The Speakers

Abstract: Large language models (LLMs) have demonstrated substantial potential in the health domain, with specialized models increasingly developed for a wide range of tasks. In this talk, I will cover: (1) systematic benchmarking of LLMs across biomedical applications, (2) advancing factuality and evidence attribution for LLMs in medicine, and (3) developing domain-specific LLMs, with a focus on ophthalmology-related clinical tasks and automated response generation for outpatient queries.

Bio: Dr. Qingyu Chen is a tenure-track Assistant Professor in the Department of Biomedical Informatics & Data Science at Yale University. Prior to joining Yale, he completed postdoctoral training at the National Library of Medicine, National Institutes of Health. His research focuses on data science and artificial intelligence in biomedicine and healthcare, with emphasis on three main areas: biomedical natural language processing and large language models; medical imaging and multimodal analysis; and accountability and trustworthy AI in medical applications.

Dr. Chen is the Principal Investigator of an R01 grant focused on improving the factuality of LLMs in medicine, as well as a K99/R00 grant on multimodal AI-assisted disease diagnosis. He has authored over 45 first/last-author publications among 100+ peer-reviewed papers.

Abstract: Large language models (LLMs) are increasingly used as “simulated subjects” and have the potential to make social science research faster and cheaper. However, how well these simulations capture real human behavior depends on many design choices. This talk discusses several challenges in generating LLM-based simulations, including prompt ambiguity and missing persona information. We then outline an evaluation framework for comparing LLM behavior to human data and for understanding when simulation errors matter for the research question. Finally, we highlight promising applications where LLMs can be safely and productively used despite imperfect realism.
Bio: https://www.georgegui.com/

Abstract:  Large Language Models have reshaped the landscape of artificial intelligence research. Large Language Models are simultaneously very powerful and yet, at times, incredibly brittle. One particular area of research, automated story generation, seems to expose this brittleness because story generation requires planning in language, commonsense  reasoning, cultural reasoning,  theory of mind, emotion modeling, and more. LLMs have uneven competencies across this list of capabilities. In this talk, I will attempt to draw some conclusions about what is missing from LLMs in the grand pursuit of artificial intelligence, and some possible paths forward.
Bio: Dr. Mark Riedl is a Professor in the Georgia Tech School of Interactive Computing and Associate Director of the Georgia Tech Machine Learning Center. Dr. Riedl’s research focuses on human-centered artificial intelligence—the development of artificial intelligence and machine learning technologies that understand and interact with human users in more natural ways. Dr. Riedl’s recent work has focused on story understanding and generation, computational creativity, explainable AI, and teaching virtual agents to behave safely.