First International Conference on Probabilistic Numerics

The First International Conference on Probabilistic Numerics brings together researchers from a field that reconceives classical numerical methods — quadrature, differential equation solvers, linear system solvers, optimization — as inference problems under uncertainty. Rather than treating numerical algorithms as producing exact answers subject to deterministic error bounds, probabilistic numerics asks what probability distributions over solutions can be constructed from finite computational budgets, enabling the principled propagation of numerical uncertainty through larger computations.

This inaugural conference marks the maturation of a research program that began as a provocative theoretical proposal and has grown into a active community with practical applications in scientific computing, Bayesian inference, and machine learning. Mathematicians, statisticians, numerical analysts, and machine learning researchers will all find the conference rewarding, as will graduate students interested in the deep connections between computation and statistical inference.