streaming_sparse_gp

KO | EN

gitlite — search

#python #java #python3 #arduino #golang #machine-learning #rust #html #flask #javascript #seismology #nodejs

streaming_sparse_gp

★ 70 Open GitHub ↗

Streaming sparse Gaussian process approximations

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

## Streaming sparse Gaussian process approximations

This repository contains an implementation of several online/streaming sparse GP approximations for regression and classification ([Bui, Nguyen and Turner, NIPS 2017](https://proceedings.neurips.cc/paper/2017/hash/f31b20466ae89669f9741e047487eb37-Abstract.html)). In particular, [osvgp.py](code/osvgpc.py) implements the uncollapsed variational free-energy for regression and classification, and [osgpr.py](code/osgpr.py) implements the collapsed variational free-energy and Power-EP energy for the regression case.

We also provide an implementation of the collapsed batch Power-EP sparse approximation of Bui, Yan and Turner (2017).

The code was tested using GPflow 2.5 and 2.6 and TensorFlow 2.5.

## Usage

We provide several test scripts ([regression](code/run_reg_toy.py) and [classification](code/run_cla_toy.py)) to demonstrate the usage. Running these examples should the results similar to the following figures:

![regression](https://raw.githubusercontent.com/thangbui/streaming_sparse_gp/master/tmp/reg_VFE_M_10_iid_False.png)

![classification](https://raw.githubusercontent.com/thangbui/streaming_sparse_gp/master/tmp/cla_VFE_M_30_iid_False.png)

## Contributors

[Thang D. Bui](https://thangbui.github.io/)\
[Cuong V. Nguyen](https://nvcuong.github.io/)\
[Richard E. Turner](http://learning.eng.cam.ac.uk/Public/Turner/WebHome)\
[ST John](https://github.com/st--/)

## References:

```
@inproceedings{BuiNguTur17,
  title =  {Streaming sparse {G}aussian process approximations},
  author =   {Bui, Thang D. and Nguyen, Cuong V. and Turner, Richard E.},
  booktitle = {Advances in Neural Information Processing Systems 30},
  year =   {2017}
}

@article{BuiYanTur16,
  title={A Unifying Framework for Sparse {G}aussian Process Approximation using {P}ower {E}xpectation {P}ropagation},
  author={Thang D. Bui and Josiah Yan and Richard E. Turner},
  journal={arXiv preprint arXiv:1605.07066},
  year={2016}
}
```

streaming_sparse_gp

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README