Fast Probabilistic Inference for ODEs with ProbNumDiffEq.jl

Explore probabilistic numerical methods for solving ordinary differential equations (ODEs) in this JuliaCon 2024 talk by Nathanael Bosch. Dive into the ProbNumDiffEq.jl package, which implements efficient Bayesian filtering and smoothing techniques for ODE solving within the DifferentialEquations.jl ecosystem. Learn about the foundations of probabilistic numerics, examine example usages for various problem types, and discover implementation details crucial for stability and speed. Gain insights into integrating with other Julia packages and discuss the future of probabilistic numerics in Julia. Discover how ProbNumDiffEq.jl offers familiar ODE solver features alongside unique probabilistic capabilities, including priors, calibration approaches, and likelihood models for parameter estimation.