VOCALPOINT WEB AI FOR SPOKEN ENGLISH PLATFORM WITH AI ASSIST: A WEB-BASED INTELLIGENT TUTORING SYSTEM FOR PRONUNCIATION AND FLUENCY TRAINING

This paper presented the design, implementation, and preliminary evaluation of VocalPoint Web AI For Spoken English Platform With AI Assist, a web-based intelligent tutoring system for spoken English. The platform combines browser-based audio capture, cloud-hosted ASR, neural pronunciation and fluency scoring, learner modeling, and adaptive recommendation to deliver personalized speaking practice and feedback. Pilot evaluation results indicate learners achieve meaningful gains in pronunciation accuracy and speaking rate within 4–6 weeks of use, and they generally perceive the system as both useful and engaging. The proposed multi-tier architecture supports scalability, modularity, and seamless integration with institutional user management and learning management systems. Future work will focus on: • Enhanced Conversational Capabilities: Integrating generative AI for open-ended, context-rich role-play conversations. • Richer Pedagogical Models: Incorporating phonological acquisition theory, task-based learning, and communicative competence frameworks. • Expanded Analytics: Cohort-level error pattern dashboards enabling targeted classroom interventions. • Robustness and Fairness: Improving robustness to diverse accents, background noise, and demographic fairness. • Large-Scale Trials: Conducting controlled studies with larger, more diverse populations and longer intervention periods.

[1] D. Nunan, Teaching English to Speakers of Other Languages: An Introduction. New York, NY, USA: Routledge, 2015. [2] G. E. Hinton et al., "Deep neural networks for acoustic modeling in speech recognition," IEEE Signal Process. Mag., vol. 29, no. 6, pp. 82–97, Nov. 2012. [3] S. Witt and S. Young, "Phone-level pronunciation scoring and assessment for interactive language learning," Speech Communication, vol. 30, no. 2–3, pp. 95–108, 2000. [4] J. D. Fletcher, "Evidence for learning from technology-assisted instruction," in Technology and Learning, H. F. O'Neil and R. S. Perez (Eds.), Lawrence Erlbaum, 2003. [5] A. Ruan, J. Jiang, G. Xu, and X. Xu, "QuizBot: A dialogue-based adaptive learning system for factual knowledge," in Proc. CHI Conf., Glasgow, UK, 2019. [6] A. Hannun et al., "Deep speech: Scaling up end-to-end speech recognition," arXiv preprint arXiv:1412.5567, 2014. [7] A. Baevski, Y. Zhou, A. Mohamed, and M. Auli, "wav2vec 2.0: A framework for self-supervised learning of speech representations," in Advances in NeurIPS, vol. 33, 2020, pp. 12449–12460. [8] K. Zechner, D. Higgins, X. Xi, and D. M. Williamson, "Automatic scoring of non-native spontaneous speech in tests of spoken English," Speech Communication, vol. 51, no. 10, pp. 883–895, Oct. 2009. [9] L. Chen and K. Zechner, "Applying automated evaluation of children's oral reading fluency," in Proc. INTERSPEECH, Florence, Italy, 2011, pp. 3013–3016. [10] P. Boersma and D. Weenink, Praat: Doing Phonetics by Computer, Version 6.4, 2024. [Online]. Available: https://www.praat.org/