Learning Representations of High-Dimensional Stochastic Systems

Explore how to learn effective representations of complex, high-dimensional stochastic systems in this conference talk by Gautam Reddy from Princeton University. Discover mathematical frameworks and computational approaches for extracting low-dimensional structure from high-dimensional biological data, including applications in neuroscience, development, ecology, and evolution. Learn about the theoretical foundations underlying dimensionality reduction techniques and their role in understanding emergent simplicity in biological systems. Examine how machine learning methods can identify meaningful patterns in complex stochastic processes while maintaining biological interpretability. Understand the challenges and opportunities in developing unifying theoretical frameworks that bridge different biological domains through common mathematical principles. Gain insights into the balance between model complexity and explanatory power when dealing with high-dimensional biological phenomena, and explore how concepts like evolvability and functional robustness enable effective low-dimensional descriptions of complex systems.