SDE Conditioning

Condition Linear SDEs on Gaussian observations to sample from posterior distributions over trajectories.

When to Use

• Interpolating sparse time series observations with uncertainty
• Creating Brownian bridges or other conditioned stochastic processes
• Performing Bayesian inference on latent states given noisy measurements
• State estimation with process and observation noise

Key Concepts

Available SDEs

linsdex provides several built-in Linear SDEs: