https://github.com/alshedivat/keras-gp
Keras + Gaussian Processes: Learning scalable deep and recurrent kernels.
https://github.com/alshedivat/keras-gp
gaussian-processes keras machine-learning neural-networks tensorflow theano
Last synced: 2 months ago
JSON representation
Keras + Gaussian Processes: Learning scalable deep and recurrent kernels.
- Host: GitHub
- URL: https://github.com/alshedivat/keras-gp
- Owner: alshedivat
- License: mit
- Created: 2016-09-30T22:07:01.000Z (over 8 years ago)
- Default Branch: master
- Last Pushed: 2024-08-02T16:12:47.000Z (11 months ago)
- Last Synced: 2025-03-30T09:06:19.484Z (3 months ago)
- Topics: gaussian-processes, keras, machine-learning, neural-networks, tensorflow, theano
- Language: Python
- Homepage:
- Size: 90.8 KB
- Stars: 249
- Watchers: 11
- Forks: 54
- Open Issues: 11
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
Gaussian Processes for Keras
----------------------------
[](https://travis-ci.org/alshedivat/keras-gp)
[](https://coveralls.io/github/alshedivat/keras-gp)
[](https://github.com/alshedivat/keras-gp/blob/master/LICENSE)
KGP extends [Keras](https://github.com/fchollet/keras/) with Gaussian Process (GP) layers.
It allows one to build flexible GP models with kernels structured with [deep](http://jmlr.org/proceedings/papers/v51/wilson16.pdf) and [recurrent](https://arxiv.org/abs/1610.08936) networks built with Keras.
The structured part of the model (the neural net) runs on [Theano](http://deeplearning.net/software/theano/) or [Tensorflow](https://www.tensorflow.org/).
The GP layers use a custom backend based on [GPML 4.0](http://www.gaussianprocess.org/gpml/code/matlab/doc/) library, and builds on [KISS-GP](http://www.jmlr.org/proceedings/papers/v37/wilson15.pdf) and [extensions](https://arxiv.org/abs/1511.01870).
The models can be trained in stages or jointly, using full-batch or semi-stochastic optimization approaches (see [our paper](https://arxiv.org/abs/1610.08936)).
For additional resources and tutorials on Deep Kernel Learning and KISS-GP see
[https://people.orie.cornell.edu/andrew/code/](https://people.orie.cornell.edu/andrew/code/)
KGP is compatible with: Python **2.7-3.5**.
In particular, this package implements the method described in our paper:
**Learning Scalable Deep Kernels with Recurrent Structure**
Maruan Al-Shedivat, Andrew Gordon Wilson, Yunus Saatchi, Zhiting Hu, Eric P. Xing
[Journal of Machine Learning Research](https://arxiv.org/abs/1610.08936), 2017.
## Getting started
KGP allows to build models in the same fashion as Keras, using the [functional API](https://keras.io/getting-started/functional-api-guide/).
For example, a simple GP-RNN model can be built and compiled in just a few lines of code:
```python
from keras.layers import Input, SimpleRNN
from keras.optimizers import Adam
from kgp.layers import GP
from kgp.models import Model
from kgp.losses import gen_gp_loss
input_shape = (10, 2) # 10 time steps, 2 dimensions
batch_size = 32
nb_train_samples = 512
gp_hypers = {'lik': -2.0, 'cov': [[-0.7], [0.0]]}
# Build the model
inputs = Input(shape=input_shape)
rnn = SimpleRNN(32)(inputs)
gp = GP(gp_hypers,
batch_size=batch_size,
nb_train_samples=nb_train_samples)
outputs = [gp(rnn)]
model = Model(inputs=inputs, outputs=outputs)
# Compile the model
loss = [gen_gp_loss(gp) for gp in model.output_layers]
model.compile(optimizer=Adam(1e-2), loss=loss)
```
Note that KGP models support arbitrary off-the-shelf optimizers from Keras.
**Further resources:**
- A [quick tutorial](https://github.com/alshedivat/keras-gp/tree/master/tutorials) that walks you through the key components of the library.
- A few more [examples](https://github.com/alshedivat/kgp/tree/master/examples).
## Installation
KGP depends on [Keras](https://github.com/fchollet/keras/) and requires either [Theano](http://deeplearning.net/software/theano/) or [TensorFlow](http://tensorflow.org/) being installed.
The GPML backend requires either MATLAB or Octave and a corresponding Python interface package: [Oct2Py](https://blink1073.github.io/oct2py/) for Octave or the [MATLAB engine for Python](https://www.mathworks.com/help/matlab/matlab-engine-for-python.html).
Generally, MATLAB backend seems to provide faster runtime.
However, if you compile the latest version of Octave with JIT and OpenBLAS support, the overhead gets reduced to minimum.
If you are using Octave, you will need the `statistics` package.
You can install the package using Octave-Forge:
```bash
$ octave --eval "pkg install -forge -verbose io"
$ octave --eval "pkg install -forge -verbose statistics"
```
The requirements can be installed via [pip](https://pypi.python.org/pypi/pip) as follows (use `sudo` if necessary):
```bash
$ pip install -r requirements.txt
```
To install the package, clone the repository and run `setup.py` as follows:
```bash
$ git clone --recursive https://github.com/alshedivat/kgp
$ cd kgp
$ python setup.py develop [--user]
```
The `--user` flag (optional) will install the package for a given user only.
**Note:** Recursive clone is required to get GPML library as a submodule.
If you already have a copy of GPML, you can set `GPML_PATH` environment variable to point to your GPML folder instead.
## Contribution
Contributions and especially bug reports are more than welcome.
## Citation
```bibtex
@article{alshedivat2017srk,
title={Learning scalable deep kernels with recurrent structure},
author={Al-Shedivat, Maruan and Wilson, Andrew Gordon and Saatchi, Yunus and Hu, Zhiting and Xing, Eric P},
journal={Journal of Machine Learning Research},
volume={18},
number={1},
year={2017},
}
```
## License
For questions about the code and licensing details, please contact [Maruan Al-Shedivat](https://www.cs.cmu.edu/~mshediva/) and [Andrew Gordon Wilson](https://people.orie.cornell.edu/andrew).