CSE Community Seminar | September 18, 2026

Dynestyx: A Probabilistic Programming Library for Dynamical Systems

Daniel Waxman, Research Affiliate at LIDS, MIT

Abstract: State-space models (SSMs) are the standard formalism for Bayesian treatment of dynamical systems, with natural applications in statistics, signal processing, and machine learning. Despite their importance in both theory and application, dynamical systems have proven difficult to incorporate in modern probabilistic programming languages (PPLs), making state-of-the-art methods less accessible to practitioners and introducing friction in following the “Bayesian workflow.” We introduce dynestyx, a probabilistic programming library with first-class support for SSMs, including state-of-the-art methods in the estimation of both states and parameters. Through a single, unified interface, users may specify arbitrary priors for discrete-time or continuous-time dynamical systems, perform inference over mixed-effect data, and make state and parameter estimates with principled uncertainty quantification. We discuss problems, applications, and opportunities that dynestyx introduces.

Parallel Computing for Coupled-Process Modeling of Subsurface Energy Systems

Tianjia Huang, Postdoctoral Associate, Department of Civil and Environmental Engineering, MIT

Abstract: Efficient utilization of subsurface energy resources while minimizing environmental impacts represents one of the most pressing scientific challenges. Numerical modeling has become an indispensable tool for addressing this challenge. However, geoscientific and geoengineering applications are characterized by coupled thermal, hydraulic, mechanical, and chemical processes, as well as geological heterogeneity and uncertainty. Capturing these complexities requires the continued advancement of computational capabilities. Integrating reservoir simulation with high-performance computing (HPC) is therefore essential for developing robust and scalable simulation tools for complex subsurface systems.

In this talk, I will concentrate on the workflow used to parallelize the TOUGH family codes (https://tough.lbl.gov/), which model the coupled transport of water, vapor, no-condensable gases, and heat in porous and fractured media. I will first review fundamental concepts in reservoir simulation and then discuss key challenges in developing high-performance subsurface simulators. I will demonstrate how parallel computing can accelerate scientific discovery through an application involving mineral precipitation and dissolution in enhanced geothermal systems. The talk will address three questions: What are the principal challenges in developing high-performance subsurface simulators? How can parallel computing accelerate subsurface energy research? And what capabilities should define the next generation of subsurface simulation tools in my mind?