Session 2
Chair: Tianxi Cai, PhD (Harvard University)
Speaker: Hongfang Liu, PhD (UT Health, Houston)
Professor, Health Data Science and Artificial Intelligence
Title: Translational Science in Healthcare AI
Abstract: The advancement of AI and digital technology provides a huge opportunity in integrating big data analytics and predictive modeling into healthcare delivery. In this talk, I will discuss the effort on advancing translational science leveraging data science, informatics, and AI and propose an implementation framework in healthcare AI. Insights from the effort of the Open Health Natural Language Processing Consortium will be shared, highlighting shifts in focus due to emerging trends and funding. Additionally, we will discuss the impact of Large Language Models (LLMs) and generative AI on healthcare, the practical challenges of deploying these technologies, and their potential to transform healthcare delivery.
Speaker: Hua Xu, PhD (Yale University)
Robert T. McCluskey Professor
Vice Chair for Research and Development, Department of Biomedical Informatics and Data Science
Assistant Dean for Biomedical Informatics, School of Medicine
Title: Building Ethical Large Language Models for Biomedical Applications
Abstract: The landscape of natural language processing (NLP) has been significantly transformed by recent advancements in Large Language Models (LLMs). In the biomedical domain, LLMs-based approaches and solutions have demonstrated its potential to revolutionize biomedical research and clinical practice. This presentation will concentrate on our recent endeavors in developing methodologies and software tailored for important biomedical applications, based on state-of-the-art LLMs. We will explore the utilization of both open-source and closed-source LLMs, including LLaMA2/3 and GPT-4, on diverse tasks such as information extraction, questions answering, and literature mining. Additionally, we will delve into ethical issues and corresponding technologies to address ethical concerns, when employing LLM-based approaches in biomedical applications.
Speaker: Marzyeh Ghassemi, PhD (Massachusetts Institute of Technology)
Associate Professor, Electrical Engineering and Computer Science
Title: The Pulse Of Ethical Machine Learning in Health
Abstract: Dr. Marzyeh Ghassemi focuses on creating and applying machine learning to understand and improve health in ways that are robust, private and fair. Dr. Ghassemi will talk about her work trying to train models that do not learn biased rules or recommendations that harm minorities or minoritized populations. The Healthy ML group tackles the many novel technical opportunities for machine learning in health, and works to make important progress with careful application to this domain.
Speaker: Jessica Gronsbell, PhD (University of Toronto)
Assistant Professor of Statistics
Title: Revisiting the Classics: Trustworthy Statistical Inference with Machine Learning Derived Data
Abstract: Predictions from machine learning models increasingly complement costly gold-standard data in scientific inquiry. Appropriately accounting for inaccuracies in these predictions is critical to achieve trustworthy conclusions from downstream statistical inference. In this talk, I will present our ongoing work on prediction-based inference methods. I will explore the impact of machine learning-derived predictions on statistical inference across various biomedical applications. I will also draw connections between recently proposed methods and classical statistical approaches dating back to the 1960s and highlight the utility of classical methods in ensuring reliable scientific research.
