[4Xin1] Poster session 2

[4Xin1-07] Toward Programming Assist AI for Error Diagnosis and Fixes

〇Momoka Obara², Nao Souma², Mai Takahashi², Kimio Kuramitsu¹ (1.Japan Women's University, 2.Graduate School of Science, Japan Women's University)

Keywords:Language Model, T5, Programming Assist AI

Programming errors are often difficult to fix for novices and they may disturb the continuous study of programming. We aim to provide programming assistance to help novices solve errors by using large language models to provide hints for solving errors or fixing error-causing code. In this presentation, we report on a method for building error corpora to support error fixing and an AI model constructed by training the built error corpora on large language models.