Dora simultaneously builds a non-parametric model of an observable process, and directs new measurements to learn about the process. Inherently, it takes actions to improve the model (exploration), and to achieve desired goals given its knowledge (exploration).

The models utilised are non-parametric models such as bilinear interpolation of a Delaunay triangulation or Gaussian Process regression. We are building this tool to be able to plug in other models, so it will shortly be possible to link in our revrand package for fast Bayesian linear regression that behaves similar to GP regression so that kernel-regression like modeling can be applied to large sampled datasets.

The key features of Dora are:

• a Python library to design experimental sampling to simultaneously explore and exploit an underlying process
• a RESTful web service that can be called from any other launguages providing the core functionality of the Python library
• The ability to use a variety of process models including Gaussian Process regression [1], Delaunay triangulation, and models from the Revrand library [2].
• a suite of strategies to conduct information, risk or value based sampling to learn these models efficiently or target phenomenon of interest.

To install, simply run setup.py:

$ python setup.py install

or install with pip:

$ pip install git+https://github.com/nicta/dora.git@release

Refer to docs/installation.rst for advanced installation instructions.

Have a look at some of the demos, e.g.:

$ python demo/demo_gp_python_api.py

The demos include examples of different underlying models (GP and Delaunay), and examples of both Python 3 code calling the api directly, and the general case of calling a server through a HTTP REST interface.

Here is a very brief example of how to use active sampling with a Gaussian Process model in Python. We are assuming we can collect some limited training data with noisy targets y_train, inputs X_train for the purpose of selecting initial model hyperparameters, and we now want to use targeted sampling to efficiently continue the exploration.

Home Page

http://github.com/nicta/dora

Documentation

http://nicta.github.io/dora

Issue tracking

https://github.com/nicta/dora/issues

For bugs, questions and discussions, please use Github Issues.

Copyright 2015 National ICT Australia.

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at:

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.