Towards Predictive Communication: The Fusion of Large Language Models and Brain-Computer Interface.
邁向預測性溝通:大型語言模型與腦機介面(Brain-Computer Interface, BCI)的融合
Sensors (Basel) 2025-07-12
Chatbots in Radiology: Current Applications, Limitations and Future Directions of ChatGPT in Medical Imaging.
放射科中的聊天機器人:ChatGPT 在醫學影像的現有應用、限制與未來發展方向
Diagnostics (Basel) 2025-07-12
Exploring the Potential of ChatGPT-4o in Thyroid Nodule Diagnosis Using Multi-Modality Ultrasound Imaging: Dual- vs. Triple-Modality Approaches.
運用多模態超音波影像探討 ChatGPT-4o 在甲狀腺結節診斷的潛力:雙模態與三模態方法之比較
Cancers (Basel) 2025-07-12
Evaluating the Accuracy of Privacy-Preserving Large Language Models in Calculating the Spinal Instability Neoplastic Score (SINS).
具備隱私保護的大型語言模型在計算脊椎不穩定腫瘤評分(SINS)準確性的評估
Cancers (Basel) 2025-07-12
這項研究發現,Claude 3.5 在從放射科報告計算脊椎不穩定腫瘤評分(SINS)時,準確度跟臨床醫師差不多,表現也比 Llama 3.1 好,幾乎和專家標準一致。顯示先進的語言模型有機會協助醫師進行 SINS 評估,但還需要更多驗證才能正式應用在臨床上。
相關文章PubMedDOI推理
AI's ability to interpret unlabeled anatomy images and supplement educational research as an AI rater.
AI 解讀未標註解剖影像及作為 AI 評分者輔助醫學教育研究之能力
Anat Sci Educ 2025-07-11
這項研究發現,ChatGPT-4o 和 Claude 3.5 Sonnet 在解讀未標註解剖圖的正確率都只有中等,差異不大。雖然 ChatGPT o1-preview 當 AI 評分員時,和專家評分有不錯的一致性,但目前這些 AI 還不夠穩定,暫時不適合直接用在解剖學教學。不過,AI 評分助手在教育研究上有發展潛力,未來還需要再改進。
相關文章PubMedDOI推理