Embedded Files
The importance of data management in the design of AI-based clinical decision support systems
The importance of data management in the design of AI-based clinical decision support systems
by Lucia Sacchi
by Lucia Sacchi
Abstract: The design and implementation of AI-based Clinical Decision Support System (CDSSs) is a complex process, which requires addressing a number of challenges. These are related to the importance of delivering a reliable and robust system, which is trusted and consistently adopted by its users. One of the crucial components of the creation of such systems is data collection and management. Data are indeed the main ingredient of such systems, both for their crucial role in model development, and for their use during runtime to guide clinical decisions. In this talk I will address the challenges of data collection from different sources and their harmonization, considering data quality and integrity, data protection and accessibility. I will also address the importance of taking into account real world data for research, and the challenges that this implies. I will illustrate these issues also with some examples taken from research projects.
Short CV: Lucia Sacchi is Associate Professor at the Department of Electrical, Computer and Biomedical Engineering at the University of Pavia, Italy. She's got a Master's Degree in Computer Engineering and a PhD in Bioengineering and Bioinformatics. She is the scientific coordinator of the OHDSI Italian National Node and she has been the President of the Italian Society of Biomedical Informatics (SIBIM) and Chair of the IMIA working group on Data Mining and Big Data Analytics. She is part of the Editorial Board of the Artificial Intelligence in Medicine Journal. Her research interests are related to data mining, with particular focus on temporal data, design and development of clinical decision support systems, and technologies for biomedical data analysis with a specific focus on real world data (RWD).
Explaining Health Predictions in the era of LLMs: an initial investigation
Explaining Health Predictions in the era of LLMs: an initial investigation
by Tommaso di Noia
by Tommaso di Noia
Abstract: LLMs are the new silver bullet in the application of AI in the health domain. Machine and Deep learning have proven to be very effective in health prediction tasks. However, there is still room for improvement when computing a human-friendly explanation for the predicted results. We will show the results of a preliminary investigation into the application of prompt engineering to generate textual explanations of predicted outcomes in the health domain. We will highlight challenges and open issues toward the creation of a conversational AI able to support clinicians and patients in health-related applications.
Short CV: Tommaso Di Noia is a full professor at the Polytechnic University of Bari. His current research activity focuses on artificial intelligence and data management, with reference to techniques and applications of machine learning and recommender systems. In recent years, he has delved into the topic of trustworthy artificial intelligence, with a focus on fairness, security, explanation, and privacy. Recently, applications of AI-based healthcare approaches have been presented with an emphasis on explainable AI and conversational agents. He has also begun to study the information encoded in brain-generated signals and its application in healthcare scenarios and creative contexts such as automatic music generation.
Page updated
Google Sites
Report abuse