A Model for Incorporating an Automatic Speech Recognition System in a Noisy Educational Environment

Automatic speech recognition technology has a high potential for improving the learning experience of students in an educational setting. This paper addresses some of the key theoretical areas involved in developing automatic speech recognition systems for educational use; namely the applications of the technology in education, prominent feature extraction and noise cancellation techniques used with audio speech data as well as some of the recent neural network based machine learning models capable of keyword spotting or continuous speech recognition. Following the theoretical background, a model for a strategy of incorporating an automatic speech recognition system in a noisy educational environment is proposed. The model is not only used to generate a lesson transcript, but also to associate named keywords identified during the lesson with the course content the lesson is based upon. Ultimately, this model is used to present a topical lesson overview to students which allows them to monitor the key topics and their prevalence as communicated in the lesson while having a transcript of what was communicated to refer back to for review purposes.