Recent Advances in Leveraging AI for Scalable and Extensible Biomedical Knowledge Management
Related Experiment Video
Updated: Nov 5, 2026
Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems
Published on: June 13, 2025
Gongbo Zhang1, Betina Idnay1, Chunhua Weng1
1Columbia University, New York, NY, USA.
Summary:
OBJECTIVES:: : To survey the recent literature on knowledge representation and management (KRM) and their biomedical applications.
Summary:
METHODS: : A bibliometric analysis of KRM methodological innovations and health applications published from January 2024 to February 2025 was carried out.
Summary:
RESULTS: : Four major methodological development trends emerge from our analysis, which are multimodal data integration, ontology-driven knowledge representation, continuous knowledge representation learning, and contrastive learning for knowledge extraction. These methods are applied across a range of biomedical domains, as represented by accelerated drug development, clinical decision support, and personalized medicine.
Summary:
DISCUSSION: : There is a growing emphasis on data-driven and learning-based approaches to biomedical knowledge representation. However, challenges remain in developing standardized evaluation benchmarks and enabling continuous updates to knowledge resources. Future innovations may arise from integrating knowledge across overlapping domains-such as precision medicine and drug repurposing-to enhance the translational impact of KRM methods.
