Prof. Daniel Huson (H-index: 86)

University of Tübingen

Daniel Huson is Professor of Algorithms in Bioinformatics at the University of Tübingen, Germany. He studied mathematics and physics at Bielefeld University, where he earned his PhD and Habilitation, and before joining Tübingen in 2002 he held research positions at Princeton University and at Celera Genomics in the United States. He has also held visiting professorships at the National University of Singapore and Nanyang Technological University.

His research designs algorithms and software for phylogenetics, genomics, and microbiome analysis. He is the author of several widely used programs, including SplitsTree for phylogenetic trees and networks, and MEGAN and MALT for the taxonomic and functional analysis of microbiome and ancient-DNA sequencing data; he also co-developed the fast sequence aligner DIAMOND. These tools are applied across fields ranging from environmental microbiology to the study of Neanderthal diet and disease from ancient dental calculus.

Huson's work has been cited more than 70,000 times. His distinctions include the PLOS Computational Biology Research Prize 2016 and first place in the 2013 DTRA Algorithms Challenge. His current research centres on phylogenetic networks and the continued development of the SplitsTree, MEGAN, and PhyloSketch software.

INVITED SPEAKERS

Assoc. Prof. Hussam Alsharif

Umm Al-Qura University

Dr. Hussam Alsharif received the B.Sc. degree from Pittsburg State University, USA, in 2010, the M.Sc. degree from Western Illinois University, USA, in 2012, and the Ph.D. degree from North Carolina Agricultural and Technical State University, USA, in 2020. He served as an Assistant Professor in the Department of Computer Science at the University College of Al Jamoum, Umm Al-Qura University, Makkah, Saudi Arabia, from 2020 to 2025, and has been an Associate Professor since 2026. His research interests include artificial intelligence, computer vision, medical image analysis, machine learning, deep learning, data science, computational biology, bioinformatics, big data analytics, and intelligent healthcare systems. He has authored numerous peer-reviewed publications in these fields and serves as a reviewer for several international journals.

Speech Title:"AI-Powered Prognostic Assessment of Impacted Maxillary Canines Using Hybrid Deep Learning"

Abstract: Artificial Intelligence (AI) is improving efficiency and consistency in dental and medical imaging, thereby advancing clinical decision-making. In orthodontics, predicting the prognosis of impacted maxillary canines remains challenging, as treatment outcomes rely heavily on clinicians’ interpretation of panoramic radiographs, often resulting in inter-observer variability. In this talk, we present a hybrid deep learning architecture that combines the strengths of the Swin Transformer and ResNet18 through an attention-based feature fusion strategy for prognostic assessment of impacted maxillary canines from panoramic radiographs. The proposed system classifies cases into three clinically meaningful prognostic categories: good, average, and poor. We will also discuss the role of hybrid deep learning systems in dental imaging, the challenges of AI implementation in routine clinical practice, and future directions for building explainable and clinically trustworthy AI systems. The proposed framework has the potential to support clinical decision-making by improving diagnostic consistency, reducing inter-observer variability, and assisting in more effective treatment planning.

Asst. Prof. Yen-Jung Chiu

Chang Gung University, Taiwan

Dr. Yen-Jung Chiu is an Assistant Professor in the Department of Biomedical Engineering at Chang Gung University, Taiwan, with a joint appointment at the Division of Cardiology, Chang Gung Memorial Hospital. His research focuses on federated learning for computational pathology, whole slide image analysis with foundation models, and cancer immunogenomics. He developed LIMPACAT, a multi-omics attention transformer for immune prediction in liver cancer using whole-slide imaging, and leads FedFM-WSI, tha is a large-scale federated learning benchmark covering 107 real-world tissue source site partitions, 7 pathology foundation models, and 9 cancer types across 5,679 TCGA whole slide images.

Speech Title:"Benchmarking Federated Learning with Pathology Foundation Models: Lessons from 107 Real-World Hospital Partitions"

Federated learning (FL) has emerged as a promising paradigm for privacy-preserving multi-institutional analysis of whole slide images (WSIs), enabling hospitals to collaboratively train models without sharing sensitive patient data. Yet rigorous benchmarks under real-world data heterogeneity remain scarce, leaving open questions about which FL algorithms and foundation models actually generalize once deployed across genuinely non-IID clinical sites. This talk presents FedFM-WSI, a comprehensive benchmark evaluating four widely used FL algorithms — FedAvg, FedProx, SCAFFOLD, and FedBN — across seven pathology foundation models and nine cancer types, built on 5,679 TCGA whole slide images partitioned by 107 tissue source sites as a natural, clinically grounded proxy for hospital-level federation. Unlike prior work relying on synthetic or artificially shuffled splits, this partitioning preserves institution-specific staining, scanning, and demographic shifts that federated systems must actually contend with in practice. I will discuss which algorithm–foundation-model combinations remain robust under class-partitioned non-IID conditions, and which degrade sharply despite performing well under standard benchmarks. I will reveal several critical failure modes, including client drift, representation collapse, and unstable convergence, invisible in conventional IID evaluations, and conclude with practical recommendations for deploying federated computational pathology systems in real-world clinical settings.

Assoc. Prof. Yihang Zhou

Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, China

Yihang Zhou is an Associate Professor and Doctoral Supervisor at the Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences. His research focuses on magnetic resonance imaging, artificial intelligence for medical imaging, low-field MRI, and neural decoding for brain-to-brain communication. He previously served as Deputy Head of the Department of Medical Physics and Research at Hong Kong Sanatorium & Hospital, where he participated in the clinical implementation and research development of MR-guided radiotherapy systems. Over the past five years, he has led major research projects funded by the National Natural Science Foundation of China, the Chinese Academy of Sciences, the National Key Research and Development Program of China, and the Shenzhen Science and Technology Major Program, with total funding exceeding RMB 20 million. He has published more than 50 papers in leading international journals and filed over 20 Chinese patent applications. He holds one US patent and six granted Chinese patents. Several of his research outcomes have been successfully commercialized and translated into clinical practice. In recent years, his research group has conducted extensive research on AI-enhanced low-field MRI, rapid four-dimensional MRI, medical image reconstruction and segmentation, and multimodal neuroimaging. These studies aim to improve the accessibility, image quality, and clinical utility of advanced medical imaging technologies.

Speech Title:"The MR Matrix: Building an Intelligent Diagnostic Foundation Model with Distributed Low-Field MRI"

The presentation would introduce our recent progress in physics-informed image reconstruction, low-field MRI system development, multimodal EEG–MRI integration, and the translation of artificial intelligence methods into practical biomedical imaging applications. Low-field magnetic resonance imaging (MRI) offers a cost-effective and accessible solution for brain imaging in community and resource-limited settings, but its clinical use is constrained by low image quality, device variability, and limited annotated data. This talk presents The MR Matrix, a distributed low-field MRI framework that integrates standardized acquisition, physics-informed image enhancement, and foundation-model-based analysis. Using self-supervised learning, cross-device feature alignment, and knowledge transfer from high-field MRI, the model learns transferable representations for brain structure segmentation, abnormality detection, disease classification, and risk assessment. By connecting distributed scanners, imaging data, and diagnostic intelligence, The MR Matrix aims to transform low-field MRI from isolated imaging devices into a scalable platform for brain disease screening and intelligent diagnosis.

PREVIOUS SPEAKERS

Prof. Yudong Zhang University of Leicester, UK	Prof. Dong Sun City University of Hong Kong, Hong Kong	Prof. Nicola Mulder University of Cape Town, South Africa	Prof. Stephen Kwok-Wing Tsui The Chinese University of Hong Kong, Hong Kong
Prof. Dong Ming Tianjin University, China	Prof. Tetsuo Shibuya The University of Tokyo, Japan	Assoc. Prof. Lin Meng Tianjin University, China	Assoc. Prof. Jie Luo Shanghai Jiao Tong University, China
Prof. Mitsuhiro Ogawa Teikyo University, Japan	Prof. Hao Jiang Renmin University of China, China	Assoc. Prof. Xingwei An Tianjin University, China	Assist. Prof. Faez Iqbal Khan Xi'an Jiaotong-Liverpool University, China
Prof. Chenjie Xu (H-index: 60) City University of Hong Kong, China	Prof. Yasukazu Nakamura (H-index: 65) National Institute of Genetics, Japan	Prof. Keiji Nakajima Nara Institute of Science and Technology, Japan	Assoc. Prof. Marwan El Rich Khalifa University, UAE