医師の監督下における会話型医療AI：実践への準備と患者体験
医師の監督下にある会話型AI医療アシスタントは、患者の満足度と医療情報の明確さを向上させながら、安全性を維持できる可能性を示している。
환자 사실 확인 AI의 윤리적 문제: 환자의 말보다 데이터를 우선시하는 경향
인공지능(AI)을 사용하여 환자의 자가 보고 정보를 검증하는 것은 환자의 개인정보보호와 의료진에 대한 신뢰를 훼손할 수 있으며, AI 시스템이 환자의 말보다 데이터 또는 다른 AI 모델의 예측을 우선시하는 "AI 자기 신뢰" 편향을 보일 수 있다.
患者の発言をAIで「事実確認」：倫理的な懸念
AIを用いて患者の自己申告情報を検証するシステムは、患者と医療従事者の信頼関係を損ない、医療における不平等を悪化させる可能性がある。
Conversational Disease Diagnosis via External Planner-Controlled Large Language Models
This study introduces an innovative approach using external planners augmented with large language models (LLMs) to develop a medical task-oriented dialogue system for conversational disease diagnosis. The system comprises a policy module for information gathering, and an LLM-based module for natural language understanding and generation, addressing the limitations of previous AI systems in these areas.
Enhancing Medical Instruction-Following Ability of Large Language Models through Diverse Machine-Generated Dataset
Tuning large language models with a diverse, machine-generated medical instruction-response dataset, MedInstruct-52k, can significantly boost their performance on medical applications while also improving their generalizability.
LLM-based Personalized Medical Assistant with Dual-Process Enhanced Memory and Parameter-Efficient Fine-Tuning
A novel computational bionic memory mechanism, equipped with a parameter-efficient fine-tuning schema, is proposed to personalize medical assistants and enhance their response quality by catering to user-specific needs.
Meerkat-7B: An Open-Source Medical AI System with Enhanced Reasoning Skills from Textbooks
Meerkat-7B, a novel 7-billion parameter open-source medical AI system, achieves state-of-the-art performance on medical benchmarks by leveraging chain-of-thought reasoning data synthesized from medical textbooks.
Addressing Fairness in Vision-Language Models for Equitable Medical Diagnosis
Significant biases exist in widely-used vision-language models, with Asian, Male, Non-Hispanic, and Spanish being the preferred subgroups across the protected attributes of race, gender, ethnicity, and language, respectively. FairCLIP, an optimal transport-based approach, achieves a favorable trade-off between performance and fairness by reducing the Sinkhorn distance between the overall sample distribution and the distributions corresponding to each demographic group.
Improving Cancer Diagnosis Accuracy through Bayesian Deep Learning Approaches
Combining the strengths of Bayesian Networks and Deep Learning models can improve the accuracy and reliability of cancer imaging diagnosis.
ArgMed-Agents: Enhancing Clinical Decision Reasoning with Argumentation Schemes
ArgMed-Agents improves clinical decision reasoning accuracy and explainability through argumentation schemes.
© 2024 by Linnk AI