket PC and the cell phone."
Now, users at the 6,300-student campus can download to their cell phone their class schedule and can access the librarys card catalog from their cell phones upon entering the library. The project also includes a feature that prevents the cell phones from ringing in class.
Another application enables students to check for a shuttle buss real-time location; another tells students which washing machines at a student laundry are available; yet another lets students listen to their notes played back to them as they walk across the campus.
"If we can capture the attention of 18- to 22-year-olds and focus it on the classroom and student life, then this is a technology thats really worth paying attention to," said Bishop.
IBM officials took the wraps off Embedded ViaVoice 4.4, which boasts what IBM calls "freeform command recognition," or the ability to understand the meaning of spoken words independent of strict word order using contextual algorithms. The companys ViaVoice technology is based on some 300 patents the company holds in speech innovations as well as open standards such as Voice XML.
At a showcase here, IBM demonstrated 12 different applications. In one, a speech recognition system in an automobile dashboard took voice commands to control air conditioning and XM satellite radio dialing. Another system dialed a cell phone using a dashboard-based speech recognition system. IBM has a partnership with Honda to develop this technology and is trying to sell the technology to various automakers as well as to auto suppliers, like Johnson Controls.
While some similar systems are in place in automobiles today, the IBM work implements conversational speech—rather than obeying only specific pre-cast commands—and larger databases of information than current systems, said Igor Jablokov, program director of multimodal speech applications at IBM.
Mike Kennewick, CEO of VoiceBox Technologies, in Kirkland Wash., said his company is using Embedded ViaVoice technology to build conversational voice search applications to deliver digital content to mobile users.
In a partnership with XM satellite radio, VoiceBox has developed a speech-recognition application that will let a driver talk to his or her satellite radio set to switch stations or ask for stock quotes or sports scores, he said.
VoiceBox is also working with Toyota to embed the technology in Toyota vehicles, although the car maker has not said when and in what models it will be available.