Tutorials

Here, the usage of PHYSBO is introduced through tutorials.

Contents:

• Basic usage of PHYSBO
  • Introduction
  • Defining the search space
  • Definition of the simulator
  • Performing optimization
  • Checking the results
  • Serializing the results
  • Prediction
  • Acquisition function
  • Parallelization
• Gaussian process
  • Preparation of search candidate data
  • Defining training data
  • Learning and Prediction of Gaussian Processes
  • Prediction by trained models
  • Bayesian linear regression with random feature map
  • Notes
• ARD (Automatic Relevance Determination) Tutorial
  • Import packages
  • Problem setup for this tutorial
  • Common parameters
  • Case: ard=True
  • Case: ard=False
  • Comparing best values
  • Comparing length scales
  • Comparing permutation importance
• Running PHYSBO interactively
  • Preparation of search candidate data
  • Definition of the simulator
  • Executing optimization
  • Suspend and restart
• Restart calculations by reading existing calculation results
  • Prepare the search candidate data
  • Preparing the precomputed data
  • Register action ID (parameter) and evaluation value to PHYSBO.
  • Get the next parameter to be executed from PHYSBO
• Search multiple candidates at once
  • Prepare the search candidate data
  • Definition of simulator
  • Performing optimizations
  • Checking results
• Multi-objective optimization
  • Test functions
  • Preparation of search candidate data
  • Definition of simulator
  • Plotting the functions
  • Performing optimizations.
  • Setting policy
  • Bayesian optimization
• Fast Computation for Multi-objective Optimization
  • Unification of Multi-objective Optimization
  • Tutorial
• Optimization in Continuous Space
  • Basic Usage
  • Acquisition Function Optimization Algorithm