Carnegie Speech is the premier developer and provider of software for assessing and teaching spoken language skills. A spin-off of Carnegie Mellon University, Carnegie Speech uses speech recognition and advanced artificial intelligence techniques to pinpoint exactly where errors are made in spoken language sounds, grammar, stress, rhythm, fluency and pace. Carnegie Speech's technologies then build an individualized training curriculum that maximizes training effectiveness and minimizes training time. With more than 20 years of research underlying its technology, Carnegie Speech has brought its products into the commercial marketplace with global success.
How It Works
Each student's experience with Carnegie Speech's products starts with a personal language assessment. Speech recognition technology, developed at Carnegie Mellon University and licensed to Carnegie Speech, assesses the student's speech against a statistical model of native speech on the same exercises. The resulting performance assessment is then used to grade the student's spoken language proficiency and ability, which is then used to create a customized curriculum focusing on each student's specific language training needs and challenges.
In the training phase, Carnegie Speech's software presents students with exercises and records their spoken-responses. Our pinpointing technology immediately grades these responses and provide students with immediate textual, graphical and audio remediation feedback. In other words, the software tells the student exactly what is wrong with their speech – and exactly how to fix it.
Changes in the student's abilities are tracked using Bayesian statistical techniques and recorded in each trainee's Student Model. This Student Model feeds extensive reporting mechanisms so that students, teachers and administrators can track student performance in an accurate and timely manner. The Student Model also gives constant feedback to the Curriculum Manager, so that the student is always working on exercises that maximize progress and minimize training time.