This project is aimed at actively learning the distribution of multiple correlated target features (e.g. concentration of minerals in a region). The distribution of target features or variables are modelled by a Multi-Output Spectral Mixture Kernels Gaussian Process proposed in https://papers.nips.cc/paper/7245-spectral-mixture-kernels-for-multi-output-gaussian-processes (https://github.com/gparracl/MOSM). After estimating the model hyperparameters by fitting the model on a training set, the agent actively determines the type and location of samples to be collected in order to minimize the uncertainty of model's prediction. The information gain criterion is entropy.
After installing the listed dependencies, simply clone this package to run scripts.
See ex.py script to setup the Jura dataset, train the GP model and perform active learning. You can setup certain arguments from command line (see arguments.py file). Execute the following to start the training:
python ex.py
A significant portion of the code is taken from Gabriel Parra's implementation available at https://github.com/gparracl/MOSM.
For any queries, feel free to raise an issue or contact me at [email protected].
This project is licensed under the MIT License - see the LICENSE.md file for details.