qml.MomentumQNGOptimizer

class MomentumQNGOptimizer(stepsize=0.01, momentum=0.9, approx='block-diag', lam=0)

Bases: pennylane.optimize.qng.QNGOptimizer

A generalization of the Quantum Natural Gradient (QNG) optimizer by considering a discrete-time Langevin equation with QNG force. For details of the theory and derivation of Momentum-QNG, please see:

Oleksandr Borysenko, Mykhailo Bratchenko, Ilya Lukin, Mykola Luhanko, Ihor Omelchenko, Andrii Sotnikov and Alessandro Lomi. “Application of Langevin Dynamics to Advance the Quantum Natural Gradient Optimization Algorithm” arXiv:2409.01978

We are grateful to David Wierichs for his generous help with the multi-argument variant of the MomentumQNGOptimizer class.

MomentumQNGOptimizer is a subclass of QNGOptimizer that requires one additional hyperparameter (the momentum coefficient) $0 \leq \rho < 1$, the default value being $\rho=0.9$. For $\rho=0$ Momentum-QNG reduces to the basic QNG. In this way, the parameter update rule in Momentum-QNG reads:

$$x^{(t+1)} = x^{(t)} + \rho (x^{(t)} - x^{(t-1)}) - \eta g(f(x^{(t)}))^{-1} \nabla f(x^{(t)}),$$

where $\eta$ is a stepsize (learning rate) value, $g(f(x^{(t)}))^{-1}$ is the pseudo-inverse of the Fubini-Study metric tensor and $f(x^{(t)}) = \langle 0 | U(x^{(t)})^\dagger \hat{B} U(x^{(t)}) | 0 \rangle$ is an expectation value of some observable measured on the variational quantum circuit $U(x^{(t)})$.

Examples:

Consider an objective function realized as a QNode that returns the expectation value of a Hamiltonian.

>>> dev = qml.device("default.qubit", wires=(0, 1, "aux"))
>>> @qml.qnode(dev)
... def circuit(params):
...     qml.RX(params[0], wires=0)
...     qml.RY(params[1], wires=0)
...     return qml.expval(qml.X(0))

Once constructed, the cost function can be passed directly to the optimizer’s step() function. In addition to the standard learning rate, the MomentumQNGOptimizer takes a momentum parameter:

>>> eta = 0.01
>>> rho = 0.93
>>> init_params = qml.numpy.array([0.5, 0.23], requires_grad=True)
>>> opt = qml.MomentumQNGOptimizer(stepsize=eta, momentum=rho)
>>> theta_new = opt.step(circuit, init_params)
>>> theta_new
tensor([0.50437193, 0.18562052], requires_grad=True)

An alternative function to calculate the metric tensor of the QNode can be provided to step via the metric_tensor_fn keyword argument, see QNGOptimizer for details.

See also

For details on quantum natural gradient, see QNGOptimizer. See MomentumOptimizer for a first-order optimizer with momentum. Also see the examples from the reference above, benchmarking the Momentum-QNG optimizer against the basic QNG, Momentum and Adam:

• QAOA

• VQE

Keyword Arguments

• stepsize=0.01 (float) – the user-defined hyperparameter $\eta$

• momentum=0.9 (float) – the user-defined hyperparameter $\rho$

• approx (str) –

  Which approximation of the metric tensor to compute.

  • If None, the full metric tensor is computed

  • If "block-diag", the block-diagonal approximation is computed, reducing the number of evaluated circuits significantly.

  • If "diag", only the diagonal approximation is computed, slightly reducing the classical overhead but not the quantum resources (compared to "block-diag").

• lam=0 (float) – metric tensor regularization $G_{ij}+\lambda I$ to be applied at each optimization step

Methods

apply_grad(grad, args)	Update the parameter array $x$ for a single optimization step.
compute_grad(objective_fn, args, kwargs[, ...])	Compute the gradient of the objective function at the given point and return it along with the objective function forward pass (if available).
step(qnode, *args[, grad_fn, ...])	Update the parameter array $x$ with one step of the optimizer.
step_and_cost(qnode, *args[, grad_fn, ...])	Update the parameter array $x$ with one step of the optimizer and return the corresponding objective function value prior to the step.

apply_grad(grad, args)

Update the parameter array $x$ for a single optimization step. Flattens and unflattens the inputs to maintain nested iterables as the parameters of the optimization.

Parameters

• grad (array) – The gradient of the objective function at point $x^{(t)}$: $\nabla f(x^{(t)})$

• args (array) – the current value of the variables $x^{(t)}$

Returns

the new values $x^{(t+1)}$

Return type

array

static compute_grad(objective_fn, args, kwargs, grad_fn=None)

Compute the gradient of the objective function at the given point and return it along with the objective function forward pass (if available).

Parameters

• objective_fn (function) – the objective function for optimization

• args (tuple) – tuple of NumPy arrays containing the current parameters for the objection function

• kwargs (dict) – keyword arguments for the objective function

• grad_fn (function) – optional gradient function of the objective function with respect to the variables args. If None, the gradient function is computed automatically. Must return the same shape of tuple [array] as the autograd derivative.

Returns

NumPy array containing the gradient $\nabla f(x^{(t)})$ and the objective function output. If grad_fn is provided, the objective function will not be evaluated and instead None will be returned.

Return type

tuple (array)

step(qnode, *args, grad_fn=None, recompute_tensor=True, metric_tensor_fn=None, **kwargs)

Update the parameter array $x$ with one step of the optimizer.

Parameters

• qnode (QNode) – the QNode for optimization

• *args – variable length argument list for qnode

• grad_fn (function) – optional gradient function of the qnode with respect to the variables *args. If None, the gradient function is computed automatically. Must return a tuple[array] with the same number of elements as *args. Each array of the tuple should have the same shape as the corresponding argument.

• recompute_tensor (bool) – Whether or not the metric tensor should be recomputed. If not, the metric tensor from the previous optimization step is used.

• metric_tensor_fn (function) – Optional metric tensor function with respect to the variables args. If None, the metric tensor function is computed automatically.

• **kwargs – variable length of keyword arguments for the qnode

Returns

the new variable values $x^{(t+1)}$

Return type

array

step_and_cost(qnode, *args, grad_fn=None, recompute_tensor=True, metric_tensor_fn=None, **kwargs)

Update the parameter array $x$ with one step of the optimizer and return the corresponding objective function value prior to the step.

Parameters

• qnode (QNode) – the QNode for optimization

• *args – variable length argument list for qnode

• grad_fn (function) – optional gradient function of the qnode with respect to the variables *args. If None, the gradient function is computed automatically. Must return a tuple[array] with the same number of elements as *args. Each array of the tuple should have the same shape as the corresponding argument.

• recompute_tensor (bool) – Whether or not the metric tensor should be recomputed. If not, the metric tensor from the previous optimization step is used.

• metric_tensor_fn (function) – Optional metric tensor function with respect to the variables args. If None, the metric tensor function is computed automatically.

• **kwargs – variable length of keyword arguments for the qnode

Returns

the new variable values $x^{(t+1)}$ and the objective function output prior to the step

Return type

tuple

code/api/pennylane.MomentumQNGOptimizer

 

Download Python script

 

Download Notebook

 

View on GitHub