Evaluating the Limitations of Large Language Models in Medical Software Systems Engineering
- سال انتشار: 1404
- محل انتشار: اولین کنفرانس بین المللی پزشکی و هوش مصنوعی در ارتقای سلامت
- کد COI اختصاصی: AIMCNFE01_006
- زبان مقاله: انگلیسی
- تعداد مشاهده: 24
نویسندگان
Professor at the Computer Engineering Department, National maharat University (Shariati Technical Faculty)
Computer Engineering Department, National maharat University (Shariati Technical Faculty)
Computer Engineering Department, National maharat University (Shariati Technical Faculty)
Computer Engineering Department, National maharat University (Shariati Technical Faculty)
چکیده
Large Language Models (LLMs) have demonstrated impressive advancements in a range of technological fields. However, their application to medical software systems engineering presents significant challenges. This paper systematically investigates the limitations of LLMs in this context, focusing on the critical nature of medical systems and the stringent regulatory requirements that constrain their use. Key limitations identified include issues with output accuracy, model interpretability, compliance with medical standards such as IEC ۶۲۳۰۴ and ISO ۱۳۴۸۵, security concerns when handling sensitive medical data, and ethical ambiguities regarding accountability. Through a detailed case study analysis and an extensive review of existing literature, the paper highlights how these challenges can negatively affect the medical software development lifecycle. The findings suggest that the direct implementation of LLMs in medical systems, without proper safeguards, could pose serious risks to patient safety and system reliability. The paper proposes several practical solutions, including the development of specialized validation frameworks, combining rule-based and deep learning approaches, and designing hybrid human-machine architectures. In conclusion, the study outlines future research directions to address these challenges and advance the safe integration of AI technologies in healthcare.کلیدواژه ها
Large Language Models (LLMs), Medical Software Engineering, AI in Healthcare, Software Reliability, Regulatory Complianceمقالات مرتبط جدید
- Artificial intelligence-based Diagnostic Approaches for Alzheimer's Disease Using Medical Imaging
- Artificial Intelligence-Based Telehealth Care in maternal health
- سیستم های هوش مصنوعی در پیشگیری و ارتقای سلامت عمومی
- Efficient Multi-Label Retinal Disease Classification with CLIP, LoRA, and Shadow Loss on the OIA-ODIR Dataset
- هوش مصنوعی در ژنتیک
اطلاعات بیشتر در مورد COI
COI مخفف عبارت CIVILICA Object Identifier به معنی شناسه سیویلیکا برای اسناد است. COI کدی است که مطابق محل انتشار، به مقالات کنفرانسها و ژورنالهای داخل کشور به هنگام نمایه سازی بر روی پایگاه استنادی سیویلیکا اختصاص می یابد.
کد COI به مفهوم کد ملی اسناد نمایه شده در سیویلیکا است و کدی یکتا و ثابت است و به همین دلیل همواره قابلیت استناد و پیگیری دارد.