Keynotes

Waldemar von Zedtwitz Professor of Medicine and Biomedical Informatics and Data Science; Deputy Dean for Biomedical Informatics; Chair, Department of Biomedical Informatics and Data Science, Yale School of Medicine

Monday | October 27, 2025 | 10:00 – 11:00

Abstract: I will review the progression of AI systems for medical applications, from early knowledge-based systems focused on a narrow medical domain, to data-driven models requiring extensive training and adaptation, to the current state where sophisticated, comprehensive AI systems can be utilized in various medical centers with little to no additional training. The adoption of AI in academic medical institutions is substantial, and the accumulated experience over the past few years allows us to discern the areas where AI has had a positive impact, been neutral, or even failed, depending on the perspectives of different users. I will also discuss how some regulatory systems have not evolved at the same pace, creating gaps in how health information can be used responsibly to train and validate various AI models. I will highlight issues related to potential privacy breaches, as well as tradeoffs between easily reproducing research results and the need to protect patient data from misuse.

Biography:

Lucila Ohno-Machado, MD, PhD, MBA, is the Deputy Dean for Biomedical Informatics and the Chair of Biomedical Informatics and Data Science. As Deputy Dean for Biomedical Informatics, Ohno-Machado oversees the infrastructure related to biomedical informatics research across the academic health system.

Biomedical Informatics and Data Science serves as the hub for biomedical collaboration at Yale. It brings informatics to the clinic and the bedside; innovates new approaches to the analysis of big data across the biomedical research spectrum from basic genetic, proteomic, cellular, and systems biology to the understanding of the social determinants of health; and works in concert with colleagues in data science.

Ohno-Machado was health sciences associate dean for informatics and technology, founding chief of the Division of Biomedical Informatics in the Department of Medicine, and distinguished professor of medicine at the University of California San Diego (UCSD). She also was founding chair of the UCSD Health Department of Biomedical Informatics and founding faculty of the UCSD Halicioğlu Data Science Institute in La Jolla, California. She received her medical degree from the University of São Paulo, Brazil; her MBA from the Escola de Administração de São Paulo, Fundação Getúlio Vargas, Brazil; and her PhD in medical information sciences and computer science at Stanford University. She has led informatics centers that were funded by various NIH initiatives and by agencies such as AHRQ, PCORI, and NSF.

She organized the first large-scale initiative to share clinical data across five UC medical systems and later extended it to various institutions in California and around the country. Prior to joining UCSD, she was distinguished chair in biomedical informatics at Brigham and Women’s Hospital, and faculty at Harvard Medical School and at MIT’s Health Sciences and Technology Division. She is an elected member of the National Academy of Medicine, the American Society for Clinical Investigation, the American Institute for Medical and Biological Engineering, the American College of Medical Informatics, and the International Academy of Health Sciences Informatics. She is a recipient of the American Medical Informatics Association leadership award, as well as the William W. Stead Award for Thought Leadership in Informatics.

Long fascinated by the combination of life science and computer science, Ohno-Machado has conducted research in predictive models and data sharing since the start of her career. Her doctoral thesis work involved neural network models for survival analysis, and she subsequently focused on new methods to evaluate predictive performance of models based on clinical and molecular data. Since AI models require large amounts of data, and institutions prefer to keep the data locally, she worked on innovative algorithms to distribute the computation so that data could stay local, but multivariate models could be built and evaluated in a federated manner.

University of Ioannina / BRI – FORTH

Monday | October 27, 2025 | 13:30 – 14:30

Abstract: The integration of Artificial Intelligence into critical domains, like media, education, healthcare, industry, and beyond, requires not only technical robustness but also societal acceptance grounded in trust. The FAITH project focuses on the design and development of a holistic and human-centric, risk-based, approach to assess and optimize AI trustworthiness, aligning with emerging EU regulations, the NIST AI Risk Management Framework, and ENISA guidelines. To this end, we introduce the FAITH AI Trustworthiness Assessment Framework (FAITH AI_TAF), a structured methodological framework that systematically evaluates socio-technical, legal, and ethical risks throughout the whole AI lifecycle. The FAITH_TAF addresses gaps often overlooked in conventional models by integrating psychological and behavioral profiling of AI participants with traditional risk assessment techniques. Central to this strategy are three operational tools developed to support practitioners: (i) the FAITH AI TrustGuard, a dynamic evaluation engine that provides a checklist-based in isolation trustworthiness risk assessments, (ii) the FAITH AI TrustSense, a profiling tool that captures human-centric vulnerabilities, such as cognitive biases and AI trustworthiness awareness, integrating them into the trustworthiness equation, and (iii) the FAITH AI Model Hub, a metadata repository and operational interface that documents the development and deployment lifecycle of AI systems, promoting transparency and traceability. Through its validation across seven large-scale pilots (LSPs) in diverse sectors, FAITH demonstrates how tailored, risk-based interventions can foster more transparent, explainable, and resilient AI systems.

Biography: Prof. Dimitrios I. Fotiadis, received the Diploma degree in chemical engineering from the National Technical University of Athens, Athens, Greece, and the Ph.D. degree in chemical engineering and materials science from the University of Minnesota, Minneapolis. He is currently a Professor of Biomedical Engineering in the Department of Materials Science and Engineering, University of Ioannina, Ioannina, Greece, where he is also the Director of the Unit of Medical Technology and Intelligent Information Systems, and is also an Affiliated Member of Foundation for Research and Technology Hellas, Biomedical Research Institute. He was a Visiting Researcher at the RWTH, Aachen, Germany, and the Massachusetts Institute of Technology, Boston. He has coordinated and participated in more than 250 R&D funded projects (in FP6, FP7, H2020, Horizon Europe and national Projects), being the coordinator (e.g. INSILC, TAXINOMISIS, HOLOBALANCE, CARDIOCARE, DECODE, etc.) and Technical coordinator (e.g. SMARTOOL, KARDIATOOL, TO_AITION, etc.). He is the author or coauthor of more than 400 papers in scientific journals, 600 papers in peer-reviewed conference proceedings, and more than 50 chapters in books. He is also the author/editor of 30 books. His work has received more than 32,800 citations (h-index=82). He is IEEE EMBS Fellow, EAMBES Fellow, Fellow of IAMBE, Fellow of AIAA, member of the IEEE Technical Committee of Biomedical Health Informatics, Editor in Chief of IEEE Journal of Biomedical and Health Informatics, Member of the Editorial Board in IEEE Reviews in Biomedical Engineering, Associate Editor for IEEE Open Journal in Engineering in Biology and Medicine and Computers in Biology and Medicine and member of the European Academy of Sciences and Arts and member of the National Academy of Artificial Intelligence (NAAI). His research interests include multiscale modelling of human tissues and organs, intelligent wearable/implantable devices for automated diagnosis, processing of big medical data, machine learning, sensor informatics, image informatics, and bioinformatics. He is the recipient of many scientific awards including the one by the Academy of Athens. He is the co-founder of PD Neurotechnology Ltd, UK and Intelligence4Rehab.

Dean and Professor
Khoury College of Computer Sciences
Northeastern University

Tuesday | October 28, 2025 | 10:00 – 11:00

Abstract: Many anticipate that AI will play a pivotal role in supporting the goals of older adults to “age in place” and sustain quality of life and independence. AI technologies have the potential to monitor and assess older adults’ behaviors while scaffolding interactive services that support daily activities and needs. Moreover, designing these technologies requires supporting the actions of older adults alongside their caregivers, spouses, adult children, and healthcare providers while being able to draw on a longitudinal understanding of routines, habits, norms, and values. In this talk, I survey the challenges and opportunities inherent in designing holistic AI systems for aging adults and their care networks. Opportunities include providing compensatory support for household safety and reminders, communicating summaries and trends to care network members, and delaying transitions to more substantial institutional care. This work grounds the use-inspired research for the NSF AI Institute for Collaborative Assistance and Responsive Interaction for Networked Groups (AI-CARING).

Biography: Dr. Elizabeth Mynatt is the Dean of Khoury College of Computer Sciences at Northeastern University. Mynatt is an internationally recognized expert in the areas of ubiquitous computing, human-centered computing, health informatics, and assistive technologies. Currently she co-leads the NSF AI-CARING Institute with the goal to create longitudinal, interactive AI technologies to empower older adults and their care networks. Mynatt was selected for induction to the American Academy of Arts and Sciences (AAA&S) in 2024. In 2015 she became a Fellow of the Association of Computing Machinery (ACM) for contributions to human-centered computing and the development of health information technologies. She has been recognized as a member of the ACM SIGCHI Academy, and an AAAS, Sloan and Kavli research fellow. Mynatt has published more than 100 scientific papers and has led grants from the NSF and NIH.

Professor of Human Genetics and Pediatrics, Faculty of Medicine. CEU San Pablo and CJC University, Madrid. Head of Research Group, Institute of Medical and Molecular Genetics

Tuesday | October 28, 2025 | 13:30 – 14:30

Abstract: This presentation will address the topics of disease classification and ontology, the current status of the many ways of classifying and naming diseases, and the problem of nomenclature and granularity of classifications. The distinction between diseases or entities and findings, signs, and symptoms will be emphasized. Finally, a new form of classification will be proposed for better exploitation of big data in health.

Biography: Pablo Lapunzina is Full Professor of Human Genetics and Pediatrics, Faculty of Medicine. CEU San Pablo and CJC University, Madrid. Head of Research Group, Institute of Medical and Molecular Genetics. La Paz University Hospital, Madrid and part of CIBERER, the Network for research on rare diseases in Spain.
Graduate in Medicine from the University of Buenos Aires. He completed his residency as a pediatric resident at the Children’s Hospital in Buenos Aires, later becoming Chief Resident and Instructor. Master in molecular genetics and Master in Hospital Management. He is a specialist in embryo-fetal medicine. He obtained a doctorate from the University of Buenos Aires. He is an expert in Artificial Intelligence from MIT (Massachusetts Institute of Technology in Boston, USA, 2021).
He is the author of more than 350 articles, 35 chapters and 11 books. His work has focused on genomic disorders, overgrowth syndromes, growth failure syndromes, and imprinting alteration syndromes. Together with his research group and in collaboration with several international groups, he has described almost 20 new diseases and syndromes along with the identification of several disease-associated genes. He has been working in classification of diseases and ontologization of diseases, signs and symptoms for the past 10 years.

Distinguished University Professor, Georgia Tech / Georgia State / Emory University

Wednesday | October 29, 2025 | 11:30 – 12:30

Abstract: The human brain is a dynamic, multimodal, and adaptive system whose complexity challenges traditional modeling approaches. Understanding it requires learning frameworks that can integrate information across time, modality, and scale while preserving interpretability and predictive power. This keynote presents a data-guided NeuroAI framework for learning the brain’s dynamic and multimodal complexity directly from data. I will trace the progression from independent component analysis and constrained decomposition to hybrid, dynamic, and multimodal fusion models that link functional, structural, and molecular data into unified representations of brain organization. These approaches reveal how brain networks evolve and interact across time and context, enabling individualized and mechanistic prediction. I will then highlight emerging deep and generative NeuroAI paradigms, including nonlinear ICA, graph-based and diffusion models, that extend independence and disentanglement principles into predictive, interpretable, and scalable domains. Applications such as multimodal brainwide risk scores, dynamic biomarkers, and cross-modal predictive modeling illustrate how these advances transform neuroimaging from descriptive mapping to predictive discovery. Together these developments define a vision for Precision NeuroAI: interpretable, generative, and multimodal frameworks that stay close to the data while learning the evolving complexity that defines the human brain.

Biography: Dr. Vince D. Calhoun is the Founding Director of the Tri-Institutional TReNDS Center and a Distinguished University Professor at Georgia Tech, Georgia State University, and Emory University. A pioneer in the development of group Independent Component Analysis (ICA) for MRI, PET, EEG, and other modalities, he has also led innovations in multimodal fusion methods that integrate diverse neuroimaging data. Dr. Calhoun’s groundbreaking work in brain connectivity and interpretable multimodal analysis has fundamentally transformed how neuroscience leverages artificial intelligence, emphasizing reproducibility, fairness, and trustworthiness across diverse populations. His contributions have been recognized with numerous honors, including the OHBM 2024 Glass Brain Lifetime Achievement Award. A Fellow of IEEE, AAAS, ISMRM, ACNP, IAMBE, OHBM, AAIA, and AIMBE, Dr. Calhoun has authored more than 1,200 publications, which have collectively garnered over 120,000 citations.