USING AUTOMATIC SPEECH RECOGNITION TO SUPPORT STUDENTS WITH DISABILITIES

Abstract

The transformative potential of digital technologies, such as using the benefits of AI and automatic speech recognition techniques, in order to improve the accessibility of students with disabilities in the education, is the main focus of this paper. As education increasingly shifts toward digital platforms, traditional barriers can be mitigated through innovative tools that cater to diverse needs. Automatic speech recognition (ASR) techniques empowers students with disabilities by providing real-time transcription, facilitating note-taking, and enhancing participation in classroom discussions. Additionally, it offers immediate feedback for language learners and supports those with writing difficulties, allowing for the expression of ideas in a verbal format. The goal of this paper is to propose an approach for implementing ASR techniques in combination with NLP translation in the educational process for students with disabilities. The proposed approach aims to create inclusive and effective learning environment for students with disabilities. Comparative analysis of several AI pre-trained models available on Hugging Face, including Wav2Vec 2.0, HuBERT, DeepSpeech and Jasper is covered in the paper. Furthermore, main challenges and best practices for implementing the proposed approach are included too.