SpeechCycle to Work with University to Expand Recognition

SpeechCycle and Ulm University, located in Germany, have forged a partnership to develop new methodologies in the area of telephony-based speech recognition. The goal of this collaboration is to advance the knowledge of how to create more interactive speech applications and to provide potential commercial value in the areas of rapid deployment for highly complex speech applications and automatic detection of caller described issues.

The Faculty of Engineering and Computer Science at Ulm University has placed a focus on the topic of advanced human-machine communication. "Our efforts in the field of statistical language modeling are targeted at making human-machine communication more adaptive and ultimately easier and more pleasing for caller interaction," said Dr. Wolfgang Minker, head of the university's Dialogue Systems Group.

SpeechCycle will contribute experience gained from automating highly complex, telephony-based technical support solutions as well as examples of how callers interact with these systems. Researchers from Ulm will collaborate with SpeechCycle to explore alternatives for creating natural language-based dialogs using statistical methods for more rapid deployment. The team will also investigate how to quickly identify when a response from a speech recognition application is needed to a request that might not originally have been anticipated by the application.