Abstract
Conventional language learning applications often lack the immersive, real-world conversational practice necessary for effective language acquisition. This disclosure describes a method for language acquisition using an extended reality (XR) platform that generates realistic environments populated by interactive avatars. These avatars can be rendered using a text-grounded Generative Adversarial Network (GAN) and are trained on datasets of native speakers’ speech and facial expressions. A large language model (LLM) facilitates real-time, audio-based conversation between the user and the avatars, with conversational difficulty adapting to the user’s proficiency level. A spaced repetition algorithm can also be used to reinforce vocabulary. The purpose of this method is to provide a more effective learning tool by simulating realistic conversational scenarios, enhancing user engagement, and improving language retention through interactive, contextual practice.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Shin, Dongeek, "Immersive Language Learning Platform Using Extended Reality and Avatars", Technical Disclosure Commons, (January 09, 2026)
https://www.tdcommons.org/dpubs_series/9155