Integrative Deep Learning for Heterogeneous Biomedical Datasets

Learn about integrative deep learning approaches for analyzing heterogeneous biomedical datasets in this 24-minute conference talk from the Computational Genomics Summer Institute 2025. Explore advanced methodologies for combining diverse biological data types through deep learning frameworks, with particular focus on multimodal attention mechanisms and graph-based reasoning approaches. Discover how these techniques can be applied to critical biomedical challenges including drug repurposing, drug interaction prediction, and Alzheimer's disease diagnosis. Examine cutting-edge research on K-Paths reasoning over graph structures for pharmaceutical applications, scalable multimodal integration strategies using one-versus-others attention mechanisms, and attention-based deep learning models for neurological disease classification. Gain insights into the computational challenges and solutions for processing complex, multi-dimensional biomedical data while understanding the practical applications of these methods in genomics and precision medicine research.