If you want to contribute to this folder, please open a new pull request.
Examples can be either notebooks that tell a story about a problem/question using Emukit, e.g the analysis of the properties of a simulator, or they can contain an implementation of some specific method. Examples can live in a .py file, and ideally have tests and come with an illustrative notebook.
Example applications or scripts should be submitted in this
emukit/examples folder. Each example must live in a separated
folder that can contain some extra files and dependencies. Please make sure that you update this
README.md file with the information
about you example when submitting a PR.
As part of making sure all our examples are running correctly with the latest version of Emukit, yor can add your own tests
emukit/tests/examples/test_example.py (if you forget, don't worry, we'll remind you during the review).
Tutorials are Jupyter notebooks that illustrate different features of the library. They are stand alone notebooks that don't require any extra file and fully sit on Emukit components (apart from the creation of the model).
If you have a tutorial idea, please download the Jupyter notebook tutorial template.
Notebook tutorials should be submitted in the
Do not forget to update the
notebooks/index.ipynb for your tutorial to show up on the website.
- Cost sensitive Bayesian optimization - Wrapper for using Bayesian optimization when there is a cost involved in the evaluation of the objective.
- Gaussian process Bayesian Optimization - Wrapper for using Bayesian optimization with Gaussian processes.
- Vanilla Bayesian Quadrature - Wrapper for vanilla Bayesian quadrature that uses a Gaussian processes with an RBF kernel.
- Models - Implementation of a variety of models that can be used in combination with other Emukit features.
- Mountain car - Optimization of the control policy of the mountain car simulator. Optimization is applied using an emulator of the reward and of the dynamics of the simulator.
Visit the index of tutorials.