NUBO: A Transparent Python Package for Bayesian Optimisation#
NUBO, short for Newcastle University Bayesian optimisation, is a Bayesian optimisation framework for the optimisation of expensive-to-evaluate black-box functions, such as physical experiments and computer simulations. It is developed and maintained by the Fluid Dynamics Lab at Newcastle University. NUBO focuses primarily on transparency and user experience to make Bayesian optimisation easily accessible to researchers from all disciplines. Transparency is ensured by clean and comprehensible code, precise references, and thorough documentation. User experience is ensured by a modular and flexible design, easy-to-write syntax, and careful selection of Bayesian optimisation algorithms. NUBO allows you to tailor Bayesian optimisation to your specific problem by writing the optimisation loop yourself using the provided building blocks or using an off-the-shelf algorithm for common problems. Only algorithms and methods that are sufficiently tested and validated to perform well are included in NUBO. This ensures that the package remains compact and does not overwhelm the user with an unnecessarily large number of options. The package is written in Python but does not require expert knowledge of Python to optimise your simulations and experiments. NUBO is distributed as open-source software under the BSD 3-Clause licence.
Contact
Thanks for considering NUBO. If you have any questions, comments, or issues feel free to email us at m.diessner2@newcastle.ac.uk. Any feedback is highly appreciated and will help make NUBO better in the future.
On this page, you can find an overview of the three main documentation sections consisting of (i) an introduction to Bayesian optimisation with NUBO, (ii) off-the-shelf algorithms for an easy and quick start with NUBO, (iii) a selection of examples that can be used as boilerplate code, and (iv) detailed references for all of NUBO’s functions and objects.
NUBO#
The NUBO section contains general information about the package, gives a concise introduction to Bayesian optimisation, and explains its components, such as the surrogate model and acquisition functions. It also provides a quick start guide to NUBO allowing you to start optimising your simulations and experiments in minutes. This is the place to start your Bayesian optimisation journey with NUBO.
Off-the-shelf algorithms#
Off-the-shelf algorithms provide the quickest and easiest way of optimising expensive black-box functions with NUBO. Basic optimiation covers a wide range of common optimisation problems such as single-point or multi-point optimisation with inputs constraints or asynchronous optimisation. If your problem is more complex, the next sections explains how to write custom optimisation loops yourself.
Off-the-shelf algorithms:
Custom loop examples#
The Examples section provides guides to some problems that NUBO has been designed to optimise and shows how to implement custom surrogate models. This boilerplate code is a good starting point when tailoring a Bayesian optimisation algorithm to your specific problem.
Custom loop examples:
- Custom Gaussian process
- Sequential single-point Bayesian optimisation
- Parallel multi-point joint Bayesian optimisation
- Parallel multi-point sequential Bayesian optimisation
- Parallel multi-point Bayesian optimisation with fixed base samples
- Asynchronous Bayesian optimisation
- Constrained single-point Bayesian optimisation
- Bayesian optimisation with continuous and discrete parameters
- Bayesian optimisation with known observational noise
Package reference#
The Package reference section gives detailed documentation of all of NUBO’s functionality. This is where you want to go first when you are not sure how a specific object or function should be used.