File size: 5,250 Bytes
9c08316 dd91014 9c08316 dd91014 9c08316 b172f73 2dc8e64 b172f73 9c08316 b172f73 9c08316 b172f73 dd91014 b172f73 2dc8e64 b172f73 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 |
---
language:
- ja
pretty_name: "Japanese National Medical Licensing Examination Dataset"
tags:
- medical
- examination
- Japan
- 医師国家試験
- 医療
- Doctor
license: "cc-by-nc-nd-4.0"
task_categories:
- multiple-choice
- question-answering
size_categories:
- 1K<n<10K
---
# 医師国家試験データセット(NMLE datasets)
## はじめに
検索してもなかったので(調べ不足の可能性あり)、医師国家試験データセットを公開します
医師として、AIのエンジニアとして、医師国家試験のデータセットはあるべきもの、あって当然のものと思っていましたが今現在(2024/6/13時点)存在が確認できず、自分で使うのに必要だったため作成
医療用タスクに特化したLLMなどの開発も活発になっていますが、そのデータ元としてUSMLE(アメリカの医師国家試験)が使われており、およそ国内の実情と違うデータセットを使うことの弊害が生じていると思っています
なので何かの役にたてばと思いこのデータセットを公開します
## 用途
用途として
- モデルの評価
- [進化的モデルマージ](https://sakana.ai/evolutionary-model-merge-jp/)のタスクにつかう(「[New Task Guide](https://github.com/EleutherAI/lm-evaluation-harness/blob/main/docs/new_task_guide.md?ref=blog.arcee.ai)」参照)
- RAGなどに用いる情報源
- 医師国家試験の俯瞰
を想定しています
## 構造
```json
data = {
"id": question_id,
"question": question_text,
"choices": choices,
"answer": answers,
"explanation": explanation
}
```
一部、構造化データにできなかった問題(画像がメインの出題など)が抜けています
## ライセンス
ライセンスはCC-BY-NC-ND4.0で、商用利用禁止となっています
改変、改善、その他ご相談については
X: @longislandtea3
までお願いします
## 免責事項
このデータセットの使用により生じた直接的、間接的、特別な損害、またはその他の損害について、一切の責任を負いません。本データセットは現状のままで提供されており、正確性、完全性、有用性について何ら保証するものではありません。本データセットの使用に際しては、自己の責任で行ってください。
## 著作権・感謝
データは医療美術部の医師国家試験過去問チャートから引用させていただきましたhttps://medical-illustration.club/kakomon-chart/med
引用を快く受け入れてくださった、医療美術部の皆様、及び米澤昌紘先生に感謝申し上げます。
またデータの中身の著作権は医療美術部に帰属します。
# Japanese National Medical Licensing Examination Dataset (NMLE datasets)
## Introduction
Since I couldn't find it through my searches (though I might have overlooked something), I am releasing a Japanese National Medical Licensing Examination dataset.
As a Doctor and AI engineer, I believe that having a dataset for the National Medical Licensing Examination is essential and expected. However, as of June 13, 2024, I couldn't confirm its existence, so I created it for my own use.
Development of LLMs specialized for medical tasks is becoming active, and the USMLE (United States Medical Licensing Examination) dataset is often used as the data source. However, I believe that using a dataset that does not reflect domestic conditions can cause issues.
Therefore, I am releasing this dataset in the hope that it will be useful to others.
## Use Cases
Intended use cases include:
- Model evaluation
- Use in tasks such as [Evolutionary Model Merge](https://sakana.ai/evolutionary-model-merge-jp/) (refer to [New Task Guide](https://github.com/EleutherAI/lm-evaluation-harness/blob/main/docs/new_task_guide.md?ref=blog.arcee.ai))
- Information source for RAG
- Overview of the National Medical Licensing Examination
## Structure
```json
data = {
"id": question_id,
"question": question_text,
"choices": choices,
"answer": answers,
"explanation": explanation
}
```
Some questions that could not be structured (e.g., those primarily based on images) are missing.
## License
The license is CC-BY-NC-ND4.0, prohibiting commercial use.
For modifications, improvements, or other inquiries, please contact:
X: @longislandtea3
## Disclaimer
We do not take any responsibility for any direct, indirect, special, or other damages arising from the use of this dataset. This dataset is provided as is, without any guarantees of accuracy, completeness, or usefulness. Use it at your own risk.
## Copyright & Acknowledgments
The data was quoted from the National Medical Licensing Examination past question charts provided by the Medical Illustration Club (https://medical-illustration.club/kakomon-chart/med).
We would like to express our gratitude to the members of the Medical Illustration Club and Dr. Masahiro Yonezawa for kindly accepting the use of their data.
The copyright of the content of this data belongs to the Medical Illustration Club.
license: cc-by-nc-nd-4.0
|