|
NativeAccent
IntroductionCarnegie Speech is a speech assessment and tutoring software company focused on testing and improving English pronunciation for non-native speakers. Using state-of-the-art speech recognition technology developed at Carnegie Mellon University, we market applications that solve specific problems in the corporate business process outsourcing (BPO) and education markets. These problems have been caused by the explosive growth of nonnative English speakers and the manual, inefficient methods for assessing and improving English pronunciation. In particular, the problems are:
To address these problems, Carnegie Speech markets two products: Carnegie Speech Assessment (English) and NativeAccentTM (English). Both products are based on an extensible software platform using the our patent-pending speech recognition and pinpointing technology. The pinpointing technology uses automated speech recognition and advanced artificial intelligence techniques to pinpoint exactly where the error is being made in spoken English sounds, rhythm, or pitch. Carnegie Speech AssessmentCarnegie Speech Assessment uses speech recognition and pinpointing technology to listen to a user’s speech and to pinpoint exactly what was correct and incorrect in the sound, rhythm and pitch of the speech. In addition to the technologies described above, this product is also based on a targeted embodiment of the linguistics and second-language learning literature. It is able to classify the quality of the individual’s speech and provides a costeffective, objective assessment capability for the corporations and educational organizations that need to perform these assessments on a large scale. The speech assessment market is diverse and includes customers who would like to classify the quality of the speech of individuals ranging from interview candidates to young students. The population of users whose speech needs to be tested is as diverse as the applications for which customers want to assess the users. For instance, a corporate customer may be interested in whether they should hire a potential job candidate and how much speech training that candidate will require to perform some particular task, while a state Board of Education may be interested in classifying students based on what classes they should be taking. Carnegie Speech’s Assessment technology can tailor the assessment to criteria given by the customer. Other assessment software “grades” users against a general notion of English fluency. This may be insufficient in a BPO/call center environment. Prospective call center agents may attain a grade of “good” after the assessment; however, these agents may not be adequate for the demands of the call center situation. Carnegie Speech uses classified customer data to adapt the assessment to categorize the prospective agents in the most useful way for each customer. The prospect then gets the most meaningful assessment for their specific situation. This ability to automatically adapt to explicit assessment classifications is unique to Carnegie Speech. The resulting applications of this product within the call center market include hiring screens and pre- and posttraining assessments. The product can also be customized to focus on known regional speech difficulties. NativeAccentTMNativeAccentTM uses automatic speech recognition and advanced artificial intelligence techniques to pinpoint a user’s pronunciation errors and to show him how to correct them while adapting to his individual abilities and needs. Specifically, NativeAccentTM differentiates itself from other English educational software by:
Learning a foreign language is not just a question of getting the words and the syntax right. A student cannot be understood until he can pronounce words well. NativeAccentTM is designed to let the student speak and then to give them performance feedback. This interactive software allows students to speak, get corrections, listen to themselves and a native speaker and try again, over and over, as many times as they want. They can practice with the privacy of their PC and then speak to others when they are ready! NativeAccentTM provides over 100 hours of training for all the sounds of English and its rhythm and pitch. The market for NativeAccentTM extends beyond non-native speakers to anyone who, due to dialect or other concerns, needs to improve the quality and understandability of his spoken English. Our product tutors students through the following process:
Unlike most other English educational software products, NativeAccentTM uses speech recognition technology to help with pronunciation. Only Carnegie Speech’s product can immediately pinpoint errors, tell the student what was wrong and deliver helpful corrective instruction. Lessons are then adapted to problems linked to the student’s native language and each lesson plan adapts according to the student’s personal progress. This intelligent tutoring greatly reduces time and effort needed for pronunciation training. Competitive AdvantagesCarnegie Speech has competitive advantages in both the English assessment market as well as the pronunciation training market. The competitive advantages that cross both applications are:
We believe that a competitive speech assessment and tutoring technology must have this entire collection of capabilities in order to produce a quality product. Our connection to Carnegie Mellon University’s premier Language Technologies Institute is a distinct advantage – they have had over 150 researchers studying these issues for 20 years. The CompanyThe company has focused its market strategy on two core markets – the BPO market (with call centers as the entry point) and the education market (with universities as the entry point). Carnegie Speech has a channel partner, Language Solutions, LLC (LS), out of Minneapolis that is addressing the Workforce Development market. LS is focusing on workplace (employment and job training) and life skills for the immigrant, Adult Basic Education, and low literacy marketplaces. They have significant customer traction with workforce development centers in places as diverse as SC, IN, IA, and MN. Carnegie Speech has an experienced management team that combines prior successes in technology software start-up companies with worldwide domain expertise in the areas of artificial intelligence and language technologies. Jaime Carbonell is the Chairman of Carnegie Speech and is one of the world’s foremost experts in artificial intelligence and language technology. He is the Director of the Language Technologies Institute at Carnegie Mellon University and is a serial entrepreneur. The CEO is Angela Kennedy, who has significant sales and marketing experience in the software industry and was previously CEO of Wisdom Technologies. Maxine Eskenazi is the CTO and has published over 40 papers in the area of speech recognition and modern language pedagogy. In addition to the many research appointments in both Europe and the United States she has held, she served as the Head of the Assessment and Variability Theme in the Speech Group at the LIMSI laboratory from 1984 to 1994. She is currently the Director of the Masters Program in Computer-Assisted Language Learning at Carnegie Mellon University. Contact:Angela Kennedy, CEO 437 Grant Street, Ste. 918
|