Emukit

https://emukit.readthedocs.io/en/latest/
Admin
Emukit is a model-agnostic Python toolkit for decision making under uncertainty, enabling flexible...
Tokens:80,490
Snippets:574
Trust Score:9.6
Update:5 months ago
Show doc for...
Tokens:
Raw
### Install Emukit from Source Step-by-Step

Source: https://emukit.readthedocs.io/en/latest/installation.html

Clone the Emukit repository, navigate to the directory, install dependencies, and then install Emukit in development mode. This provides more control over the installation process.

```bash
git clone https://github.com/emukit/emukit.git
cd Emukit
pip install -r requirements/requirements.txt
python setup.py develop
```

--------------------------------

### Install Emukit from GitHub using pip

Source: https://emukit.readthedocs.io/en/latest/installation.html

Install Emukit directly from its GitHub repository. This is useful for getting the latest development version.

```bash
pip install git+https://github.com/emukit/emukit.git
```

--------------------------------

### Install Emukit using pip

Source: https://emukit.readthedocs.io/en/latest/installation.html

Use this command to install the Emukit package directly from PyPI. Ensure Python 3.5+ and NumPy are installed.

```bash
pip install emukit
```

--------------------------------

### Enable Emukit Logging

Source: https://emukit.readthedocs.io/en/latest/usage.html

Use this code to enable logging before calling Emukit functions to monitor progress. Basic configuration is shown; refer to Python documentation for advanced setups.

```python
import logging
logging.basicConfig(level=logging.INFO)
```

--------------------------------

### Set up Emukit parameter space and experimental design loop

Source: https://emukit.readthedocs.io/en/latest/notebooks/Emukit-tutorial-custom-model.html

Defines the parameter space for the optimization and initializes the `ExperimentalDesignLoop` with the custom Emukit model. The loop is then run to collect data by optimizing the acquisition function.

```python
p = ContinuousParameter('c', x_min, x_max)
space = ParameterSpace([p])

loop = ExperimentalDesignLoop(space, emukit_model)
loop.run_loop(np.sin, 50)
```

--------------------------------

### get

Source: https://emukit.readthedocs.io/en/latest/genindex.html

Retrieves an item, likely from an AnchorPointsGenerator.

```APIDOC
## Method: get

### Description
Retrieves an item. This method is part of the AnchorPointsGenerator class.

### Class
emukit.core.optimization.anchor_points_generator.AnchorPointsGenerator
```

--------------------------------

### QuadratureProductKernel Methods

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Methods available for QuadratureProductKernel.

```APIDOC
## QuadratureProductKernel.qK()

### Description
Calculates the kernel.

### Method
N/A (Method call)

### Endpoint
N/A

### Parameters
None explicitly documented.

### Request Example
N/A

### Response
N/A

## QuadratureProductKernel.qKq()

### Description
Calculates the kernel with a second input.

### Method
N/A (Method call)

### Endpoint
N/A

### Parameters
None explicitly documented.

### Request Example
N/A

### Response
N/A

## QuadratureProductKernel.dqK_dx()

### Description
Calculates the gradient of the kernel with respect to the first input.

### Method
N/A (Method call)

### Endpoint
N/A

### Parameters
None explicitly documented.

### Request Example
N/A

### Response
N/A
```

--------------------------------

### AnchorPointsGenerator.get

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Gets anchor points using AnchorPointsGenerator.

```APIDOC
## AnchorPointsGenerator.get()

### Description
Gets anchor points using AnchorPointsGenerator.

### Method
Not specified (assumed to be a method call within a Python context).

### Parameters
None explicitly defined in the source.

### Request Example
```python
anchor_points_generator_instance.get()
```

### Response
None explicitly defined in the source.
```

--------------------------------

### emukit.core.initial_designs

Source: https://emukit.readthedocs.io/en/latest/api/emukit.core.html

Provides strategies for generating initial sample points for Bayesian optimization.

```APIDOC
## InitialDesignBase

### Description
Base class for initial design strategies.

### Methods
- `get_samples()`: Generates initial samples.
```

```APIDOC
## RandomDesign

### Description
Generates initial samples randomly within the search space.

### Methods
- `get_samples()`: Generates random samples.
```

--------------------------------

### IDifferentiable.get_prediction_gradients

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Gets the prediction gradients from the IDifferentiable interface.

```APIDOC
## IDifferentiable.get_prediction_gradients()

### Description
Gets the prediction gradients from the IDifferentiable interface.

### Method
Not specified (assumed to be a method call within a Python context).

### Parameters
None explicitly defined in the source.

### Request Example
```python
idifferentiable_instance.get_prediction_gradients()
```

### Response
None explicitly defined in the source.
```

--------------------------------

### emukit.model_wrappers.ProductMatern32GPy

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Product Matern 3/2 kernel for GPy models.

```APIDOC
## ProductMatern32GPy

### Description
Represents a product of Matern 3/2 kernels for GPy models.

### Methods
- `K(X1, X2)`: Computes the kernel matrix between points `X1` and `X2`.
- `dK_dx1(X)`: Computes the gradient of the kernel matrix with respect to the first input.

### Properties
- `lengthscales`: The lengthscales of the Matern kernels.
- `variance`: The variance of the Matern kernels.
```

--------------------------------

### ObjectiveAnchorPointsGenerator.get_anchor_point_scores

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Gets anchor point scores using ObjectiveAnchorPointsGenerator.

```APIDOC
## ObjectiveAnchorPointsGenerator.get_anchor_point_scores()

### Description
Gets anchor point scores using ObjectiveAnchorPointsGenerator.

### Method
Not specified (assumed to be a method call within a Python context).

### Parameters
None explicitly defined in the source.

### Request Example
```python
objective_anchor_points_generator_instance.get_anchor_point_scores()
```

### Response
None explicitly defined in the source.
```

--------------------------------

### OuterLoop.get_next_points

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Gets the next points to evaluate using OuterLoop.

```APIDOC
## OuterLoop.get_next_points()

### Description
Gets the next points to evaluate using OuterLoop.

### Method
Not specified (assumed to be a method call within a Python context).

### Parameters
None explicitly defined in the source.

### Request Example
```python
outer_loop_instance.get_next_points()
```

### Response
None explicitly defined in the source.
```

--------------------------------

### emukit.model_wrappers.ProductMatern12GPy

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Product Matern 1/2 kernel for GPy models.

```APIDOC
## ProductMatern12GPy

### Description
Represents a product of Matern 1/2 kernels for GPy models.

### Methods
- `K(X1, X2)`: Computes the kernel matrix between points `X1` and `X2`.
- `dK_dx1(X)`: Computes the gradient of the kernel matrix with respect to the first input.

### Properties
- `lengthscales`: The lengthscales of the Matern kernels.
- `variance`: The variance of the Matern kernels.
```

--------------------------------

### InitialDesignBase.get_samples

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Retrieves samples from the InitialDesignBase.

```APIDOC
## InitialDesignBase.get_samples()

### Description
Retrieves samples from the InitialDesignBase.

### Method
Not specified (assumed to be a method call within a Python context).

### Parameters
None explicitly defined in the source.

### Request Example
```python
initial_design_base_instance.get_samples()
```

### Response
None explicitly defined in the source.
```

--------------------------------

### IJointlyDifferentiable.get_joint_prediction_gradients

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Gets the joint prediction gradients from the IJointlyDifferentiable interface.

```APIDOC
## IJointlyDifferentiable.get_joint_prediction_gradients()

### Description
Gets the joint prediction gradients from the IJointlyDifferentiable interface.

### Method
Not specified (assumed to be a method call within a Python context).

### Parameters
None explicitly defined in the source.

### Request Example
```python
i_jointly_differentiable_instance.get_joint_prediction_gradients()
```

### Response
None explicitly defined in the source.
```

--------------------------------

### QuadratureProductKernel

Source: https://emukit.readthedocs.io/en/latest/api/emukit.quadrature.html

Product kernel function for quadrature.

```APIDOC
## QuadratureProductKernel

### Description
Product kernel function for quadrature.

### Methods
- `qK()`: Computes the quadrature kernel.
- `qKq()`: Computes the quadrature kernel with query points.
- `dqK_dx()`: Computes the gradient of the quadrature kernel.
```

--------------------------------

### EntropySearch

Source: https://emukit.readthedocs.io/en/latest/api/emukit.bayesian_optimization.acquisitions.html

The EntropySearch acquisition function aims to maximize the information gain about the location of the optimum. It approximates the distribution of the global optimum and uses this to guide the search.

```APIDOC
## class _emukit.bayesian_optimization.acquisitions.entropy_search.EntropySearch(_model_ , _space_ , _sampler =None_, _num_samples =100_, _num_representer_points =50_, _proposal_function =None_, _burn_in_steps =50_)

### Description
Bases: `Acquisition`

### Methods
#### update_parameters()
Updates p_min parameter

Return type:
`None`

#### update_pmin()
Approximates the distribution of the global optimum p(x=x_star|D) by doing the following steps:

* discretizing the input space by representer points sampled from a proposal measure (default EI)
* predicting mean and the covariance matrix of these representer points
* uses EPMGP algorithm to compute the probability of each representer point being the minimum

Return type:
`ndarray`

#### evaluate(_x_)
Computes the information gain, i.e the change in entropy of p_min if we would evaluate x.

Parameters:
* **x** (`ndarray`) – points where the acquisition is evaluated.

Return type:
`ndarray`

### Properties
#### _has_gradients 
Returns that this acquisition has gradients


```

--------------------------------

### WSABILLoop

Source: https://emukit.readthedocs.io/en/latest/api/emukit.quadrature.loop.html

The loop for WSABI-L.

```APIDOC
## WSABILLoop

### Description
The loop for WSABI-L.

### Parameters
* **model** (WSABIL) - The WSABI-L model.
* **model_updater** (Optional[ModelUpdater]) - Defines how and when the quadrature model is updated if new data arrives. Defaults to updating hyper-parameters every iteration.
* **acquisition_optimizer** (Optional[AcquisitionOptimizerBase]) - Optimizer selecting next evaluation points by maximizing acquisition. Gradient based optimizer is used if None. Defaults to None.
```

--------------------------------

### Import necessary Emukit and utility libraries

Source: https://emukit.readthedocs.io/en/latest/notebooks/Emukit-tutorial-custom-model.html

Imports core Emukit components for experimental design and parameter spaces, along with NumPy for numerical operations.

```python
import numpy as np

from emukit.experimental_design import ExperimentalDesignLoop
from emukit.core import ParameterSpace, ContinuousParameter
from emukit.core.loop import UserFunctionWrapper
```

--------------------------------

### QuadratureKernel Methods

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Methods for the QuadratureKernel class, used in quadrature methods.

```APIDOC
## QuadratureKernel

### Description
A kernel designed for use in quadrature methods.

### Methods

#### `K(X1, X2)`
Computes the kernel matrix between two sets of points `X1` and `X2`.

#### `qK(X)`
Computes the kernel matrix between a set of points `X` and the quadrature points.

#### `Kq(X)`
Computes the kernel matrix between the quadrature points and a set of points `X`.

#### `qKq()`
Computes the kernel matrix between the quadrature points.

#### `dK_dx1(X1, X2)`
Computes the gradient of the kernel matrix with respect to the first set of points `X1`.

#### `dK_dx2(X1, X2)`
Computes the gradient of the kernel matrix with respect to the second set of points `X2`.

#### `dKdiag_dx(X)`
Computes the gradient of the diagonal of the kernel matrix with respect to the input points `X`.

#### `dqK_dx(X)`
Computes the gradient of `qK` with respect to the input points `X`.

#### `dKq_dx(X)`
Computes the gradient of `Kq` with respect to the input points `X`.
```

--------------------------------

### Initialize and train scikit-learn GP model

Source: https://emukit.readthedocs.io/en/latest/notebooks/Emukit-tutorial-custom-model.html

Generates sample data and trains a scikit-learn GaussianProcessRegressor on it. This model will later be wrapped for Emukit.

```python
x_min = -30.0
x_max = 30.0

X = np.random.uniform(x_min, x_max, (10, 1))
Y = np.sin(X) + np.random.randn(10, 1) * 0.05
sklearn_gp = GaussianProcessRegressor();
sklearn_gp.fit(X, Y);
```

--------------------------------

### IBaseGaussianProcess Interface

Source: https://emukit.readthedocs.io/en/latest/api/emukit.quadrature.html

Interface for base Gaussian Process models used in quadrature.

```APIDOC
## IBaseGaussianProcess

### Description
Interface for base Gaussian Process models.

### Methods
- `observation_noise_variance`: The observation noise variance.
- `predict_with_full_covariance()`: Predicts the mean and full covariance.
- `solve_linear()`: Solves a linear system.
- `graminv_residual()`: Computes the residual of the Gram matrix inverse.
- `get_prediction_gradients()`: Computes gradients of the prediction.
```

--------------------------------

### ProductMatern32GPy

Source: https://emukit.readthedocs.io/en/latest/api/emukit.model_wrappers.html

Wrapper for GPy's Matern32 product kernel. This class is designed for EmuKit quadrature methods and supports initialization with either a GPy kernel object or by specifying lengthscales and variance. It provides methods for kernel evaluation and gradient calculation.

```APIDOC
## class emukit.model_wrappers.gpy_quadrature_wrappers.ProductMatern32GPy

### Description
Wrapper of the GPy Matern32 product kernel as required for some EmuKit quadrature methods.
The product kernel is of the form k(x,x′)=σ2∏i=1dki(x,x′) where
ki(x,x′)=(1+3ri)e−3ri.
d is the input dimensionality, ri:=|xi−xi′|λi, σ2 is the `variance` property and λi is the i th element of the `lengthscales` property.

### Parameters
* **gpy_matern** (`Union`[`Matern32`, `Prod`, `None`]) – A Matern32 product kernel from GPy. For d=1 this is equivalent to a Matern32 kernel. For d>1, this is _not_ a d-dimensional Matern32 kernel but a product of d 1-dimensional Matern32 kernels with differing active dimensions constructed as k1 * k2 * … . Make sure to unlink all variances except the variance of the first kernel k1 in the product as the variance of k1 will be used to represent σ2. If you are unsure what to do, use the `lengthscales` and `variance` parameter instead. If `gpy_matern` is not given, the `lengthscales` argument is used.
* **lengthscales** (`Optional`[`ndarray`]) – If `gpy_matern` is not given, a product Matern32 kernel will be constructed with the given lengthscales. The number of elements need to be equal to the dimensionality d. If `gpy_matern` is given, this input is disregarded.
* **variance** (`Optional`[`float`]) – The variance of the product kernel. Only used if `gpy_matern` is not given. Defaults to 1.

### Properties
* **lengthscales** (`ndarray`) – The lengthscales λ of the kernel.
* **variance** (`float`) – The scale σ2 of the kernel.

### Methods
#### K(x1, x2)
Evaluates the kernel k(x1, x2) at x1 and x2.

Parameters:
* **x1** (`ndarray`) – First argument of the kernel, shape (n_points N, input_dim)
* **x2** (`ndarray`) – Second argument of the kernel, shape (n_points M, input_dim)

Returns:
`ndarray` – Kernel evaluated at x1, x2, shape (N, M).

#### dK_dx1(x1, x2)
Computes the gradient of the kernel with respect to x1 evaluated at pair x1, x2.

Parameters:
* **x1** (`ndarray`) – First argument of the kernel, shape (n_points N, input_dim)
* **x2** (`ndarray`) – Second argument of the kernel, shape (n_points M, input_dim)

Returns:
`ndarray` – The gradient of the kernel wrt x1 evaluated at (x1, x2), shape (input_dim, N, M).
```

--------------------------------

### emukit.core.optimization

Source: https://emukit.readthedocs.io/en/latest/api/emukit.core.html

Provides algorithms and utilities for optimizing acquisition functions.

```APIDOC
## AcquisitionOptimizerBase

### Description
Base class for acquisition function optimizers.

### Methods
- `optimize()`: Optimizes the acquisition function.
```

```APIDOC
## AnchorPointsGenerator

### Description
Generates anchor points for optimization.

### Methods
- `get_anchor_point_scores()`: Computes scores for anchor points.
- `get()`: Retrieves anchor points.
```

```APIDOC
## ObjectiveAnchorPointsGenerator

### Description
Generates anchor points based on objective function values.

### Methods
- `get_anchor_point_scores()`: Computes scores for anchor points.
```

```APIDOC
## ContextManager

### Description
Manages context for optimization.

### Methods
- `expand_vector()`: Expands a vector with context.
```

```APIDOC
## GradientAcquisitionOptimizer

### Description
Optimizer that uses gradients to find the maximum of the acquisition function.
```

```APIDOC
## LocalSearchAcquisitionOptimizer

### Description
Performs local search to optimize the acquisition function.
```

```APIDOC
## MultiSourceAcquisitionOptimizer

### Description
Optimizes acquisition functions from multiple sources.

### Methods
- `optimize()`: Optimizes the acquisition function.
```

```APIDOC
## Optimizer

### Description
Generic optimizer class.

### Methods
- `optimize()`: Optimizes a function.
```

```APIDOC
## OptLbfgs

### Description
L-BFGS optimizer.

### Methods
- `optimize()`: Optimizes a function using L-BFGS.
```

```APIDOC
## apply_optimizer()

### Description
Applies a specified optimizer to a function.
```

```APIDOC
## OptimizationWithContext

### Description
Handles optimization with contextual information.

### Methods
- `f_no_context()`: Objective function without context.
- `df_no_context()`: Gradient of the objective function without context.
- `f_df_no_context()`: Objective function and its gradient without context.
```

```APIDOC
## OptTrustRegionConstrained

### Description
Trust region constrained optimizer.

### Methods
- `optimize()`: Optimizes a function with trust region constraints.
```

```APIDOC
## RandomSearchAcquisitionOptimizer

### Description
Performs random search to optimize the acquisition function.
```

--------------------------------

### emukit.model_wrappers.ProductMatern52GPy

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Product Matern 5/2 kernel for GPy models.

```APIDOC
## ProductMatern52GPy

### Description
Represents a product of Matern 5/2 kernels for GPy models.

### Properties
- `lengthscales`: The lengthscales of the Matern kernels.
- `variance`: The variance of the Matern kernels.
```

--------------------------------

### get_samples

Source: https://emukit.readthedocs.io/en/latest/genindex.html

Retrieves samples from an initial design or sampler.

```APIDOC
## Method: get_samples

### Description
Retrieves samples. This method is part of the InitialDesignBase class and is also implemented in RandomDesign, AffineInvariantEnsembleSampler, and McmcSampler.

### Class
emukit.core.initial_designs.base.InitialDesignBase

### Implementations
- emukit.core.initial_designs.random_design.RandomDesign
- emukit.samplers.mcmc_sampler.AffineInvariantEnsembleSampler
- emukit.samplers.mcmc_sampler.McmcSampler
```

--------------------------------

### IBaseGaussianProcess Interface

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Interface for base Gaussian process models.

```APIDOC
## IBaseGaussianProcess

### Description
Interface defining the essential methods for a Gaussian process model.

### Properties

#### `observation_noise_variance`
The variance of the observation noise.

### Methods

#### `predict_with_full_covariance(X)`
Predicts the mean and full covariance of the output for given input points `X`.

#### `solve_linear(A, b)`
Solves a linear system `Ax = b`.

#### `graminv_residual()`
Computes the Gram matrix inverse residual.

#### `get_prediction_gradients(X)`
Computes the gradients of the predictions with respect to the input points `X`.
```

--------------------------------

### SimpleGaussianProcessModel Methods

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Methods available for the SimpleGaussianProcessModel class, used for interacting with Gaussian process models.

```APIDOC
## SimpleGaussianProcessModel

### Description
Represents a simple Gaussian Process model.

### Methods

#### `optimize()`
Optimizes the hyperparameters of the Gaussian process model.

#### `predict(X)`
Predicts the output values and their variances for given input points `X`.

#### `set_data(X, Y)`
Sets the training data for the Gaussian process model.

### Attributes

#### `X`
Training input data.

#### `Y`
Training output data.
```

--------------------------------

### emukit.model_wrappers.GPyMultiOutputWrapper

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Wrapper for multi-output GPy models.

```APIDOC
## GPyMultiOutputWrapper

### Description
A wrapper class for multi-output GPy models.

### Methods
- `calculate_variance_reduction()`: Calculates the variance reduction.
- `get_prediction_gradients(X)`: Computes the gradients of the prediction.
- `predict(X)`: Predicts the output mean and variance.
- `set_data(X, Y)`: Sets the training data.
- `optimize()`: Optimizes the model hyperparameters.
- `predict_covariance(X)`: Predicts the covariance matrix.
- `get_covariance_between_points(X1, X2)`: Computes the covariance between two sets of points.
- `generate_hyperparameters_samples()`: Generates samples from the hyperparameter posterior.
- `fix_model_hyperparameters()`: Fixes the current model hyperparameters.

### Properties
- `X`: Training inputs.
- `Y`: Training outputs.
```

--------------------------------

### MultiSourceAcquisitionOptimizer.optimize

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Optimizes the acquisition function using MultiSourceAcquisitionOptimizer.

```APIDOC
## MultiSourceAcquisitionOptimizer.optimize()

### Description
Optimizes the acquisition function using MultiSourceAcquisitionOptimizer.

### Method
Not specified (assumed to be a method call within a Python context).

### Parameters
None explicitly defined in the source.

### Request Example
```python
multi_source_acquisition_optimizer_instance.optimize()
```

### Response
None explicitly defined in the source.
```

--------------------------------

### Product Methods

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Methods available for the Product class.

```APIDOC
## Product

### Methods

- `evaluate()`
- `evaluate_with_gradients()`
- `has_gradients`
```

--------------------------------

### VanillaBayesianQuadrature Methods

Source: https://emukit.readthedocs.io/en/latest/api/emukit.quadrature.html

Documentation for the VanillaBayesianQuadrature class, a standard implementation of Bayesian Quadrature.

```APIDOC
## VanillaBayesianQuadrature

### Description
A standard implementation of Bayesian Quadrature without specific warping.

### Methods

- **`predict_base(X)`**
  - Description: Predicts the mean of the function at the given points using the base model.
  - Parameters:
    - `X` (array-like): Points at which to predict.
  - Returns: Mean prediction.

- **`predict_base_with_full_covariance(X)`**
  - Description: Predicts the mean and full covariance of the function at the given points using the base model.
  - Parameters:
    - `X` (array-like): Points at which to predict.
  - Returns: Mean and full covariance prediction.

- **`integrate()`**
  - Description: Computes the integral of the function using the Bayesian Quadrature model.
  - Returns: The estimated integral value.

- **`get_prediction_gradients(X)`**
  - Description: Computes the gradients of the prediction at the given points.
  - Parameters:
    - `X` (array-like): Points at which to compute gradients.
  - Returns: Prediction gradients.
```

--------------------------------

### WSABILLoop

Source: https://emukit.readthedocs.io/en/latest/api/emukit.quadrature.html

Implements the WSA-BI loop for Bayesian quadrature.

```APIDOC
## WSABILLoop

### Description
WSA-BI loop for Bayesian quadrature.
```

--------------------------------

### create_emukit_model_from_gpy_model

Source: https://emukit.readthedocs.io/en/latest/api/emukit.model_wrappers.html

Wraps a GPy model to create an EmuKit quadrature model, facilitating integration with EmuKit's quadrature methods.

```APIDOC
## emukit.model_wrappers.gpy_quadrature_wrappers.create_emukit_model_from_gpy_model

### Description
Wraps a GPy model and returns an EmuKit quadrature model.

### Parameters
- **gpy_model** (GPy Model) - The GPy model to wrap.
- **integral_bounds** (Optional[Tuple[float, float]]) - The bounds for the integral.
- **measure** (Optional[Measure]) - The measure for the integral.
- **integral_name** (str) - The name of the integral. Defaults to ''.
```

--------------------------------

### ProductMatern12GPy

Source: https://emukit.readthedocs.io/en/latest/api/emukit.model_wrappers.html

Wrapper for the GPy Exponential (Matern12) product kernel, used in EmuKit quadrature methods.

```APIDOC
## class _emukit.model_wrappers.gpy_quadrature_wrappers.ProductMatern12GPy(_gpy_matern =None_, _lengthscales =None_, _variance =None_)

### Description
Wrapper of the GPy Exponential (a.k.a Matern12) product kernel as required for some EmuKit quadrature methods.
The product kernel is of the form k(x,x′)=σ2∏i=1dki(x,x′) where
ki(x,x′)=e−ri.
d is the input dimensionality, ri:=|xi−xi′|λi, σ2 is the `variance` property and λi is the i th element of the `lengthscales` property.

### Parameters
* **gpy_matern** (`Union`[`Exponential`, `Prod`, `None`]) – An Exponential (a.k.a. Matern12) product kernel from GPy. For d=1 this is equivalent to an Exponential kernel. For d>1, this is _not_ a d-dimensional Exponential kernel but a product of d 1-dimensional Exponential kernels with differing active dimensions constructed as k1 * k2 * … . Make sure to unlink all variances except the variance of the first kernel k1 in the product as the variance of k1 will be used to represent σ2. If you are unsure what to do, use the `lengthscales` and `variance` parameter instead. If `gpy_matern` is not given, the `lengthscales` argument is used.
```

--------------------------------

### InitialDesignBase.get_samples

Source: https://emukit.readthedocs.io/en/latest/api/emukit.core.initial_designs.html

Generates a specified number of sample points within the parameter space. This is a base method intended to be implemented by subclasses.

```APIDOC
## InitialDesignBase.get_samples

### Description
Generates requested amount of points.

### Parameters
#### Path Parameters
- **point_count** (int) - Required - Number of points required.

### Response
#### Success Response (200)
- **samples** (ndarray) - A numpy array of generated samples, shape (point_count x space_dim)
```

--------------------------------

### QuadratureProductMatern32

Source: https://emukit.readthedocs.io/en/latest/api/emukit.quadrature.html

Product of Matern 3/2 kernels for quadrature.

```APIDOC
## QuadratureProductMatern32

### Description
Product of Matern 3/2 kernels for quadrature.

### Methods
- `nu`: The smoothness parameter nu.
- `lengthscales`: The lengthscales of the Matern 3/2 kernel.
- `variance`: The variance of the Matern 3/2 kernel.
```

--------------------------------

### ProductMatern52GPy

Source: https://emukit.readthedocs.io/en/latest/api/emukit.model_wrappers.html

Wrapper for GPy's Matern52 product kernel. This class is designed for EmuKit quadrature methods and supports initialization with either a GPy kernel object or by specifying lengthscales and variance. It provides methods for kernel evaluation and gradient calculation.

```APIDOC
## class emukit.model_wrappers.gpy_quadrature_wrappers.ProductMatern52GPy

### Description
Wrapper of the GPy Matern52 product kernel as required for some EmuKit quadrature methods.
The product kernel is of the form k(x,x′)=σ2∏i=1dki(x,x′) where
ki(x,x′)=(1+5ri+53ri2)exp⁡(−5ri).
d is the input dimensionality, ri:=|xi−xi′|λi, σ2 is the `variance` property and λi is the i th element of the `lengthscales` property.

### Parameters
* **gpy_matern** (`Union`[`Matern52`, `Prod`, `None`]) – A Matern52 product kernel from GPy. For d=1 this is equivalent to a Matern52 kernel. For d>1, this is _not_ a d-dimensional Matern52 kernel but a product of d 1-dimensional Matern52 kernels with differing active dimensions constructed as k1 * k2 * … . Make sure to unlink all variances except the variance of the first kernel k1 in the product as the variance of k1 will be used to represent σ2. If you are unsure what to do, use the `lengthscales` and `variance` parameter instead. If `gpy_matern` is not given, the `lengthscales` argument is used.
* **lengthscales** (`Optional`[`ndarray`]) – If `gpy_matern` is not given, a product Matern52 kernel will be constructed with the given lengthscales. The number of elements need to be equal to the dimensionality d. If `gpy_matern` is given, this input is disregarded.
* **variance** (`Optional`[`float`]) – The variance of the product kernel. Only used if `gpy_matern` is not given. Defaults to 1.

### Properties
* **lengthscales** (`ndarray`) – The lengthscales λ of the kernel.
* **variance** (`float`) – The scale σ2 of the kernel.

### Methods
#### K(x1, x2)
Evaluates the kernel k(x1, x2) at x1 and x2.

Parameters:
* **x1** (`ndarray`) – First argument of the kernel, shape (n_points N, input_dim)
* **x2** (`ndarray`) – Second argument of the kernel, shape (n_points M, input_dim)

Returns:
`ndarray` – Kernel evaluated at x1, x2, shape (N, M).

#### dK_dx1(x1, x2)
Computes the gradient of the kernel with respect to x1 evaluated at pair x1, x2.

Parameters:
* **x1** (`ndarray`) – First argument of the kernel, shape (n_points N, input_dim)
* **x2** (`ndarray`) – Second argument of the kernel, shape (n_points M, input_dim)

Returns:
`ndarray` – The gradient of the kernel wrt x1 evaluated at (x1, x2), shape (input_dim, N, M).
```

--------------------------------

### AcquisitionOptimizerBase.optimize

Source: https://emukit.readthedocs.io/en/latest/api/emukit.core.optimization.html

Abstract method to optimize the acquisition function. Subclasses must implement the specific optimization logic.

```APIDOC
## optimize

### Description
Optimizes the acquisition function.

### Parameters
* **acquisition** (Acquisition) – The acquisition function to be optimized
* **context** (Optional[Dict[str, Any]]) – Optimization context. Determines whether any variable values should be fixed during the optimization

### Returns
Tuple of (location of maximum, acquisition value at maximizer)
```

--------------------------------

### QuadratureKernel

Source: https://emukit.readthedocs.io/en/latest/api/emukit.quadrature.html

Kernel function for quadrature.

```APIDOC
## QuadratureKernel

### Description
Kernel function for quadrature.

### Methods
- `K()`: Computes the kernel matrix.
- `qK()`: Computes the quadrature kernel.
- `Kq()`: Computes the kernel matrix with a query point.
- `qKq()`: Computes the quadrature kernel with query points.
- `dK_dx1()`: Computes the gradient of the kernel with respect to the first input.
- `dK_dx2()`: Computes the gradient of the kernel with respect to the second input.
- `dKdiag_dx()`: Computes the gradient of the diagonal of the kernel matrix.
- `dqK_dx()`: Computes the gradient of the quadrature kernel.
- `dKq_dx()`: Computes the gradient of the kernel with respect to the query point.
```

--------------------------------

### OptimizationWithContext Class

Source: https://emukit.readthedocs.io/en/latest/api/emukit.core.optimization.html

Handles optimization objectives that involve context variables, providing wrappers to manage context addition.

```APIDOC
## class emukit.core.optimization.optimizer.OptimizationWithContext(x0, f_, df=None, f_df=None, context_manager=None)

### Description
Handles optimization objectives that involve context variables.

### Methods
#### f_no_context(x)

##### Parameters
* **x** (`ndarray`) – Input without context variables

##### Return type
`ndarray`

#### df_no_context(x)

##### Parameters
* **x** (`ndarray`) – Input without context variables

##### Return type
`ndarray`

#### f_df_no_context(x)

##### Parameters
* **x** (`ndarray`) – Input without context variables

##### Return type
`Tuple`[`ndarray`, `ndarray`]
```

--------------------------------

### emukit.core.acquisition

Source: https://emukit.readthedocs.io/en/latest/api/emukit.core.html

Provides base classes and implementations for acquisition functions used in Bayesian optimization.

```APIDOC
## Acquisition

### Description
Base class for acquisition functions.

### Methods
- `evaluate()`: Evaluates the acquisition function.
- `evaluate_with_gradients()`: Evaluates the acquisition function and its gradients.
- `update_parameters()`: Updates the parameters of the acquisition function.

### Properties
- `has_gradients` (bool): Whether the acquisition function provides gradients.
```

```APIDOC
## Quotient

### Description
Acquisition function that computes the quotient of two other acquisition functions.

### Methods
- `evaluate()`: Evaluates the acquisition function.
- `evaluate_with_gradients()`: Evaluates the acquisition function and its gradients.
- `update_parameters()`: Updates the parameters of the acquisition function.

### Properties
- `has_gradients` (bool): Whether the acquisition function provides gradients.
```

```APIDOC
## Product

### Description
Acquisition function that computes the product of two other acquisition functions.

### Methods
- `evaluate()`: Evaluates the acquisition function.
- `evaluate_with_gradients()`: Evaluates the acquisition function and its gradients.
- `update_parameters()`: Updates the parameters of the acquisition function.

### Properties
- `has_gradients` (bool): Whether the acquisition function provides gradients.
```

```APIDOC
## Sum

### Description
Acquisition function that computes the sum of two other acquisition functions.

### Methods
- `evaluate()`: Evaluates the acquisition function.
- `evaluate_with_gradients()`: Evaluates the acquisition function and its gradients.
- `update_parameters()`: Updates the parameters of the acquisition function.

### Properties
- `has_gradients` (bool): Whether the acquisition function provides gradients.
```

```APIDOC
## acquisition_per_expected_cost()

### Description
Calculates the acquisition per expected cost.
```

```APIDOC
## CostAcquisition

### Description
Base class for acquisition functions that consider cost.

### Methods
- `evaluate()`: Evaluates the acquisition function.
- `evaluate_with_gradients()`: Evaluates the acquisition function and its gradients.

### Properties
- `has_gradients` (bool): Whether the acquisition function provides gradients.
```

```APIDOC
## IntegratedHyperParameterAcquisition

### Description
Acquisition function that integrates over hyperparameter distributions.

### Methods
- `evaluate()`: Evaluates the acquisition function.
- `evaluate_with_gradients()`: Evaluates the acquisition function and its gradients.
- `update_parameters()`: Updates the parameters of the acquisition function.
- `update_batches()`: Updates the batches for hyperparameter integration.

### Properties
- `has_gradients` (bool): Whether the acquisition function provides gradients.
```

--------------------------------

### emukit.core.ParameterSpace

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Defines a space of parameters for experiments.

```APIDOC
## ParameterSpace

### Description
Represents a space of parameters, defining the search domain for experiments.

### Methods
- `find_parameter_index_in_model(parameter_name)`: Finds the index of a parameter in the model.
- `get_parameter_by_name(parameter_name)`: Retrieves a parameter object by its name.
- `get_bounds()`: Returns the bounds of all parameters.
- `round(X)`: Rounds the given points `X` to the nearest valid values within the parameter space.
- `check_points_in_domain(X)`: Checks if the given points `X` are within the domain of the parameter space.
- `sample_uniform(num_samples)`: Samples uniformly from the parameter space.

### Properties
- `dimensionality`: The total number of dimensions in the parameter space.
- `parameters`: A list of parameter objects in the space.
- `parameter_names`: A list of names of the parameters.
```

--------------------------------

### LocalPenalizationPointCalculator Methods

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Methods available for the LocalPenalizationPointCalculator class.

```APIDOC
## LocalPenalizationPointCalculator

### Methods

- `compute_next_points()`
```

--------------------------------

### AcquisitionOptimizerBase.optimize

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Optimizes the acquisition function using AcquisitionOptimizerBase.

```APIDOC
## AcquisitionOptimizerBase.optimize()

### Description
Optimizes the acquisition function using AcquisitionOptimizerBase.

### Method
Not specified (assumed to be a method call within a Python context).

### Parameters
None explicitly defined in the source.

### Request Example
```python
acquisition_optimizer_base_instance.optimize()
```

### Response
None explicitly defined in the source.
```

--------------------------------

### QuadratureKernel

Source: https://emukit.readthedocs.io/en/latest/api/emukit.quadrature.kernels.html

Abstract base class for a kernel augmented with integrability. Specific implementations require a standard kernel inheriting from IStandardKernel and an IntegrationMeasure.

```APIDOC
## Class QuadratureKernel

### Description
Abstract class for a kernel augmented with integrability. Each specific implementation of this class must go with a specific standard kernel as input which inherits from `IStandardKernel`. This is because we both want the `QuadratureKernel` to be backend agnostic but at the same time `QuadratureKernel` needs access to specifics of the standard kernel. For example a specific pair of `QuadratureKernel` and `IStandardKernel` is `QuadratureRBF` and `IRBF`. The kernel embeddings are implemented w.r.t. a specific integration measure, for example the `LebesgueMeasure`.

### Parameters
* **kern** (`IStandardKernel`) – Standard EmuKit kernel.
* **measure** (`IntegrationMeasure`) – The integration measure.
* **variable_names** (`str`) – The (variable) name(s) of the integral.

### Methods
#### K(x1, x2)

The kernel k(x1, x2) evaluated at x1 and x2.

Parameters:
* **x1** (`ndarray`) – First argument of the kernel.
* **x2** (`ndarray`) – Second argument of the kernel.

Returns:
* `ndarray` – The kernel at x1, x2.

#### qK(x2)

The kernel with the first argument integrated out (kernel mean) evaluated at x2.

Parameters:
* **x2** (`ndarray`) – The locations where the kernel mean is evaluated, shape (n_points, input_dim).

Returns:
* `ndarray` – The kernel mean at x2, shape (1, n_points).

#### Kq(x1)

The kernel with the second argument integrated out (kernel mean) evaluated at x1.

Parameters:
* **x1** (`ndarray`) – The locations where the kernel mean is evaluated, shape (n_points, input_dim).

Returns:
* `ndarray` – The kernel mean at x1, shape (n_points, 1).

#### qKq()

The kernel integrated over both arguments x1 and x2.

Returns:
* `float` – Double integrated kernel.

#### dK_dx1(x1, x2)

The gradient of the kernel wrt x1 evaluated at pair x1, x2.

Parameters:
* **x1** (`ndarray`) – First argument of the kernel, shape (n_points N, input_dim).
* **x2** (`ndarray`) – Second argument of the kernel, shape (n_points M, input_dim).

Returns:
* `ndarray` – The gradient at (x1, x2), shape (input_dim, N, M).

#### dK_dx2(x1, x2)

The gradient of the kernel wrt x2 evaluated at pair x1, x2. Note that this is equal to the transposed gradient of the kernel wrt x1 evaluated at x2 and x1 (swapped arguments).

Parameters:
* **x1** (`ndarray`) – First argument of the kernel, shape (n_points N, N, input_dim).
* **x2** (`ndarray`) – Second argument of the kernel, shape (n_points N, M, input_dim).

Returns:
* `ndarray` – The gradient at (x1, x2), shape (input_dim, N, M).

#### dKdiag_dx(x)

The gradient of the diagonal of the kernel v(x):=k(x,x) evaluated at x.

Parameters:
* **x** (`ndarray`) – The locations where the gradient is evaluated, shape = (n_points M, input_dim).

Returns:
* `ndarray` – The gradient at x, shape (input_dim, M).

#### dqK_dx(x2)

The gradient of the kernel mean (integrated in first argument) evaluated at x2.

Parameters:
* **x2** (`ndarray`) – The locations where the gradient is evaluated, shape (n_points N, N, input_dim).

Returns:
* `ndarray` – The gradient at x2, shape (input_dim, N).

#### dKq_dx(x1)

The gradient of the kernel mean (integrated in second argument) evaluated at x1.

Parameters:
* **x1** (`ndarray`) – The locations where the gradient is evaluated, shape (n_points N, N, input_dim).

Returns:
* `ndarray` – The gradient with shape (N, input_dim).
```

--------------------------------

### IStandardKernel

Source: https://emukit.readthedocs.io/en/latest/api/emukit.quadrature.interfaces.html

Interface for a standard kernel k(x, x') that in principle can be integrated.

```APIDOC
## class emukit.quadrature.interfaces.IStandardKernel

### Description
Interface for a standard kernel k(x, x’) that in principle can be integrated.

See also
  * `emukit.quadrature.interfaces.IRBF`
  * `emukit.quadrature.interfaces.IProductMatern52`
  * `emukit.quadrature.interfaces.IProductMatern32`
  * `emukit.quadrature.interfaces.IProductBrownian`

### Methods
#### K
The kernel k(x1, x2) evaluated at x1 and x2.

Parameters:
    * **x1** (`ndarray`) – First argument of the kernel, shape (n_points N, input_dim)
    * **x2** (`ndarray`) – Second argument of the kernel, shape (n_points M, input_dim)

Returns:
    `ndarray` – Kernel evaluated at x1, x2, shape (N, M).

#### dK_dx1
Gradient of the kernel wrt x1 evaluated at pair x1, x2.

Parameters:
    * **x1** (`ndarray`) – First argument of the kernel, shape (n_points N, input_dim)
    * **x2** (`ndarray`) – Second argument of the kernel, shape (n_points M, input_dim)

Returns:
    `ndarray` – The gradient of the kernel wrt x1 evaluated at (x1, x2), shape (input_dim, N, M).

#### dKdiag_dx
The gradient of the diagonal of the kernel (the variance) v(x):=k(x, x) evaluated at x.

Parameters:
    * **x** (`ndarray`) – The locations where the gradient is evaluated, shape (n_points, input_dim).

Returns:
    `ndarray` – The gradient of the diagonal of the kernel evaluated at x, shape (input_dim, n_points).
```

--------------------------------

### WSABIL Methods

Source: https://emukit.readthedocs.io/en/latest/api/emukit.quadrature.html

Documentation for the WSABIL class, likely related to warping and Bayesian integration.

```APIDOC
## WSABIL

### Description
Class related to Warping and Bayesian Integration, potentially for specific algorithms.

### Methods

- **`compute_warping_params()`**
  - Description: Computes the parameters for the warping transformation.
```

--------------------------------

### QuadratureProductMatern32

Source: https://emukit.readthedocs.io/en/latest/api/emukit.quadrature.kernels.html

Represents a product Matern32 kernel augmented with integrability. This class is compatible with the standard kernel IProductMatern32 and embeds the kernel with respect to a specific integration measure.

```APIDOC
## Class: QuadratureProductMatern32

### Description
Base class for a product Matern32 kernel augmented with integrability. The kernel is of the form k(x,x′)=σ2∏i=1dki(x,x′) where ki(x,x′)=(1+3ri)e−3ri. Above, d is the input dimensionality, ri=|xi−xi′|λi, is the scaled distance, σ2 is the `variance` property and λi is the i th element of the `lengthscales` property.

### Parameters
* **matern_kernel** (`IProductMatern32`) – The standard EmuKit product Matern32 kernel.
* **measure** (`IntegrationMeasure`) – The integration measure.
* **variable_names** (`str`) – The (variable) name(s) of the integral.

### Properties
* **nu** (`float`) – The smoothness parameter of the kernel.
* **lengthscales** (`ndarray`) – The lengthscales λ of the kernel.
* **variance** (`float`) – The scale σ2 of the kernel.
```

--------------------------------

### emukit.core.Parameter

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Base class for parameters.

```APIDOC
## Parameter

### Description
Base class for defining parameters in a parameter space.

### Methods
- `round(x)`: Rounds a value `x` to a valid representation for this parameter.
- `check_in_domain(x)`: Checks if a value `x` is within the domain of this parameter.
- `sample_uniform(num_samples)`: Samples uniformly from the domain of this parameter.

### Properties
- `dimension`: The dimensionality of the parameter.
- `model_parameters`: Parameters relevant to the model.
- `bounds`: The bounds of the parameter.
```

--------------------------------

### IStandardKernel Interface

Source: https://emukit.readthedocs.io/en/latest/api/emukit.html

Interface for standard kernels.

```APIDOC
## IStandardKernel

### Description
Interface for standard kernel functions.

### Methods

#### `K(X1, X2)`
Computes the kernel matrix between two sets of points `X1` and `X2`.

#### `dK_dx1(X1, X2)`
Computes the gradient of the kernel matrix with respect to the first set of points `X1`.

#### `dKdiag_dx(X)`
Computes the gradient of the diagonal of the kernel matrix with respect to the input points `X`.
```

--------------------------------

### IConstraint

Source: https://emukit.readthedocs.io/en/latest/api/emukit.core.html

Abstract base class for constraints. Subclasses must implement the evaluate method.

```APIDOC
## class emukit.core.constraints.IConstraint

### Description
Abstract base class for constraints. Subclasses must implement the evaluate method.

### Methods
- **evaluate(x)**: Evaluates the constraint at given input locations.

### Parameters
- **x** (ndarray): Array of shape (n_points x n_dims) containing input locations to evaluate constraint at.
```
Tokens:
Raw