bolib 0.22.0

Python library for Bayesian Optimization.

BOlib

A python library for Bayesian Optimization.

Setup BOlib

• Create and activate venv

  python3 -m venv .env
  
  source .env/bin/activate
  

• Upgrade pip

  python -m pip install --upgrade pip
  

• Install BOlib package

  python -m pip install bolib
  

• Matplotlib requires to install a backend to work interactively

  (See https://matplotlib.org/faq/virtualenv_faq.html). The easiest solution is to install the Tk framework, which can be found as python-tk (or python3-tk) on certain Linux distributions.

Use BOlib

• Import BOlib to use it in your python script.

  import bolib
  

• Some well-known objetive functions have been included.

  of = bolib.ofs.Branin()
  
  of.evaluate([1.0, 1.0])  # 27.702905548512433
  

• To use Bayesian Optimization we need a probabilistic model. In this example we will use Gaussian Processes.

  import gplib
  
  model = gplib.GP(
      mean_function=gplib.mea.Fixed(),
      covariance_function=gplib.ker.SquaredExponential(ls=([1.] * of.d))
  )
  
  metric = gplib.me.LML()
  
  fitting_method = gplib.fit.MultiStart(
      obj_fun=metric.fold_measure,
      max_fun_call=300,
      nested_fit_method=gplib.fit.LocalSearch(
          obj_fun=metric.fold_measure,
          max_fun_call=75,
          method='Powell'
      )
  )
  
  validation = gplib.vm.Full()
  

• Bayesian Optimization also needs an acquisition function.

  af = bolib.afs.ExpectedImprovement()
  

• Finally, we can initialize our optimization model and start the optimization process.

  bo = bolib.methods.BayesianOptimization(
      model, fitting_method, validation, af
  )
  
  bo.set_seed(seed=1)
  
  x0 = bo.random_sample(of.get_bounds(), batch_size=10)
  
  bo.minimize(
      of.evaluate, x0,
      bounds=of.get_bounds(),
      tol=1e-5,
      maxiter=of.get_max_eval(),
      disp=True
  )
  

• BOlib is also Scipy compatible.

  import scipy.optimize as spo
  
  bo.set_seed(seed=1)
  
  x0 = bo.random_sample(of.get_bounds(), batch_size=1)[0]
  
  result = spo.minimize(
      of.evaluate,
      x0,
      bounds=of.get_bounds(),
      method=bo.minimize,
      tol=1e-5,
      options={
          'maxiter': of.get_max_eval(),
          'disp': True
      }
  )
  

• There are more examples in examples/ directory. Check them out!

Develop BOlib

• Update API documentation

  source ./.env/bin/activate
  pip install Sphinx
  cd docs/
  sphinx-apidoc -f -o ./ ../bolib