[1Q4-GS-10] AI application:

[1Q4-GS-10-04] Automatic Generation of Reading Comprehension Questions in a Rehabilitation Support System for Individuals with Aphasia

〇Yusuke Sakurai¹, Tomoyuki Uchida¹, Kayo Kawamoto¹, Kazuharu Kosaki², Yusuke Suzuk¹, Tetsuhiro Miyahara¹ (1. Hiroshima City University, 2. Hiroshima City Rehabilitation Hospital)

Keywords:Aphasia, Rehabilitation Support System, Machine Learning, Bidirectional Encoder Representations from Transformers, Text-to-Text Transfer Transformer

Aphasia is an acquired language disorder caused by damage to the language centers in the cerebral cortex, resulting from conditions such as cerebral infarction, cerebral hemorrhage, traumatic brain injury, or brain tumors. Speech-language pathologists assess, advise, and guide individuals with aphasia (aphasic patients) to help maintain and improve their language functions.The authors are developing a rehabilitation support system for aphasic patients, called the Aphasia Rehabilitation Support System (AphaRehaSys), aimed at facilitating their social reintegration.In AphaRehaSys, aphasic patients engage in rehabilitation to recover their language functions by solving problems based on plans created by speech-language pathologists. These problems are primarily designed by speech-language pathologists.To reduce the burden on speech-language pathologists, this paper proposes a method for automatically generating reading comprehension questions using the natural language processing models BERT and T5.Furthermore, the effectiveness of the proposed method is demonstrated through evaluation experiments on the automatically generated reading comprehension questions.