Current Events

Date: Tuesday, February 3, 2026 | 12:00PM-1:00PM, BMI Classroom 4004 or on Zoom
Speakers: Satish E. Viswanath, PhD
               Associate Professor, Pediatrics and Biomedical Engineering, Emory University

Abstract: Developing artificial intelligence (AI) schemes to assist the clinician towards enabling precision medicine approaches requires development of objective markers that are predictive of disease response to treatment or prognostic of longer-term patient survival. The solutions being developed in my group in this regard involve designing computational imaging features together with histology or molecular data for detailed tissue and disease characterization in vivo as well be associated with patient outcomes. The key innovation in this approach lies in “handcrafting” unique tools that can capture biologically relevant and clinically intuitive measurements from routinely acquired imaging (MRI, CT, PET) or digitized images of tissue specimens. Further, by conducting cross-scale associations between imaging, pathology, and -omics, we can not only “unlock” and integrate the information captured by these different, disparate data modalities but also develop an interpretable and intuitive understanding of what drives their performance. Specific problems addressed via the new computerized imaging markers we have developed include: (a) predicting response to treatment to identify optimal therapeutic pathways, as well as (b) evaluating therapeutic response to guide follow-up procedures. We will further examine how to account for differences between sites, scanners, and acquisition parameters to ensure generalizable performance of AI tools and computational imaging features; crucial for wider clinical translation and widespread adoption. These will be discussed in the context of clinical applications in colorectal and renal cancers, digestive diseases, as well as pediatric conditions.

Date: Tuesday, November 4, 2025 | 12:00PM-1:00PM, BMI Classroom 4004 or on Zoom
Speakers: R. James Cotton, MD, PhD
                Assistant Professor, Physical Medicine and Rehabilitation
                Assistant Professor, McCormick School of Engineering
                Northwestern University
 
Abstract: This talk will review recent advances in AI-enabled biomechanical human pose estimation and analysis. It will first cover differentiable biomechanics for end-to-end recovery of kinematics with confidence intervals and kinetics. It will show how this approach can be extended to track high quality arm and hand biomechanics, and to monocular whole-body gait analysis. It will also demonstrate new opportunities that emerge with large rehabilitation data such as imitation learning and training multimodal language models that are fluent in human movement.

Date: Tuesday, October 21, 2025 | 12:00PM-1:00PM, BMI Classroom 4004 or on Zoom
Speakers: Carl Yang, PhD
                Assistant Professor, Department of Computer Science
                Emory University
 
Abstract: Large language models (LLM) have brought disruptive progress to information technology from accessing data to performing analytical tasks. While demonstrating unprecedented capabilities, LLMs have been found unreliable in tasks requiring factual knowledge and rigorous reasoning, posing critical challenges in domains such as healthcare. Knowledge graphs (KG) have been widely used for explicitly organizing and indexing biomedical knowledge, but the quality and coverage of KG are hard to scale up given the notoriously complex and noisy healthcare data with multiple modalities from multiple institutions. Existing approaches show promises in combining LLMs and KGs to enhance each other, but they do not study the techniques in real healthcare contexts and scenarios. In this talk, I will introduce our research vision and agenda towards KG-LLM co-learning for healthcare, followed by successful examples from our recent exploration on LLM-aided KG construction, KG-guided LLM enhancement, and federated multi-agent systems. I will conclude the talk with discussions on future directions that can benefit from further collaborations with researchers interested in data mining or biomedical informatics in general.

Date: Tuesday, October 7, 2025 | 12:00PM-1:00PM, BMI Classroom 4004 or on Zoom
Speakers: Ramakanth Kavuluru, PhD
                Professor, Division of Biomedical Informatics
                Department of Internal Medicine
                University of Kentucky
 
Abstract: While LLMs are dominating the field of NLP, biomedical information extraction (named entity recognition, relation extraction, slot filling) remains a relevant task for knowledge discovery. A recent example is the use of knowledge graphs extracted from scientific literature to facilitate grounding and mitigate hallucinations in LLMs. In this context, are tailored supervised models needed for information extraction or can we simply use zero-shot extraction from frontier LLMs? What about encoders (e.g., BERT) vs decoders (e.g., GPT-x)? Are we better off pretraining biomedical language models from scratch or can we simply continue pretraining on top of general LMs? Culling examples from my lab’s work in recent years, Dr. Kavuluru will try to answer these questions and present recent (and likely future) trends in biomedical information extraction.

Date: Tuesday, September 16, 2025 | 12:00PM-1:00PM, BMI Classroom 4004 or on Zoom
Speakers: Muhammed Idris, PhD
                Assistant Professor for Department of Medicine
                Morehouse School of Medicine
 
Abstract: This talk explores how biomedical informatics can be applied to investigate and address health disparities. I will introduce the Morehouse Model 3.0—a framework that combines community engagement, data science, and implementation research to develop culturally responsive, community-centered digital health interventions. The talk will feature two in-depth case studies: one on natural language processing (NLP) for clinical trial recruitment and another on EHR-based heart failure risk prediction, highlighting research and lessons learned in designing, developing, and evaluating AI/ML systems in resource-constrained settings. I will conclude with research areas and projects that I am excited about including LLM-based risk prediction, Small Language Models for health communication, and real-world performance evaluations of algorithms deployed in clinical and community health settings. Finally, I talk will also highlight key ethical considerations and how community partnerships are essential to implementing responsible AI and can help ensure equitable outcomes in health informatics.

Date: Tuesday, August 19, 2025 | 12:00PM-1:00PM, BMI Classroom 4004 or on Zoom
Speakers: Temiloluwa Prioleau 
                 Assistant Professor of Computer Science
                 Emory University 
Abstract: The prevalence of digital technology in daily living creates astounding amounts of user-generated data that can be helpful for understanding complex health conditions in a new way. These large data sources are also foundational for developing AI-driven solutions that are necessary to realize personalized healthcare. Yet, significant gaps exist around data analytic solutions capable of eliciting clinically meaningful and actionable insights from continuously generated user data. Additionally, challenges exist around the reproducibility of health-centered AI methods in literature.

To bridge these gaps, this talk will present our research efforts on opportunistic use of digital health data in the context of diabetes management. I will discuss how large volumes of user-generated data can improve our understanding of unique lifestyle factors that play a critical role in outcomes. The talk will be organized around three core topics including context-aware sensing for digital health interventions, wearable data analytics and AI for personalized care, and reproducibility in machine learning research for health. Finally, I will present concluding thoughts on how these research efforts can be extended to enable digital and computational health in other application domains.

Date: Tuesday, June 3, 2025 | 12:00PM-1:00PM, BMI Classroom 4004 or on Zoom
Speakers: David L. Minkoff, MD 
                 Assistant Professor, Division of Hospital Medicine
                 Emory University School of Medicine
Abstract: Although Hospital Medicine is now one of the largest physician specialties in the country, research into hospitalist practice patterns and performance remains limited.  Significant variability exists in Hospital Medicine outcomes; for example, between two hospitalists in a given practice, readmission rates and length-of-stay can vary by as much as a factor of two. Data to explain such variability are extremely limited. Previously, research into clinical practice patterns was limited both by the data available in paper charts and the need for manual chart reviews. However, the widespread adoption of electronic health records (EHRs) in the past decade has allowed for the study of novel data types, such as audit logs, and recent advances in natural language processing will reduce the need for manual chart reviews, which will allow for superior studies in both detail and scale. Development of next-generation performance metrics, both for individual physicians and for the practice environment at large, will allow for improved patient care and bolster the financial sustainability of the healthcare system. Additional aims of the current work include:

• Development of novel clinical practice metrics incorporating EHR metadata to identify "practice phenotypes"
• Addition of data from clinical notes using natural language processing for clinical decision support
• Development of hospital state metrics to guide operations and quality improvement initiatives