Congratulations to Chaoyue Sun, a Ph.D. candidate in the Smart Medical Informatics Learning & Evaluation (SMILE) Lab, on receiving the Certificate of Outstanding Merit from the UF International Center. @RuoguFang

A Comprehensive Survey of Foundation Models in Medicine 1/ This survey provides an extensive review of foundation models (FMs) in medicine, covering their evolution, learning strategies, flagship models, applications, and challenges. It demonstrates how large-scale pretrained models like BERT and GPT are transforming healthcare. 2/ Foundation models (FMs) are revolutionizing healthcare by providing a versatile base for tasks like clinical language modeling, medical image analysis, and omics research. This paper presents a detailed taxonomy of FM-enabled applications, spanning NLP, computer vision, graph learning, and more. 3/ Key models such as GatorTronGPT, MedSAM, and BioBERT are specifically designed for healthcare, achieving remarkable success in clinical NLP tasks like named entity recognition (NER), question answering (QA), and medical relation extraction (RE), alongside their contributions to medical image segmentation. 4/ One of the significant challenges FMs face in healthcare is the lack of high-quality, labeled medical data. The paper delves into strategies to overcome this limitation, including the use of self-supervised learning (SSL) techniques and the creation of domain-specific corpora like PubMed and MIMIC-III. 5/ The paper also discusses the increasing need for specialized FMs in healthcare due to the complexity of medical data and the fine-grained differences between general-purpose and domain-specific applications. It highlights models like BioMegatron, Med-PaLM, and GeneGPT designed to cater specifically to medical datasets. 6/ A major opportunity for FMs in medicine is their ability to be generalized across different tasks. For example, GatorTronGPT is trained on a massive corpus of clinical texts and performs well across multiple medical tasks, illustrating the versatility of large-scale language models in clinical settings. 7/ Despite their transformative potential, the paper identifies several challenges, including high computational costs, model interpretability, and the risk of social biases. These issues must be addressed to ensure responsible deployment and to enhance trust in FM-based healthcare solutions. 8/ The survey concludes by outlining open research questions and potential directions for the future of FMs in healthcare. It emphasizes the need for robust, scalable models capable of handling diverse medical data while ensuring privacy, security, and ethical considerations. @RuoguFang 📜Paper: #FoundationModels #HealthcareAI #MedicalAI #MachineLearning #Bioinformatics #ClinicalAI #MedicalImageAnalysis #Omics

🙏 Thank you Dr. Maria A. Zuluaga @marazul for your informative presentation! 👏 📢 The recording and presentation slides are now available on the Women in MICCAI YouTube channel 🔗

Congratulations to Joseph Cox, recipient of the 2024 Supervised Teaching Excellence Award! Your commitment to mentoring and excellence is truly commendable.

Congratulations to Skylar Stolte and Suzanne Lightsey, the 2024 Graduate Excellence Awardees! Your achievements make the UF BME community incredibly proud.

Congratulations to Dr. Ruogu Fang, the 2024 Faculty Teaching Excellence Awardee! Your passion for teaching continues to inspire and empower the next generation. @RuoguFang

Women in MICCAI (#WiM) is happy to present our slack channel. Please join us to make our community stronger. Invitation Link:

Women in #MICCAI (#WiM) is happy to announce the second invited talk in the "Health Equity in the AI Era" Webinar Series 2024 by Dr. Maria Zuluaga (@marazul ) from @KingsCollegeLon on "Building Trustworthiness through Robustness" on Monday, December 9, 2024, 4:00 - 5:00 PM UTC.