AI-Powered Foreign Language Conversation Practice with LLMs

Tuesday, April 30, 2:35-2:55 pm
Room 236
Presenter: Owen Sizemore
Modality: Traditional Talk

Abstract

Large Language Models (LLMs) are a versatile, sophisticated technology which has brought the complex task of natural language generation and understanding to a variety of academic, economic, and social domains. In the context of education, LLMs offer promising solutions to offload various responsibilities from educators and enhance the quality of education delivered to students, such as generating unique lesson plans, explaining confusing course concepts in easy-to-follow language, and providing personalized feedback to students at scale. The discipline of foreign language education stands to benefit enormously from the adoption of LLMs in the classroom: current literature suggests that students desire personalized learning environments to keep them engaged as well as greater access to practice speaking and writing even when an instructor is unavailable. While there are existing solutions that aim to offer LLM-enhanced learning resources in foreign language classrooms, such solutions struggle to simulate authentic language experiences and are often confined to rigid, text-only user interfaces with a lack of personalization. My research from the CS 6460: Educational Technology builds upon this work by prioritizing multi-modal, authentic foreign language conversation practice by implementing LLMs to engage in text-, voice-, and image-based dialogues to simulate a rich conversation with a native speaker. In this talk, I will present PolyTalk, a basic web application for foreign language conversation practice, and discuss how LLMs can be applied to spaces beyond traditional computer science environments and offer positive outcomes in a field of education that is often lagging in its adoption of educational technology.

Bio

Owen Sizemore is an Android Software Engineer at Wealthfront, an automated investment firm in Palo Alto, California. Owen grew up in Greensboro, North Carolina and attended the University of North Carolina at Charlotte, studying Computer Science with minors in Political Science and Spanish. Owen is specializing in Interactive Intelligence in the OMSCS program and currently serves as a TA for CS 8803-O17: Global Entrepreneurship. Outside of work and school, Owen enjoys playing soccer and cooking vegan meals.

Program

Check out the Program page for the full program!

Questions About the Conference?

Check out our FAQ page for answers and contact information!