README
Data for instruction fine-tuning:
data-ift.csv - data prepared for instruction fine-tuning.
Each sample in the instruction fine-tuning dataset is represented as:
"instruction": "Some kind of instruction."
"input": "Some prior information."
"output": "Desirable output."
Data sources:
| Data | Number of samples | Number of created samples | Description |
|---|---|---|---|
| Almazov anamneses | 2356 | 6861 | Set of anonymized EMRs of patients with acute coronary syndrome (ACS) from Almazov National Medical Research Centre. |
| Diagnosis data | 286 | 286 | Diagnosis data that contains a description of the diagnosis and corresponding symptoms. |
| Diseases data | 2945 | 4890 | Diseases data includes a description of diseases and their symptoms from Wikipedia. |
| GPT-3.5 data | 11480 | 11480 | Clinical notes generated with GPT-3.5 based on disease-relevant symptoms. |
| GPT-4 data with symptoms | 5006 | 5006 | Clinical notes generated with GPT-4 based on disease-relevant symptoms. |
| ISA anamneses | 161 | 483 | Set of large fragments of patients’ anamnesis from Research Institute of the Russian Academy of Sciences. |
| MedElement | 1741 | 1741 | Pairs of the abbreviation and its definition from the MedElement portal. |
| Medical QA | 10937 | 10937 | Q&A posts from a medical-related forum |
| MedSpellcheck | 1054 | 1054 | MedSpellcheck benchmark contains pairs of a sentence that may contain spelling errors and corresponding correction. |
| Medical Knowledge Graph | 13582 | 6000 | Complex instructions with descriptions of diseases, drugs, and symptoms, extracted from the MKG. |
| MSD Manual | 162 | 1789 | Set of detailed descriptions of diseases and associated symptoms from MSD Manuals. |
| RuMedDaNet | 1564 | 1564 | Q&A pairs in the range of medical-related domains (pharmacology, anatomy, therapeutic medicine, etc). |
| RuMedNLI | 12627 | 12627 | NLI task in the clinical domain, full translated counterpart of MedNLI data. |
| RuMedPrime | 7625 | 33465 | Anonymized medical anamneses of SSMU hospital visitors. |
| RuMedTest SOGMA | 770 | 355 | RuMedTest is a 5-question test on different medical fields (gastroenterology, cardiovascular diseases, etc.). |
| In-house clinical notes | 9024 | 19719 | Set of anonymized in-house patient clinical notes. |
| WikiMed | 13324 | 33951 | Set of descriptions of diseases and drug names from WikiMed. |
Citation
@InProceedings{10.1007/978-3-031-70381-2_14,
author="Kumichev, Gleb
and Blinov, Pavel
and Kuzkina, Yulia
and Goncharov, Vasily
and Zubkova, Galina
and Zenovkin, Nikolai
and Goncharov, Aleksei
and Savchenko, Andrey",
editor="Bifet, Albert
and Krilavi{\v{c}}ius, Tomas
and Miliou, Ioanna
and Nowaczyk, Slawomir",
title="MedSyn: LLM-Based Synthetic Medical Text Generation Framework",
booktitle="Machine Learning and Knowledge Discovery in Databases. Applied Data Science Track",
year="2024",
publisher="Springer Nature Switzerland",
address="Cham",
pages="215--230",
isbn="978-3-031-70381-2"
}
Similar repos
Linked in READMEs
README
Data for instruction fine-tuning:
data-ift.csv - data prepared for instruction fine-tuning.
Each sample in the instruction fine-tuning dataset is represented as:
"instruction": "Some kind of instruction."
"input": "Some prior information."
"output": "Desirable output."
Data sources:
| Data | Number of samples | Number of created samples | Description |
|---|---|---|---|
| Almazov anamneses | 2356 | 6861 | Set of anonymized EMRs of patients with acute coronary syndrome (ACS) from Almazov National Medical Research Centre. |
| Diagnosis data | 286 | 286 | Diagnosis data that contains a description of the diagnosis and corresponding symptoms. |
| Diseases data | 2945 | 4890 | Diseases data includes a description of diseases and their symptoms from Wikipedia. |
| GPT-3.5 data | 11480 | 11480 | Clinical notes generated with GPT-3.5 based on disease-relevant symptoms. |
| GPT-4 data with symptoms | 5006 | 5006 | Clinical notes generated with GPT-4 based on disease-relevant symptoms. |
| ISA anamneses | 161 | 483 | Set of large fragments of patients’ anamnesis from Research Institute of the Russian Academy of Sciences. |
| MedElement | 1741 | 1741 | Pairs of the abbreviation and its definition from the MedElement portal. |
| Medical QA | 10937 | 10937 | Q&A posts from a medical-related forum |
| MedSpellcheck | 1054 | 1054 | MedSpellcheck benchmark contains pairs of a sentence that may contain spelling errors and corresponding correction. |
| Medical Knowledge Graph | 13582 | 6000 | Complex instructions with descriptions of diseases, drugs, and symptoms, extracted from the MKG. |
| MSD Manual | 162 | 1789 | Set of detailed descriptions of diseases and associated symptoms from MSD Manuals. |
| RuMedDaNet | 1564 | 1564 | Q&A pairs in the range of medical-related domains (pharmacology, anatomy, therapeutic medicine, etc). |
| RuMedNLI | 12627 | 12627 | NLI task in the clinical domain, full translated counterpart of MedNLI data. |
| RuMedPrime | 7625 | 33465 | Anonymized medical anamneses of SSMU hospital visitors. |
| RuMedTest SOGMA | 770 | 355 | RuMedTest is a 5-question test on different medical fields (gastroenterology, cardiovascular diseases, etc.). |
| In-house clinical notes | 9024 | 19719 | Set of anonymized in-house patient clinical notes. |
| WikiMed | 13324 | 33951 | Set of descriptions of diseases and drug names from WikiMed. |
Citation
@InProceedings{10.1007/978-3-031-70381-2_14,
author="Kumichev, Gleb
and Blinov, Pavel
and Kuzkina, Yulia
and Goncharov, Vasily
and Zubkova, Galina
and Zenovkin, Nikolai
and Goncharov, Aleksei
and Savchenko, Andrey",
editor="Bifet, Albert
and Krilavi{\v{c}}ius, Tomas
and Miliou, Ioanna
and Nowaczyk, Slawomir",
title="MedSyn: LLM-Based Synthetic Medical Text Generation Framework",
booktitle="Machine Learning and Knowledge Discovery in Databases. Applied Data Science Track",
year="2024",
publisher="Springer Nature Switzerland",
address="Cham",
pages="215--230",
isbn="978-3-031-70381-2"
}