2023 CHOICE Institute Annual Symposium: скачать в хорошем качестве

2023 CHOICE Institute Annual Symposium:

The Comparative Health Outcomes, Policy, & Economics (CHOICE) Institute 5th Annual Research Symposium Artificial Intelligence, Real Consequences: The side effects of AI in healthcare Date: Thursday, October 12, 2023 Speakers: -Judy Wawira Gichoya, MD, MS, Associate Professor – Interventional Radiology & Informatics, Emory University School of Medicine Talk title: Hidden in plain sight : Harnessing AI's ability for population health from medical images Dr. Judy Gichoya will explore the revolutionary potential of using artificial intelligence to derive and analyze population health data from medical images, highlighting the boundless possibilities of image-based prediction. She will emphasize the crucial need to confront and correct biases within AI systems, ensuring that they offer accurate and equitable insights and evaluations of diverse populations. Concluding her discussion, Dr. Gichoya will outline the anticipated advancements and necessary future work in this field, focusing on enhancing AI methodologies and fostering interdisciplinary collaborations to advance population health management. -Noémi Kreif, PhD, Senior Research Fellow, Center for Health Economics, University of York Talk Title: Machine learning to answer causal questions: how to avoid misleading interpretation? Dr. Noémi Kreif will first discuss the risks of the naive use of machine learning prediction models to inform treatment decisions, using published examples from clinical medicine. Next, she will bring examples from her own research on data-driven health policy targeting rules, showing the importance of careful adjustment for confounding via causal machine learning. Her case study will demonstrate the potential for even state of the art methods to provide counterintuitive and even harmful recommendations, and will discuss the utility of interpretable ML models in avoiding misleading results. -Michael Willingham, Director of Regulatory Affairs, Google Health Talk Title: Considerations for Large Language Models and Generative AI in Medical Applications Recent advancements in artificial intelligence (AI) have led to the public introduction of sophisticated large language models (LLMs) such as GPT-4 and Bard. The application of LLMs in healthcare settings has drawn considerable attention because of their transformative potential, including use cases for clinical documentation, summarizing research information and chatbots for clinical decision making. However, LLMs pose a unique set of risks and considerations, especially within the critical context of patient care, since these models are trained differently from previous AI and machine learning models. Michael will present on these topics using Google's research for Med-PaLM as a case study. Annual Symposiums: https://sop.washington.edu/choice/who...