emukit.bayesian_optimization package

Subpackages

• emukit.bayesian_optimization.acquisitions package
  • Submodules
    • EntropySearch
      • EntropySearch.update_parameters()
      • EntropySearch.update_pmin()
      • EntropySearch.evaluate()
      • EntropySearch.has_gradients
    • MultiInformationSourceEntropySearch
    • ExpectedImprovement
      • ExpectedImprovement.evaluate()
      • ExpectedImprovement.evaluate_with_gradients()
      • ExpectedImprovement.has_gradients
    • MeanPluginExpectedImprovement
    • get_standard_normal_pdf_cdf()
    • MultipointExpectedImprovement
      • MultipointExpectedImprovement.evaluate()
      • MultipointExpectedImprovement.evaluate_with_gradients()
    • get_covariance_given_smallest()
    • get_covariance_given_value_of_i()
    • get_correlations_given_value_of_i()
    • decompose_mvn()
    • Phi_gradient()
    • Phi_hessian()
    • LocalPenalization
      • LocalPenalization.has_gradients
      • LocalPenalization.update_batches()
      • LocalPenalization.evaluate()
      • LocalPenalization.evaluate_with_gradients()
    • LogAcquisition
      • LogAcquisition.evaluate()
      • LogAcquisition.evaluate_with_gradients()
      • LogAcquisition.has_gradients
    • MaxValueEntropySearch
      • MaxValueEntropySearch.update_parameters()
      • MaxValueEntropySearch.evaluate()
      • MaxValueEntropySearch.has_gradients
    • MUMBO
      • MUMBO.update_parameters()
      • MUMBO.evaluate()
    • NegativeLowerConfidenceBound
      • NegativeLowerConfidenceBound.evaluate()
      • NegativeLowerConfidenceBound.evaluate_with_gradients()
      • NegativeLowerConfidenceBound.has_gradients
    • ProbabilityOfFeasibility
      • ProbabilityOfFeasibility.evaluate()
      • ProbabilityOfFeasibility.evaluate_with_gradients()
      • ProbabilityOfFeasibility.has_gradients
    • ProbabilityOfImprovement
      • ProbabilityOfImprovement.evaluate()
      • ProbabilityOfImprovement.evaluate_with_gradients()
      • ProbabilityOfImprovement.has_gradients
  • Module contents
• emukit.bayesian_optimization.interfaces package
  • Submodules
    • IEntropySearchModel
      • IEntropySearchModel.predict_covariance()
      • IEntropySearchModel.get_covariance_between_points()
  • Module contents
• emukit.bayesian_optimization.loops package
  • Submodules
    • BayesianOptimizationLoop
      • BayesianOptimizationLoop.get_results()
    • BayesianOptimizationResults
    • CostSensitiveBayesianOptimizationLoop
    • UnknownConstraintBayesianOptimizationLoop
  • Module contents

Submodules

emukit.bayesian_optimization.epmgp.joint_min(mu, var, with_derivatives=False)

Computes the probability of every given point to be the minimum based on the EPMGP[1] algorithm. [1] J. Cunningham, P. Hennig, and S. Lacoste-Julien. Gaussian probabilities and expectation propagation. under review. Preprint at arXiv, November 2011.

Parameters

• mu (ndarray) – Mean value of each of the N points, dims (N,).

• var (ndarray) – Covariance matrix for all points, dims (N, N).

• with_derivatives (bool) – If True than also the gradients are computed.

Return type

ndarray

Returns

pmin distribution, dims (N,1).

emukit.bayesian_optimization.epmgp.min_factor(Mu, Sigma, k, gamma=1)

emukit.bayesian_optimization.epmgp.lt_factor(s, l, M, V, mp, p, gamma)

emukit.bayesian_optimization.epmgp.log_relative_gauss(z)

class emukit.bayesian_optimization.local_penalization_calculator.LocalPenalizationPointCalculator(acquisition, acquisition_optimizer, model, parameter_space, batch_size, fixed_lipschitz_constant=None, fixed_minimum=None)

Bases: CandidatePointCalculator

Candidate point calculator that computes a batch using local penalization from:

Batch Bayesian Optimization via Local Penalization. Javier González, Zhenwen Dai, Philipp Hennig, Neil D. Lawrence

compute_next_points(loop_state, context=None)

Computes a batch of points using local penalization.

Parameters

• loop_state (LoopState) – Object containing the current state of the loop

• context (Optional[dict]) – Contains variables to fix through optimization of acquisition function. The dictionary key is the parameter name and the value is the value to fix the parameter to.

Return type

ndarray

Module contents