| 研究課題/領域番号 |
23K00767
|
|---|
| 研究種目 |
基盤研究(C)
|
| 配分区分 | 基金 |
|---|
| 応募区分 | 一般 |
|---|
| 審査区分 |
小区分02100:外国語教育関連
|
|---|
| 研究機関 | 東京医科歯科大学 |
|---|
研究代表者 |
デニソン ジェネット 東京医科歯科大学, 教養部, 准教授 (60764793)
|
|---|
| 研究分担者 |
Gary Ross 金沢大学, 薬学系, 准教授 (10708142)
|
|---|
| 研究期間 (年度) |
2023-04-01 – 2027-03-31
|
| 研究課題ステータス |
交付 (2023年度)
|
| 配分額 *注記 |
4,680千円 (直接経費: 3,600千円、間接経費: 1,080千円)
2026年度: 910千円 (直接経費: 700千円、間接経費: 210千円)
2025年度: 1,040千円 (直接経費: 800千円、間接経費: 240千円)
2024年度: 1,690千円 (直接経費: 1,300千円、間接経費: 390千円)
2023年度: 1,040千円 (直接経費: 800千円、間接経費: 240千円)
|
|---|
| キーワード | speech recognition / chatbot / ChatGPT / medical English / medical interview / medical communication / speaking practice / online learning / autonomy / verbal engagement |
|---|
| 研究開始時の研究の概要 |
This study will investigate the feasibility of innovative methodologies using speech recognition and chatbot technologies in an autonomous, self-evaluating learning platform for increasing verbal engagement and improving communication skills within medical English education programs.
|
| 研究実績の概要 |
This study is investigating the feasibility of innovative methodologies using speech recognition (SR) and chatbot technologies in an autonomous, self-evaluating learning platform for increasing verbal engagement and improving communication skills within medical English education programs. Our study has proceeded through Stages I and II.
Stage I involved development of content through ChatGPT prompt engineering for an existing SR+ChatGPT speaking program. The first set of content reflected four different tiers of medical English speaking tasks. Task 1 was pronunciation practice of medical terminology based on body system themes. Task 2 was the teaching of interview questions common for doctor-patient medical interview. Task 3 was the practice in mini scenarios of the interview questions learned in task 2. Task 4 was a full simulated interview with a "AI patient" based on constructed medical case.
In Stage II, we tested the first content set on a small cohort of non-Japanese (international) students from medical-related graduate programs at one Japanese university. This first test allowed us to measure the functionality and content appropriateness. The survey results and observations of the interactions with the program suggest that while current SR technology is still struggle to properly recognize some medical terminology, the combined SR+ChatGPT program could provide a good opportunity for self-study speaking practice. With improvements in SR technology and medical-specific LLMs, we will be able to provide accessible medical communication to a variety of English speakers.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
As planned, we have completed Stage I (program content development) and Stage II (pilot test on content with non-Japanese cohort).
|
| 今後の研究の推進方策 |
We will start Stage III of the study which involves extensive content development for medical communication coursework within a Japanese medical undergraduate program. First, we will begin testing the pilot test content in Japanese medical students based on the following parameters: level-appropriate English, functionality, user friendliness, task instructions, etc. Based on this pilot study with this small cohort of Japanese medical students we will then create additional content and expand the study to a larger cohort of students. Third, we will also look at the ability of ChatGPT to evaluate, provide feedback and track the progress of individual students. The results of Stage III will provide insight on the development of a full course of content for medical English communication programs that utilize the SR+ChatGPT program. As SR and LLM technology advances, we expect better interactions and fewer SR-related technical issues allowing for a more motivating and effective interaction.
|
