Prompt Engineering Paradigms for Medical Applications: Scoping Review

Journal of Medical Internet Research, Journal Year: 2024, Volume and Issue: 26, P. e60501 - e60501

Published: Sept. 10, 2024

Prompt engineering, focusing on crafting effective prompts to large language models (LLMs), has garnered attention for its capabilities at harnessing the potential of LLMs. This is even more crucial in medical domain due specialized terminology and technicity. Clinical natural processing applications must navigate complex ensure privacy compliance. engineering offers a novel approach by designing tailored guide exploiting clinically relevant information from texts. Despite promise, efficacy prompt remains be fully explored.

Language: Английский

---

Comparative Efficacy of AI LLMs in Clinical Social Work: ChatGPT-4, Gemini, Copilot

Research on Social Work Practice, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 17, 2025

Purpose This study examines the comparative efficacy of three AI large language models (LLMs)—ChatGPT-4, Gemini, and Microsoft Copilot—in clinical social work. Method By presenting scenarios varying complexities, assessed their performance using Ateşman Readability Index a Likert-type accuracy scale. Results showed that Gemini had highest accuracy, while Copilot excelled in readability. Significant differences were found scores ( p = .003), although readability not statistically significant .054). No correlation was between case complexity either or Discussion Despite differences, none fully met all standards, indicating areas for further improvement. The findings suggest LLMs offer promise work, they require refinement to better meet field's needs.

Language: Английский

---

Development and Evaluation of an Auditory VR Generative System via Natural Language Interaction to Aid Exposure Therapy for PTSD Patients

ACM Transactions on Computing for Healthcare, Journal Year: 2025, Volume and Issue: unknown

Published: March 19, 2025

Post-Traumatic Stress Disorder (PTSD) is a prevalent disorder triggered by life-threatening trauma, and exposure therapy, which involves confronting traumatic stimuli, has been proven to be highly effective for treating PTSD. However, therapy not widely adopted. Virtual Reality (VR) shown comparable effectiveness that of traditional methods, therefore advancing. this broadly implemented, partly because the time required create VR experiences tailored patient’s specific trauma. To address problem, study proposes system generates auditory using Large Language Model (LLM) natural language interaction. This system, built on LLM an audio dataset, sounds matching user-provided themes corresponding scenarios coordinates. An experiment with clinicians generate stimuli was conducted assess usability therapeutic potential generated audio. The results indicated high quality, requiring minimal adjustments applications. Notably, within duration standard clinical session. challenges remain, particularly complex themes, highlighting need further research enhance verify system’s feasibility efficacy.

Language: Английский

---

Medical students’ perception of the use of artificial intelligence in medical education

International Journal of Research in Medical Sciences, Journal Year: 2024, Volume and Issue: 13(1), P. 82 - 89

Published: Dec. 31, 2024

Background: Artificial Intelligence (AI) refers to technology that can efficiently perform tasks typically require human intelligence, such as decision-making, teaching, object detection, and solving complex problems. As a veritable tool in medical education, this study was conducted assess students’ awareness, perception usage of AI learning. Methodology: This at the Department Paediatrics, University Port Harcourt Teaching Hospital (UPTH). The subjects were one hundred thirty-nine 5th-year students who had completed 3 months Paediatrics Obstetrics/Gynaecology clinical rotations. Data collected using semi-structured, open-ended questionnaire. analysed IBM SPSS Statistics version 26. Statistical significance set p value<0.05. Results: 64 (46%) respondents are aware AI. 57 (44%) applied during their training. most commonly used is Chatbots. 57(100%). major limitations use unreliable internet connectivity (62%) high cost hardware software (53%). Most (68%) expressed ethical concerns about There statistically significant relationship between awareness learning (p=0.0001) Conclusion: demonstrates average AI's benefits among students. unlimited tools. To maximize education developing countries, schools need increase infrastructural capacity.

Language: Английский

---

Comparative analysis of artificial intelligence-driven assistance in diverse educational queries: ChatGPT vs. Google Bard

Frontiers in Education, Journal Year: 2024, Volume and Issue: 9

Published: Sept. 26, 2024

Artificial intelligence tools are rapidly growing in education, highlighting the imperative need for a thorough and critical evaluation of their performance. To this aim, study tests effectiveness ChatGPT Google Bard answering range questions within engineering health sectors. True/false, multiple choice (MCQs), matching, short answer, essay, calculation among question types investigated. Findings showed that 4 surpasses both 3.5 terms creative problem-solving accuracy across various types. The highest achieved by was true/false questions, reaching 97.5%, while its least accurate performance noted with an 82.5%. Prompting to provide responses apparently prevented them from hallucinating unrealistic or nonsensical responses. majority problems which provided incorrect answers demonstrated correct approach; however, AI models struggled accurately perform simple calculations. In MCQs related sciences, seemed have challenge discerning answer several plausible options. While all three managed essay competently, avoiding any blatantly (unlike other types), some nuanced differences were noticed. consistently adhered more closely prompts, providing straightforward essential responses, superiority over adaptability. ChatGPT4 fabricated references, creating nonexistent authors research titles response prompts sources. utilizing education holds promise, even latest most advanced versions not able questions. There remains significant human cognitive skills further advancements capabilities.

Language: Английский

---

Bibliometric Analysis of Studies on Chat GPT with Vosviewer

Black Sea Journal of Engineering and Science, Journal Year: 2024, Volume and Issue: 7(6), P. 1194 - 1203

Published: Oct. 9, 2024

Chat GPT, which suddenly entered our lives at the end of 2022, is a large language model and has attracted attention people from all walks life. Following this, researchers have also conducted studies in many areas related to Chat-GPT literature. For this purpose, bibliometric data different types works scanned Web Science database published between 2020-2024 was taken as basis. When we look distribution 415 GPT according publication years, it seen that most were produced 2024 (222 works), 2023 (186 works) each other years (2 works). It there been concentration last two years; authors who are Viroj Wiwanitkit, Wisit Cheungpasitporn Jerome R. Lechien; The type mainly journal article (286), paper (48), early appearance (45), review (30) letter(6); top five research fields examined, found computer science (101), general medicine (50), educational (32), surgery (26); leadership publications by country with publishers USA (123), China (47), India (33); primarily English (405), Spanish (6), Portuguese (3); SCI-E (227), ESCI (115), SSCI (45) indexes predominate. frequently used keywords following leading expressions: “artificial intelligence” 101 occurrences, “Chat gpt” 90 “large models” 43 “chat 42 “ai” 29 “chat-gpt” 27 “chatbot” 17 occurrences.

Language: Английский

---