Advanced Examples

Advanced Examples

These notebooks showcase more specialized and advanced use cases in BoFire. These examples are not necessarily better strategies, but represent more complex uses of components within the library.

Available Tutorials

Custom SOBO Strategy

Create custom single-objective Bayesian optimization strategies.

Desirability Functions for Multi-Objective Optimization

Working with desirability functions for multi-criteria optimization.

Genetic Algorithm

Using genetic algorithms for optimization in BoFire.

Merging Objectives

Techniques for combining multiple objectives in optimization.

Multi-fidelity Bayesian Optimization

Leveraging multiple fidelity levels for efficient optimization.

Multi-Objective, Multi-Fidelity BO

Optimize multiple objectives across fidelity levels.

Input Features as Output Objectives

Defining optimization objectives directly on input parameters.

Random Forest in BoFire

Using Random Forest as a surrogate model instead of Gaussian Processes.

Transfer Learning BO

Applying transfer learning techniques to Bayesian optimization.

Predict Motor Octane Number of Hydrocarbon Mixtures

Predict the motor octane number of hydrocarbon mixtures with a surrogate model.

Conditional Features

Define input features that are conditionally active.

LLM-driven Molecular Optimization

Propose candidates with an LLMStrategy that prompts a large language model with the optimization problem and prior experiments.

Preference Learning with the Pairwise GP

Learn a latent utility function from pairwise comparison data using Pairwise Gaussian Processes.