Data Sheet 1_Fine-tuning a local LLaMA-3 large language model for automated privacy-preserving physician letter generation in radiation oncology.pdf
Introduction<p>Generating physician letters is a time-consuming task in daily clinical practice.</p>Methods<p>This study investigates local fine-tuning of large language models (LLMs), specifically LLaMA models, for physician letter generation in a privacy-preserving manner within...
محفوظ في:
| المؤلف الرئيسي: | |
|---|---|
| مؤلفون آخرون: | , , , , , , , , , , , , , |
| منشور في: |
2025
|
| الموضوعات: | |
| الوسوم: |
إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
|
| _version_ | 1852023628260966400 |
|---|---|
| author | Yihao Hou (20555675) |
| author2 | Christoph Bert (34117) Ahmed Gomaa (4115773) Godehard Lahmer (20555678) Daniel Höfler (10512040) Thomas Weissmann (9960221) Raphaela Voigt (20555681) Philipp Schubert (9577160) Charlotte Schmitter (6449204) Alina Depardon (20555684) Sabine Semrau (9577172) Andreas Maier (6397244) Rainer Fietkau (757368) Yixing Huang (16324230) Florian Putz (4048246) |
| author2_role | author author author author author author author author author author author author author author |
| author_facet | Yihao Hou (20555675) Christoph Bert (34117) Ahmed Gomaa (4115773) Godehard Lahmer (20555678) Daniel Höfler (10512040) Thomas Weissmann (9960221) Raphaela Voigt (20555681) Philipp Schubert (9577160) Charlotte Schmitter (6449204) Alina Depardon (20555684) Sabine Semrau (9577172) Andreas Maier (6397244) Rainer Fietkau (757368) Yixing Huang (16324230) Florian Putz (4048246) |
| author_role | author |
| dc.creator.none.fl_str_mv | Yihao Hou (20555675) Christoph Bert (34117) Ahmed Gomaa (4115773) Godehard Lahmer (20555678) Daniel Höfler (10512040) Thomas Weissmann (9960221) Raphaela Voigt (20555681) Philipp Schubert (9577160) Charlotte Schmitter (6449204) Alina Depardon (20555684) Sabine Semrau (9577172) Andreas Maier (6397244) Rainer Fietkau (757368) Yixing Huang (16324230) Florian Putz (4048246) |
| dc.date.none.fl_str_mv | 2025-01-14T06:07:49Z |
| dc.identifier.none.fl_str_mv | 10.3389/frai.2024.1493716.s001 |
| dc.relation.none.fl_str_mv | https://figshare.com/articles/dataset/Data_Sheet_1_Fine-tuning_a_local_LLaMA-3_large_language_model_for_automated_privacy-preserving_physician_letter_generation_in_radiation_oncology_pdf/28201901 |
| dc.rights.none.fl_str_mv | CC BY 4.0 info:eu-repo/semantics/openAccess |
| dc.subject.none.fl_str_mv | Knowledge Representation and Machine Learning radiation oncology data privacy parameter-efficient fine-tuning LLaMA fine-tuning physician letter large language model LLM |
| dc.title.none.fl_str_mv | Data Sheet 1_Fine-tuning a local LLaMA-3 large language model for automated privacy-preserving physician letter generation in radiation oncology.pdf |
| dc.type.none.fl_str_mv | Dataset info:eu-repo/semantics/publishedVersion dataset |
| description | Introduction<p>Generating physician letters is a time-consuming task in daily clinical practice.</p>Methods<p>This study investigates local fine-tuning of large language models (LLMs), specifically LLaMA models, for physician letter generation in a privacy-preserving manner within the field of radiation oncology.</p>Results<p>Our findings demonstrate that base LLaMA models, without fine-tuning, are inadequate for effectively generating physician letters. The QLoRA algorithm provides an efficient method for local intra-institutional fine-tuning of LLMs with limited computational resources (i.e., a single 48 GB GPU workstation within the hospital). The fine-tuned LLM successfully learns radiation oncology-specific information and generates physician letters in an institution-specific style. ROUGE scores of the generated summary reports highlight the superiority of the 8B LLaMA-3 model over the 13B LLaMA-2 model. Further multidimensional physician evaluations of 10 cases reveal that, although the fine-tuned LLaMA-3 model has limited capacity to generate content beyond the provided input data, it successfully generates salutations, diagnoses and treatment histories, recommendations for further treatment, and planned schedules. Overall, clinical benefit was rated highly by the clinical experts (average score of 3.4 on a 4-point scale).</p>Discussion<p>With careful physician review and correction, automated LLM-based physician letter generation has significant practical value.</p> |
| eu_rights_str_mv | openAccess |
| id | Manara_85fe389b4eecbd752d13284edfba2417 |
| identifier_str_mv | 10.3389/frai.2024.1493716.s001 |
| network_acronym_str | Manara |
| network_name_str | ManaraRepo |
| oai_identifier_str | oai:figshare.com:article/28201901 |
| publishDate | 2025 |
| repository.mail.fl_str_mv | |
| repository.name.fl_str_mv | |
| repository_id_str | |
| rights_invalid_str_mv | CC BY 4.0 |
| spelling | Data Sheet 1_Fine-tuning a local LLaMA-3 large language model for automated privacy-preserving physician letter generation in radiation oncology.pdfYihao Hou (20555675)Christoph Bert (34117)Ahmed Gomaa (4115773)Godehard Lahmer (20555678)Daniel Höfler (10512040)Thomas Weissmann (9960221)Raphaela Voigt (20555681)Philipp Schubert (9577160)Charlotte Schmitter (6449204)Alina Depardon (20555684)Sabine Semrau (9577172)Andreas Maier (6397244)Rainer Fietkau (757368)Yixing Huang (16324230)Florian Putz (4048246)Knowledge Representation and Machine Learningradiation oncologydata privacyparameter-efficient fine-tuningLLaMAfine-tuningphysician letterlarge language modelLLMIntroduction<p>Generating physician letters is a time-consuming task in daily clinical practice.</p>Methods<p>This study investigates local fine-tuning of large language models (LLMs), specifically LLaMA models, for physician letter generation in a privacy-preserving manner within the field of radiation oncology.</p>Results<p>Our findings demonstrate that base LLaMA models, without fine-tuning, are inadequate for effectively generating physician letters. The QLoRA algorithm provides an efficient method for local intra-institutional fine-tuning of LLMs with limited computational resources (i.e., a single 48 GB GPU workstation within the hospital). The fine-tuned LLM successfully learns radiation oncology-specific information and generates physician letters in an institution-specific style. ROUGE scores of the generated summary reports highlight the superiority of the 8B LLaMA-3 model over the 13B LLaMA-2 model. Further multidimensional physician evaluations of 10 cases reveal that, although the fine-tuned LLaMA-3 model has limited capacity to generate content beyond the provided input data, it successfully generates salutations, diagnoses and treatment histories, recommendations for further treatment, and planned schedules. Overall, clinical benefit was rated highly by the clinical experts (average score of 3.4 on a 4-point scale).</p>Discussion<p>With careful physician review and correction, automated LLM-based physician letter generation has significant practical value.</p>2025-01-14T06:07:49ZDatasetinfo:eu-repo/semantics/publishedVersiondataset10.3389/frai.2024.1493716.s001https://figshare.com/articles/dataset/Data_Sheet_1_Fine-tuning_a_local_LLaMA-3_large_language_model_for_automated_privacy-preserving_physician_letter_generation_in_radiation_oncology_pdf/28201901CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/282019012025-01-14T06:07:49Z |
| spellingShingle | Data Sheet 1_Fine-tuning a local LLaMA-3 large language model for automated privacy-preserving physician letter generation in radiation oncology.pdf Yihao Hou (20555675) Knowledge Representation and Machine Learning radiation oncology data privacy parameter-efficient fine-tuning LLaMA fine-tuning physician letter large language model LLM |
| status_str | publishedVersion |
| title | Data Sheet 1_Fine-tuning a local LLaMA-3 large language model for automated privacy-preserving physician letter generation in radiation oncology.pdf |
| title_full | Data Sheet 1_Fine-tuning a local LLaMA-3 large language model for automated privacy-preserving physician letter generation in radiation oncology.pdf |
| title_fullStr | Data Sheet 1_Fine-tuning a local LLaMA-3 large language model for automated privacy-preserving physician letter generation in radiation oncology.pdf |
| title_full_unstemmed | Data Sheet 1_Fine-tuning a local LLaMA-3 large language model for automated privacy-preserving physician letter generation in radiation oncology.pdf |
| title_short | Data Sheet 1_Fine-tuning a local LLaMA-3 large language model for automated privacy-preserving physician letter generation in radiation oncology.pdf |
| title_sort | Data Sheet 1_Fine-tuning a local LLaMA-3 large language model for automated privacy-preserving physician letter generation in radiation oncology.pdf |
| topic | Knowledge Representation and Machine Learning radiation oncology data privacy parameter-efficient fine-tuning LLaMA fine-tuning physician letter large language model LLM |