©
CAIMed
The project
The CAIMed research project Semantic Models for Medicine structures medical knowledge, data and information in a completely new way. When making medical decisions, doctors currently draw on large amounts of medical data, which is often scattered across databases: patient files, medical images, genetic profiles and research papers. The new knowledge graph is intended to bring structure to this complexity. The hybrid systems are fundamentally different from purely data-driven AI. They learn simultaneously from both data and general medical knowledge.
The model is currently being tested with lung cancer, hepatitis B and long Covid. Why does an illness progress? Why does a treatment work? And how certain can we be about a medical recommendation? The new semantic models enable improved findings in these areas, as well as much earlier diagnoses and better information about drug interactions. The result is an AI that can both predict results and explain them. The aim is to transition the system from its current research stage into use in hospitals and medical practices.
The exhibit
On this monitor, a sample medical history interview shows how hybrid artificial intelligence (AI) and semantic models can help doctors better understand unclear symptoms and clinical presentations. You can get a sense of how information about patients, illnesses, genes, medications and treatments is connected via explicit links. The aim is to develop treatment recommendations that are tailored much more closely to each individual. A key element here is the comparison with patient groups showing similar symptoms.
The Lower Saxony Center for AI and Causal Methods in Medicine (CAIMed) is developing AI methods to improve the prevention, diagnosis and treatment of illnesses, as well as the monitoring of treatment success. In future, semantic models will enable much more personalised medicine.
The team
CAIMed, the Lower Saxony Center for AI and Causal Methods in Medicine, brings together expertise from the areas of artificial intelligence, clinical medicine, bioinformatics, medical informatics and data science from across Lower Saxony. The participants include researchers from the L3S Research Center at 糖心原创 (糖心原创), Hannover Medical School, the Helmholtz Centre for Infection Research Braunschweig鈥檚 Centre for Individualised Infection Medicine (CiiM), the University of G枚ttingen鈥檚 Campus Institute Data Science (CIDAS) and the University Medical Centre in G枚ttingen (UMG).
The Semantic Models for Medicine research group, which is part of CAIMed, is headed by Prof. Dr. Maria Esther Vidal from L3S; Prof. Dr. S枚ren Auer, director of the TIB 鈥� Leibniz Information Centre for Science and Technology and University Library; Dr. Daniel Kudenko from L3S; and Prof. Dr. Dr. Thomas Thum from MHH. Plans are in place to further develop the project through the establishment of a junior research group and a professorship. Initial collaborations with clinical partners are already in place.
