Source code for pennylane.optimize.qng

# Copyright 2018 Xanadu Quantum Technologies Inc.

# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at


# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# See the License for the specific language governing permissions and
# limitations under the License.
"""Quantum natural gradient optimizer"""
#pylint: disable=too-many-branches
import autograd.numpy as np
from scipy import linalg

from pennylane.utils import _flatten, unflatten

from .gradient_descent import GradientDescentOptimizer

[docs]class QNGOptimizer(GradientDescentOptimizer): r"""Optimizer with adaptive learning rate, via calculation of the diagonal or block-diagonal approximation to the Fubini-Study metric tensor. A quantum generalization of natural gradient descent. The QNG optimizer uses a step- and parameter-dependent learning rate, with the learning rate dependent on the pseudo-inverse of the Fubini-Study metric tensor :math:`g`: .. math:: x^{(t+1)} = x^{(t)} - \eta g(f(x^{(t)}))^{-1} \nabla f(x^{(t)}), where :math:`f(x^{(t)}) = \langle 0 | U(x^{(t)})^\dagger \hat{B} U(x^{(t)}) | 0 \rangle` is an expectation value of some observable measured on the variational quantum circuit :math:`U(x^{(t)})`. Consider a quantum node represented by the variational quantum circuit .. math:: U(\mathbf{\theta}) = W(\theta_{i+1}, \dots, \theta_{N})X(\theta_{i}) V(\theta_1, \dots, \theta_{i-1}), where all parametrized gates can be written of the form :math:`X(\theta_{i}) = e^{i\theta_i K_i}`. That is, the gate :math:`K_i` is the *generator* of the parametrized operation :math:`X(\theta_i)` corresponding to the :math:`i`-th parameter. For each parametric layer :math:`\ell` in the variational quantum circuit containing :math:`n` parameters, the :math:`n\times n` block-diagonal submatrix of the Fubini-Study tensor :math:`g_{ij}^{(\ell)}` is calculated directly on the quantum device in a single evaluation: .. math:: g_{ij}^{(\ell)} = \langle \psi_\ell | K_i K_j | \psi_\ell \rangle - \langle \psi_\ell | K_i | \psi_\ell\rangle \langle \psi_\ell |K_j | \psi_\ell\rangle where :math:`|\psi_\ell\rangle = V(\theta_1, \dots, \theta_{i-1})|0\rangle` (that is, :math:`|\psi_\ell\rangle` is the quantum state prior to the application of parameterized layer :math:`\ell`). Combining the quantum natural gradient optimizer with the analytic parameter-shift rule to optimize a variational circuit with :math:`d` parameters and :math:`L` layers, a total of :math:`2d+L` quantum evaluations are required per optimization step. For more details, see: James Stokes, Josh Izaac, Nathan Killoran, Giuseppe Carleo. "Quantum Natural Gradient." `arXiv:1909.02108 <>`_, 2019. .. note:: The QNG optimizer **only supports single QNodes** as objective functions. In particular: * For hybrid classical-quantum models, the "mixed geometry" of the model makes it unclear which metric should be used for which parameter. For example, parameters of quantum nodes are better suited to one metric (such as the QNG), whereas others (e.g., parameters of classical nodes) are likely better suited to another metric. * For multi-QNode models, we don't know what geometry is appropriate if a parameter is shared amongst several QNodes. .. seealso:: See the :ref:`quantum natural gradient example <quantum_natural_gradient>` for more details on Fubini-Study metric tensor and this optimization class. Args: stepsize (float): the user-defined hyperparameter :math:`\eta` diag_approx (bool): If ``True``, forces a diagonal approximation where the calculated metric tensor only contains diagonal elements :math:`G_{ii}`. In some cases, this may reduce the time taken per optimization step. tol (float): tolerance used when finding the inverse of the quantum gradient tensor """ def __init__(self, stepsize=0.01, diag_approx=False): super().__init__(stepsize) self.diag_approx = diag_approx self.metric_tensor_inv = None
[docs] def step(self, qnode, x, recompute_tensor=True): """Update x with one step of the optimizer. Args: qnode (QNode): the QNode for optimization x (array): NumPy array containing the current values of the variables to be updated recompute_tensor (bool): Whether or not the metric tensor should be recomputed. If not, the metric tensor from the previous optimization step is used. Returns: array: the new variable values :math:`x^{(t+1)}` """ # pylint: disable=arguments-differ if not hasattr(qnode, "metric_tensor"): raise ValueError("Objective function must be encoded as a single QNode") if recompute_tensor or self.metric_tensor is None: # pseudo-inverse metric tensor metric_tensor = qnode.metric_tensor(x, diag_approx=self.diag_approx) self.metric_tensor_inv = linalg.pinvh(metric_tensor) g = self.compute_grad(qnode, x) x_out = self.apply_grad(g, x) return x_out
[docs] def apply_grad(self, grad, x): r"""Update the variables x to take a single optimization step. Flattens and unflattens the inputs to maintain nested iterables as the parameters of the optimization. Args: grad (array): The gradient of the objective function at point :math:`x^{(t)}`: :math:`\nabla f(x^{(t)})` x (array): the current value of the variables :math:`x^{(t)}` Returns: array: the new values :math:`x^{(t+1)}` """ grad_flat = np.array(list(_flatten(grad))) x_flat = np.array(list(_flatten(x))) x_new_flat = x_flat - self._stepsize * self.metric_tensor_inv @ grad_flat return unflatten(x_new_flat, x)