放射学中大型语言模型的最佳实践

Christian Bluethgen¹, Dave Van Veen¹, Cyril Zakka¹+7

• ¹From the Stanford Center for Artificial Intelligence in Medicine and Imaging, Palo Alto, Calif (C.B., D.V.V., C.P.L., S.G., A.C.); Institute for Diagnostic and Interventional Radiology, University Hospital Zurich, University of Zurich, Rämistrasse 100, 8005 Zurich, Switzerland (C.B., T.F.); Department of Electrical Engineering, Stanford University, Stanford, Calif (D.V.V.); Department of Cardiothoracic Surgery, Stanford Medicine, Stanford, Calif (C.Z.); Department of Medical Education, Icahn School of Medicine at Mount Sinai, New York, NY (K.E.L.); NVIDIA, New York, NY (K.E.L.); UT Health San Antonio, San Antonio, Tex (A.H.F.); Department of Biomedical Data Science, Stanford Medicine, Stanford, Calif (A.H.F., R.D., C.P.L., A.C.); Department of Dermatology, Stanford Medicine, Redwood City, Calif (R.D.); Department of Medicine, Stanford Medicine, Stanford, Calif (C.P.L., A.C.); and Department of Radiology, Stanford University, Stanford, Calif (C.P.L., S.G., A.C.).