Implementation of kernelized Stein discrepancy and experiements with Stein variational gradient descent.
This contains implementations of the following papers:
[Q. Liu, D. Wang] Stein variational gradient descent: a general purpose bayesian inference algorithm -- arXiv:1608.04471v2
[Q. Liu] Stein variational gradient descent as gradient flow -- arXiv:1704.07520v2
[D. Wang, Q. Liu] Learning to draw samples: with application to amortized MLE for generative adversarial learning -- arXiv: 1611.01722v2
