Paper Image

Automating conversational tutors from text

Published on:

26 September 2023

Primary Category:

Computation and Language

Paper Authors:

Robin Schmucker,

Meng Xia,

Amos Azaria,

Tom Mitchell

Bullets

Key Details

Proposes a novel workflow to induce full conversational tutors from text via LLMs

Orchestrates tutoring via two agents in student and professor roles

Improved helpfulness/coherence ratings but not test scores in user study

Enables more scalable authoring of conversational tutors

Provides insights for future LLM-based educational systems

AI generated summary

Automating conversational tutors from text

This paper introduces a new type of conversational tutoring system that uses large language models to automatically generate tutoring scripts from lesson texts, and orchestrate them via two AI agents acting as a student and professor. In a user study, the system improved helpfulness and coherence ratings, though didn't significantly impact test scores.

Answers from this paper

Comments

No comments yet, be the first to start the conversation...

Sign up to comment on this paper

Sign Up