MATLAB and Python code for Stein Variational sampling methods
Switch branches/tags
Nothing to show
Clone or download
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
BNN Add files via upload Oct 27, 2018
conditional_diffusion Update script_cond_diff.m Nov 18, 2018
double_banana Add files via upload Oct 27, 2018
fastfins Add files via upload Jun 12, 2018
library Add files via upload Jun 12, 2018
nonlin_regression Add files via upload Oct 27, 2018
python Delete nonIntrusiveSeqSVGD.ipynb Nov 18, 2018
samplers/stein Add files via upload Oct 14, 2018
NeurIPS2018Poster.pdf Add files via upload Nov 23, 2018
README.md Update README.md Nov 23, 2018

README.md

Stein-Variational-samplers

ArXiv: https://arxiv.org/abs/1806.03085 -- NIPS 2018

MATLAB and Python code for Stein Variational Newton method and comparisons with Stein Variational Gradient Descent.

Test cases:

  • Double banana -- two-dimensional, multi-modal banana shaped.
  • Non-linear regression -- two-dimension, skewed shaped.
  • Conditional diffusion -- 100 dimensional, Brownian path reconstruction out of noisy observations of the solution of a Langevin stochastic differential equation.
  • Bayesian Neural Network -- 2951 dimensional on a real data-set with 308 data points.