Gaussian SDE Filter (Algorithm 2) implementation. This script is used to do pathwise comparisons of Algorithm. 2 and Euler-Maruyama scheme. This script is similar to pathwise_convergence_experiment_2.py but uses less RAM by computing the Euler-Maruyama solution on the fly instead of storing Brownian increments in memory and then computing the Euler-Maruyama solution. See Section 3.1

Gaussian Mixture SDE Filter (Algorithm 3) implementation. Similar to the previous scheme. Propagating uncertainty through the posterior variance. Very similar convergence rates to Algorithm 2. See Section 3.2

Marginalised Gaussian SDE Filter (Algorithm 4) implementation. This is a low-order weakly convergent scheme. However, it allows to compute exact (under our model) transition densities. See Section 3.3

Reading .npy output files containing paths, to compute both strong and weak errors.