Kalyani Ramadurgam

👉🏼 Prompt engineering on leveraging large language models in generating response to InBasket messages 🤓 Sherry Yan 👇🏻 https://lnkd.in/eZpReUeH 🔍 Focus on data insights: - 📈 The study demonstrated a significant increase in physician acceptance of LLM-generated draft responses, rising from 62% to 84% after three iterations of prompt engineering. - 🗣️ Patients reported a notable improvement in the tone (78%) and overall quality (80%) of messages generated by the optimized prompts compared to the original ones. - 🤖 A majority of patients (76%) could not distinguish between human and LLM-generated responses, indicating high-quality output from the model. 💡 Main outcomes and implications: - 🔄 Iterative prompt engineering proved effective in enhancing the quality of responses to Patient Medical Advice Requests (PMARs), suggesting a viable method for improving clinician-patient communication. - 🧠 Clinicians expressed that the use of LLMs could significantly reduce cognitive load when managing InBasket messages, potentially leading to improved efficiency in clinical settings. - 💬 The findings highlight the potential for LLMs to bridge gaps in patient care by providing timely and relevant information, thus improving patient satisfaction. 📚 Field significance: - 🌐 This research underscores the transformative role of AI and LLMs in healthcare, particularly in managing communication and administrative tasks. - 🏥 The study contributes to the growing body of evidence supporting the integration of AI tools in clinical workflows, paving the way for future innovations in patient care. - 📊 It emphasizes the importance of user feedback in refining AI applications, ensuring that they meet the needs of both clinicians and patients effectively. 🗄️: [#AI #LargeLanguageModels #Healthcare #PatientCare #PromptEngineering #ClinicalCommunication #CognitiveLoad #PatientSatisfaction #InBasketMessages]

👉🏼 Evaluating Large Language Model (LLM) Performance on Established Breast Classification Systems 🤓 Syed Ali Haider 👇🏻 https://lnkd.in/eb3hb8se 🔍 Focus on data insights: - 📊 Gemini achieved 98% accuracy in breast classification, outperforming ChatGPT-4's 71%. - 🧠 Both models excelled in specific classifications, indicating strengths in certain areas of medical diagnostics. - 🔬 The study utilized fifty clinical scenarios to rigorously assess LLM performance across five established classification systems. 💡 Main outcomes and implications: - 🚀 The superior performance of Gemini suggests its potential as a valuable tool for plastic surgeons in diagnostic processes. - 🏥 Enhanced accuracy in classification can lead to better-informed treatment decisions, ultimately improving patient outcomes. - 🔄 Integration of LLMs into clinical practice may revolutionize diagnostic support in the field of plastic surgery. 📚 Field significance: - 🌟 This research highlights the growing role of artificial intelligence in medicine, particularly in enhancing diagnostic accuracy. - 🔗 The findings could pave the way for further studies on LLM applications in various medical specialties. - 📈 As LLM technology evolves, its integration into healthcare practices may significantly impact patient care and surgical outcomes. 🗄️: [#diagnostics #artificialintelligence #machinelearning #plasticurgery #breastclassification #medicalresearch #patientoutcomes #healthcareinnovation]

👉🏼 Fine-Tuned Bidirectional Encoder Representations From Transformers Versus ChatGPT for Text-Based Outpatient Department Recommendation: Comparative Study 🤓 Eunbeen Jo 👇🏻 https://lnkd.in/gx9UH89R 🔍 Focus on data insights: - 📊 ChatGPT demonstrated a 6.2% answer avoidance rate, indicating areas for improvement in responsiveness. - ⚖️ The fine-tuned KM-BERT model achieved an accuracy of 97.7%, highlighting its superior performance in medical specialty recommendations compared to ChatGPT's 93.9%. - 🚦 Precision and recall metrics indicate KM-BERT's effectiveness in correctly identifying relevant specialties, with precision at 57.0% and recall at 65.2%, whereas ChatGPT struggled with lower figures

👉🏼 A Systematic Review of ChatGPT and Other Conversational Large Language Models in Healthcare 🤓 Leyao Wang 👇🏻 https://lnkd.in/e--t65tY 🔍 Focus on data insights: - ChatGPT from OpenAI was the most popular conversational LLM in healthcare applications. - Concerns about reliability and bias were prevalent in research papers utilizing LLMs. - LLMs showed promising results in summarization and providing medical knowledge with high accuracy. 💡 Main outcomes and implications: - Future studies should prioritize improving the reliability of LLM applications in complex health-related tasks. - Investigation into how LLM applications introduce bias and privacy issues is crucial for further development. - Legal, social, and technical efforts are necessary to address concerns and enhance the application of LLMs in healthcare. 📚 Field significance: - Healthcare applications - Conversational Large Language Models - Data insights 🗄️: [#healthcare #LLMs #datainsights]

👉🏼 Evaluating the efficacy of few-shot learning for GPT-4Vision in neurodegenerative disease histopathology: A comparative analysis with convolutional neural network model 🤓 Daisuke Ono 👇🏻 https://lnkd.in/e2RK6-xr 🔍 Focus on data insights: - GPT-4Vision accurately recognized staining techniques and tissue origin but struggled with specific lesion identification. - The interpretation of images was notably influenced by the provided textual context, sometimes leading to diagnostic inaccuracies. - Few-shot learning significantly improved GPT-4Vision's diagnostic capabilities, enhancing accuracy from 40% in zero-shot learning to 90% with 20-shot learning. 💡 Main outcomes and implications: - GPT-4Vision faces challenges in independently interpreting histopathological images. - Few-shot learning is a promising approach to enhance GPT-4Vision's performance, especially in neuropathology where acquiring extensive labeled datasets is challenging. 📚 Field significance: - Artificial intelligence advancements like GPT-4Vision have expanded the potential for medical image interpretation. - Comparative analysis with traditional convolutional neural network models provides insights into improving diagnostic accuracy in neurodegenerative disease histopathology. 🗄️: #neurodegenerative #histopathology #GPT-4Vision #fewshotlearning

👉🏼 Performance evaluation of ChatGPT in detecting diagnostic errors and their contributing factors: an analysis of 545 case reports of diagnostic errors 🤓 Yukinori Harada 👇🏻 https://lnkd.in/eWwYv8t8 🔍 Focus on data insights: - ChatGPT correctly detected diagnostic errors in 95% of cases - ChatGPT coded statistically larger numbers of factors contributing to diagnostic errors per case compared to physicians - The most important contributing factors of diagnostic errors coded by ChatGPT were 'failure/delay in considering the diagnosis', 'atypical presentation', and 'atypical presentation' 💡 Main outcomes and implications: - ChatGPT accurately detects diagnostic errors from case presentations - ChatGPT may be more sensitive than manual reviewing in detecting factors contributing to diagnostic errors, especially for 'atypical presentation' 📚 Field significance: - Improved efficiency in detecting diagnostic errors using artificial intelligence like ChatGPT - Potential for enhancing diagnostic accuracy and reducing errors in medical practice 🗄️: #datainsights #diagnosticerrors #ChatGPT #medicalpractice

👉🏼 ChatGPT-4 Performs Clinical Information Retrieval Tasks Utilizing Consistently More Trustworthy Resources Than Does Google Search for Queries Concerning the Latarjet Procedure 🤓 Jacob F Oeding 👇🏻 https://lnkd.in/eXi69Fdk 🔍 Focus on data insights: - ChatGPT-4 had a mean accuracy of 2.9±0.9 for numeric-based answers, outperforming Google's 2.5±1.4 (p=0.65). - ChatGPT-4 sourced information solely from academic resources, significantly different from Google's use of non-academic sources. - 40% of FAQs were identical between ChatGPT-4 and Google, showcasing comparable general information retrieval capabilities. 💡 Main outcomes and implications: - ChatGPT-4 provided accurate and reliable information on the Latarjet procedure, relying on academic sources exclusively. - Google Search Engine often utilized individual surgeon and medical practice websites, potentially impacting information reliability. - Despite differences in sourcing, both platforms offered clinically relevant and accurate information to users. 📚 Field significance: - Information retrieval tools like ChatGPT-4 can enhance patient education and understanding of medical procedures. - Emphasizes the importance of utilizing trustworthy academic sources for accurate medical information dissemination. 🗄️: [#clinicalinformatics #informationretrieval #medicalresearch]

👉🏼 SPROUT: an Interactive Authoring Tool for Generating Programming Tutorials with the Visualization of Large Language Models 🤓 Yihan Liu 👇🏻 https://lnkd.in/eeJfTY3C 🔍 Focus on data insights: - The use of large language models (LLMs) like ChatGPT has significantly improved the efficiency of creating programming tutorials. - LLMs can generate comprehensive text descriptions for code snippets based on text prompts provided by users. - Transparency in the end-to-end generation process is crucial for understanding model behavior and enhancing user control over results. 💡 Main outcomes and implications: - Introducing a novel approach that breaks down the programming tutorial creation task into actionable steps. - Implementing the tree-of-thought method to engage LLMs in an exploratory process for generating diverse and faithful programming tutorials. - Presenting SPROUT, an authoring tool with interactive visualizations that empower users to have greater control and understanding of the programming tutorial creation process. 📚 Field significance: - Enhancing user experience and improving the overall quality of programming tutorials. - Providing users with more control and understanding leads to more reliable and customizable results. - Empowering users to actively participate in the programming tutorial creation process. 🗄️: #LargeLanguageModels #ProgrammingTutorials #SPROUT #DataInsights

👉🏼 Me-LLaMA: Foundation Large Language Models for Medical Applications 🤓 Qianqian Xie 👇🏻 https://lnkd.in/eQJREuAf 🔍 Focus on data insights: - Me-LLaMA models achieve better performance than existing open-source medical LLMs in zero-shot, few-shot, and supervised learning abilities. - Task-specific instruction tuning improves performance compared to other models. - Me-LLaMA models outperform ChatGPT and GPT-4 on various datasets. 💡 Main outcomes and implications: - Me-LLaMA is a significant advancement in medical language models, showcasing superior performance across general and medical tasks. - The study addresses limitations in applying LLMs to clinical settings due to a lack of specialized training on medical-specific data. - Me-LLaMA models mitigate the catastrophic forgetting problem better than other medical LLMs. 📚 Field significance: - Me-LLaMA introduces a novel approach to medical language models, leveraging large datasets and continual pre-training for improved performance. - The study highlights the potential of Me-LLaMA in revolutionizing medical applications and advancing AI in healthcare. 🗄️: [#medicalapplications #LLMs #datainsights #MeLLaMA #healthcareAI]

👉🏼 Exploring the capacities of ChatGPT: A comprehensive evaluation of its accuracy and repeatability in addressing Helicobacter pylori-related queries 🤓 Yongkang Lai 👇🏻 https://lnkd.in/eG-2XQ8s 🔍 Focus on data insights: - ChatGPT responses to H. pylori-related questions were generally satisfactory, with 61.9% marked as "completely correct" and 33.33% as "correct but inadequate." - The repeatability of ChatGPT responses to H. pylori-related questions was 95.23%. - Responses related to prevention (comprehensive: 75%) had the best response rate, followed by treatment (comprehensive: 66.7%), basic knowledge (comprehensive: 60%), and diagnosis (comprehensive: 50%). 💡 Main outcomes and implications: - ChatGPT can provide correct answers to the majority of H. pylori-related queries. - It exhibited good reproducibility and delivered easily comprehensible responses to patients. - Further enhancements in real-time information updates and correction of inaccurate data can make ChatGPT an essential tool for providing accurate H. pylori-related health information. 📚 Field significance: - Improving ChatGPT's knowledge on H. pylori resistance and the use of sensitive antibiotics can enhance its utility in providing comprehensive health information to patients. 🗄️: [#Helicobacter #ChatGPT #Hpylori #DataInsights]