{"693303":{"#nid":"693303","#data":{"type":"event","title":"ISyE Statistics Seminar - Xiao Hu","body":[{"value":"\u003Cp\u003ETitle: UNIPHY: A Unified Prediction and Forecasting Architecture Using Physiological Signals\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EAbstract: Despite the ubiquity of continuous physiological monitoring in critically ill patients, clinicians lack tools to use high-resolution electrocardiogram (ECG) and photoplethysmography (PPG) signals in clinical care, especially combined with the context of patient characteristics, treatments, and evolving clinical state.\u0026nbsp; As a result, current patient ECG and PPG data remain largely under-utilized in clinical practice. Artificial intelligence (AI) methods for ECG and PPG signals could address this gap, but current approaches are often task-specific, of a short-window nature, and insufficiently contextualized for critical care. \u0026nbsp;As a result, they typically do not model long-timescale physiologic dynamics or account for how comorbidities, treatments, and evolving illness shape observed signals and outcomes. To address these methodological gaps, we are developing the \u003Cstrong\u003EUNIPHY\u003C\/strong\u003E framework, which combines a pretrained physiologic foundation model (physioFM) \u0026nbsp;for local signal representation with an efficient state-space model (SSM) for long-range temporal integration and contextualized patient modeling. In this talk, I will share our most recent progress and lessons learned in training and testing UNIPHY on multiple use cases in various inpatient care settings, and example ambulatory monitoring applications.\u003C\/p\u003E\u003Cp\u003EBio: Dr. Xiao Hu is a Professor at the Nell Hodgson Woodruff School of Nursing, an associated professor at department of Computer Sciences at College of Arts and Sciences and a program faculty member of the joint GA Tech and Emory Biomedical Engineering department. Dr. Hu is also the Associate Director of the Center for Data Science. Dr. Hu earned bachelor\u2019s and master\u2019s of science degrees in biomedical engineering at the University of Electronic Science and Technology in Chengdu, China. He then studied at UCLA, receiving a PhD in biomedical engineering in 2004. Before joining Emory, he held faculty positions at UCLA, UCSF, and Duke University.\u003C\/p\u003E\u003Cp\u003EHu conducts research at the intersection of computational and health sciences where he develops signal processing, mathematical modeling, machine learning, and deep learning algorithms to transform physiological and clinical data into actionable information for improving patient care. Through his research, Hu has created knowledge advancing the relationship among clinicians, scientists, engineers, innovators, policy makers, and healthcare administrators.\u003C\/p\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003EDespite the ubiquity of continuous physiological monitoring in critically ill patients, clinicians lack tools to use high-resolution electrocardiogram (ECG) and photoplethysmography (PPG) signals in clinical care, especially combined with the context of patient characteristics, treatments, and evolving clinical state.\u0026nbsp; As a result, current patient ECG and PPG data remain largely under-utilized in clinical practice. Artificial intelligence (AI) methods for ECG and PPG signals could address this gap, but current approaches are often task-specific, of a short-window nature, and insufficiently contextualized for critical care. \u0026nbsp;As a result, they typically do not model long-timescale physiologic dynamics or account for how comorbidities, treatments, and evolving illness shape observed signals and outcomes. To address these methodological gaps, we are developing the \u003Cstrong\u003EUNIPHY\u003C\/strong\u003E framework, which combines a pretrained physiologic foundation model (physioFM) \u0026nbsp;for local signal representation with an efficient state-space model (SSM) for long-range temporal integration and contextualized patient modeling. In this talk, I will share our most recent progress and lessons learned in training and testing UNIPHY on multiple use cases in various inpatient care settings, and example ambulatory monitoring applications.\u003C\/p\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"UNIPHY: A Unified Prediction and Forecasting Architecture Using Physiological Signals "}],"uid":"36873","created_gmt":"2026-10-19 17:53:44","changed_gmt":"2026-10-19 17:58:26","author":"csheppard39","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2026-10-28T11:00:00-04:00","event_time_end":"2026-10-28T12:00:00-04:00","event_time_end_last":"2026-10-28T12:00:00-04:00","gmt_time_start":"2026-10-28 15:00:00","gmt_time_end":"2026-10-28 16:00:00","gmt_time_end_last":"2026-10-28 16:00:00","rrule":null,"timezone":"America\/New_York"},"location":"George Tower 1502","extras":[],"groups":[{"id":"1242","name":"School of Industrial and Systems Engineering (ISYE)"}],"categories":[],"keywords":[],"core_research_areas":[],"news_room_topics":[],"event_categories":[{"id":"1795","name":"Seminar\/Lecture\/Colloquium"}],"invited_audience":[{"id":"194945","name":"Alumni"},{"id":"78761","name":"Faculty\/Staff"},{"id":"177814","name":"Postdoc"},{"id":"78771","name":"Public"},{"id":"174045","name":"Graduate students"},{"id":"78751","name":"Undergraduate students"}],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[],"email":[],"slides":[],"orientation":[],"userdata":""}}}