SOURCE-LINKED INTELLIGENCE

Propensity Straight-Through Gradients for Discrete Stochastic Systems

arXiv · AI, language, vision and robotics · article · Aug 26, 2026 · UTC

Continuous-time Markov chains (CTMCs) provide the backbone for modeling discrete stochastic dynamics across applied, physical, and biological sciences. Their integration with modern gradient-based machine learning, however, is limited by the hard categorical event selection intrinsic to Gillespie-type simulation algorithms. We exploit the affine state update to obtain the exact one-step conditional-mean sensitivity by differentiating normalized reaction propensities. We pair this backward rule with exact forward trajectories to define the propensity straight-through (PST) estimator. At the tra

Read original source ↗ Open in workspace

recordType

paper

region

Global