Skip to content

ssydasheng/Neural-Kernel-Network

Repository files navigation

Neural Kernel Network

This code is jointly contributed by Shengyang Sun, Guodong Zhang, Chaoqi Wang and Wenyuan Zeng

Introduction

Code for "Differentiable Compositional Kernel Learning for Gaussian Processes" (https://arxiv.org/abs/1806.04326)

Dependencies

This project runs with Python 3.6. Before running the code, you have to install

Experiments

Below we shows some examples to run the experiments. We also provide experiment figures and logging files in results folder, as a reference.

Time Series

python exp/time-series.py --name airline --kern nkn

Regression

python exp/regression.py --data energy --split uci_woval --kern nkn
python exp/regression.py --data energy --split uci_woval_pca --kern nkn

Bayesian Optimization

python exp/bayes-opt.py --name sty --kern nkn --run 0

Texture Extrapolation

python exp/texture.py --data pave --kern nkn

Citation

To cite this work, please use

@article{sun2018differentiable,
  title={Differentiable Compositional Kernel Learning for Gaussian Processes},
  author={Sun, Shengyang and Zhang, Guodong and Wang, Chaoqi and Zeng, Wenyuan and Li, Jiaman and Grosse, Roger},
  journal={arXiv preprint arXiv:1806.04326},
  year={2018}
}

Releases

No releases published

Packages

No packages published

Languages