Intelligent Robotic Tutoring: Integrating Verbal Input for Personalizing Learning Responses

EasyChair Preprint no. 12401

Abstract

Previous research has demonstrated that robots can be effective teaching aids, yet they lack the ability to understand and respond to students' speech in the way human teachers do. In this paper, we introduce an innovative approach that utilizes Large Language Models (LLMs) to enable more intelligent teaching robots, thereby enhancing the learning experience. Initially, we explore the capability of LLMs to evaluate students' learning progress through their speech, surpassing traditional student progress models like Bayesian Knowledge Tracing, which depend solely on the correctness of students' answers. We then proceed to discuss the second phase, which involves comparing various mechanisms of using LLMs to generate personalized feedback, and incorporating the most appropriate one into the final framework.

Keyphrases: human-robot interaction, Intelligent Tutoring System, Large Language Model

@Booklet{EasyChair:12401,
  author = {Puming Jiang and Nicole Salomons},
  title = {Intelligent Robotic Tutoring: Integrating Verbal Input for Personalizing Learning Responses},
  howpublished = {EasyChair Preprint no. 12401},

  year = {EasyChair, 2024}}