Generative Models and Their Applications in Computational Biology

Explore the intersection of generative modeling and computational biology in this 35-minute conference talk from the Computational Genomics Summer Institute. Delve into the fundamental principles of generative models, including variational autoencoders, denoising diffusion probabilistic models, and flow matching techniques. Examine how these advanced machine learning approaches are being applied to solve complex problems in bioinformatics and computational biology. Learn about the theoretical foundations established by seminal papers in the field, from Kingma & Welling's auto-encoding variational Bayes framework to recent advances in diffusion models and flow matching for generative modeling. Discover practical applications of these methods in biological data analysis, protein structure prediction, genomic sequence generation, and other computational biology challenges. Gain insights into how diffusion models are revolutionizing bioinformatics workflows and their potential for advancing our understanding of biological systems through generative approaches.