Skip to main contentIBM Quantum Documentation
This page is from an old version of Qiskit SDK and does not exist in the latest version. We recommend you migrate to the latest version. See the release notes for more information.
Important

IBM Quantum Platform is moving and this version will be sunset on July 1. To get started on the new platform, read the migration guide.

qiskit.aqua.operators.gradients.NaturalGradient

class NaturalGradient(grad_method='lin_comb', qfi_method='lin_comb_full', regularization=None, **kwargs)

GitHub

Convert an operator expression to the first-order gradient.

Given an ill-posed inverse problem

x = arg min{||Ax-C||^2} (1)

one can use regularization schemes can be used to stabilize the system and find a numerical solution

x_lambda = arg min{||Ax-C||^2 + lambda*R(x)} (2)

where R(x) represents the penalization term.

Parameters

  • grad_method (Union[str, CircuitGradient]) – The method used to compute the state gradient. Can be either 'param_shift' or 'lin_comb' or 'fin_diff'.
  • qfi_method (Union[str, CircuitQFI]) – The method used to compute the QFI. Can be either 'lin_comb_full' or 'overlap_block_diag' or 'overlap_diag'.
  • regularization (Optional[str]) – Use the following regularization with a least square method to solve the underlying system of linear equations Can be either None or 'ridge' or 'lasso' or 'perturb_diag' 'ridge' and 'lasso' use an automatic optimal parameter search If regularization is None but the metric is ill-conditioned or singular then a least square solver is used without regularization
  • kwargs (dict) – Optional parameters for a CircuitGradient

__init__

__init__(grad_method='lin_comb', qfi_method='lin_comb_full', regularization=None, **kwargs)

Parameters

  • grad_method (Union[str, CircuitGradient]) – The method used to compute the state gradient. Can be either 'param_shift' or 'lin_comb' or 'fin_diff'.
  • qfi_method (Union[str, CircuitQFI]) – The method used to compute the QFI. Can be either 'lin_comb_full' or 'overlap_block_diag' or 'overlap_diag'.
  • regularization (Optional[str]) – Use the following regularization with a least square method to solve the underlying system of linear equations Can be either None or 'ridge' or 'lasso' or 'perturb_diag' 'ridge' and 'lasso' use an automatic optimal parameter search If regularization is None but the metric is ill-conditioned or singular then a least square solver is used without regularization
  • kwargs (dict) – Optional parameters for a CircuitGradient

Methods

__init__([grad_method, qfi_method, …])type grad_methodUnion[str, CircuitGradient]
convert(operator[, params])type operatorOperatorBase
gradient_wrapper(operator, bind_params[, …])Get a callable function which provides the respective gradient, Hessian or QFI for given parameter values.
parameter_expression_grad(param_expr, param)Get the derivative of a parameter expression w.r.t.

Attributes

grad_methodReturns CircuitGradient.
qfi_methodReturns CircuitQFI.
regularizationReturns the regularization option.

convert

convert(operator, params=None)

Parameters

  • operator (OperatorBase) – The operator we are taking the gradient of.
  • params (Union[ParameterVector, ParameterExpression, List[ParameterExpression], None]) – The parameters we are taking the gradient with respect to.

Return type

OperatorBase

Returns

An operator whose evaluation yields the NaturalGradient.

Raises

  • TypeError – If operator does not represent an expectation value or the quantum state is not CircuitStateFn.
  • ValueError – If params contains a parameter not present in operator.

grad_method

Returns CircuitGradient.

Return type

CircuitGradient

Returns

CircuitGradient.

gradient_wrapper

gradient_wrapper(operator, bind_params, grad_params=None, backend=None)

Get a callable function which provides the respective gradient, Hessian or QFI for given parameter values. This callable can be used as gradient function for optimizers.

Parameters

  • operator (OperatorBase) – The operator for which we want to get the gradient, Hessian or QFI.
  • bind_params (Union[ParameterExpression, ParameterVector, List[ParameterExpression]]) – The operator parameters to which the parameter values are assigned.
  • grad_params (Union[ParameterExpression, ParameterVector, List[ParameterExpression], Tuple[ParameterExpression, ParameterExpression], List[Tuple[ParameterExpression, ParameterExpression]], None]) – The parameters with respect to which we are taking the gradient, Hessian or QFI. If grad_params = None, then grad_params = bind_params
  • backend (Union[BaseBackend, QuantumInstance, None]) – The quantum backend or QuantumInstance to use to evaluate the gradient, Hessian or QFI.

Returns

Function to compute a gradient, Hessian or QFI. The function takes an iterable as argument which holds the parameter values.

Return type

callable(param_values)

parameter_expression_grad

static parameter_expression_grad(param_expr, param)

Get the derivative of a parameter expression w.r.t. the given parameter.

Parameters

  • param_expr (ParameterExpression) – The Parameter Expression for which we compute the derivative
  • param (ParameterExpression) – Parameter w.r.t. which we want to take the derivative

Return type

Union[ParameterExpression, float]

Returns

ParameterExpression representing the gradient of param_expr w.r.t. param

qfi_method

Returns CircuitQFI.

Returns: CircuitQFI

Return type

CircuitQFI

regularization

Returns the regularization option.

Returns: the regularization option.

Return type

Optional[str]

Was this page helpful?
Report a bug or request content on GitHub.