[1M5-GS-10] AI application

[1M5-GS-10-02] Automatic Generation of Multiple-Choice Questions in a Rehabilitation Support System for Persons with Aphasia using BERT

〇Yusuke Sakurai¹, Tomoyuki Uchida¹, Kayo Kawamoto¹, Daimon Shigeta, Kazuharu Kosaki² (1. Hiroshima City University, 2. Hiroshima City Rehabilitation Hospital)

Keywords:Aphasia, BERT, Rehabilitation Support System, Machine Learning

Aphasia is an acquired language disorder caused by damage to the language center of the cerebrum, which may be caused by cerebral infarction, brain hemorrhage, brain trauma, or brain tumor. A speech-language-hearing therapist (ST, for short) provides examination, advice, and guidance to maintain and improve the language function of a person with aphasia. Aphasia Rehabilitation Support System (AphaRehaSys, for short) is an Android application developed to help people with aphasia reintegrate into society. AphaRehaSys provides rehabilitation through listening, speaking, and reading tasks created by STs. In this paper, we propose a method for automatically creating multiple-choice questions for AphaRehaSys using BERT, a natural language processing model.