qiskit.aqua.algorithms.QGAN

class QGAN(data, bounds=None, num_qubits=None, batch_size=500, num_epochs=3000, seed=7, discriminator=None, generator=None, tol_rel_ent=None, snapshot_dir=None, quantum_instance=None)

The Quantum Generative Adversarial Network algorithm.

The qGAN [1] is a hybrid quantum-classical algorithm used for generative modeling tasks.

This adaptive algorithm uses the interplay of a generative GenerativeNetwork and a discriminative DiscriminativeNetwork network to learn the probability distribution underlying given training data.

These networks are trained in alternating optimization steps, where the discriminator tries to differentiate between training data samples and data samples from the generator and the generator aims at generating samples which the discriminator classifies as training data samples. Eventually, the quantum generator learns the training data’s underlying probability distribution. The trained quantum generator loads a quantum state which is a model of the target distribution.

References:

[1] Zoufal et al.,

Quantum Generative Adversarial Networks for learning and loading random distributions

Parameters

• data (ndarray) – Training data of dimension k

• bounds (Optional[ndarray]) – k min/max data values [[min_0,max_0],…,[min_k-1,max_k-1]] if univariate data: [min_0,max_0]

• num_qubits (Optional[ndarray]) – k numbers of qubits to determine representation resolution, i.e. n qubits enable the representation of 2**n values [num_qubits_0,…, num_qubits_k-1]

• batch_size (int) – Batch size, has a min. value of 1.

• num_epochs (int) – Number of training epochs

• seed (int) – Random number seed

• discriminator (Optional[DiscriminativeNetwork]) – Discriminates between real and fake data samples

• generator (Optional[GenerativeNetwork]) – Generates ‘fake’ data samples

• tol_rel_ent (Optional[float]) – Set tolerance level for relative entropy. If the training achieves relative entropy equal or lower than tolerance it finishes.

• snapshot_dir (Optional[str]) – Directory in to which to store cvs file with parameters, if None (default) then no cvs file is created.

• quantum_instance (Union[QuantumInstance, Backend, BaseBackend, None]) – Quantum Instance or Backend

Raises

AquaError – invalid input

__init__(data, bounds=None, num_qubits=None, batch_size=500, num_epochs=3000, seed=7, discriminator=None, generator=None, tol_rel_ent=None, snapshot_dir=None, quantum_instance=None)

Parameters

• data (ndarray) – Training data of dimension k

• bounds (Optional[ndarray]) – k min/max data values [[min_0,max_0],…,[min_k-1,max_k-1]] if univariate data: [min_0,max_0]

• num_qubits (Optional[ndarray]) – k numbers of qubits to determine representation resolution, i.e. n qubits enable the representation of 2**n values [num_qubits_0,…, num_qubits_k-1]

• batch_size (int) – Batch size, has a min. value of 1.

• num_epochs (int) – Number of training epochs

• seed (int) – Random number seed

• discriminator (Optional[DiscriminativeNetwork]) – Discriminates between real and fake data samples

• generator (Optional[GenerativeNetwork]) – Generates ‘fake’ data samples

• tol_rel_ent (Optional[float]) – Set tolerance level for relative entropy. If the training achieves relative entropy equal or lower than tolerance it finishes.

• snapshot_dir (Optional[str]) – Directory in to which to store cvs file with parameters, if None (default) then no cvs file is created.

• quantum_instance (Union[QuantumInstance, Backend, BaseBackend, None]) – Quantum Instance or Backend

Raises

AquaError – invalid input

Methods

__init__(data[, bounds, num_qubits, …])	type data ndarray
get_rel_entr()	Get relative entropy between target and trained distribution
run([quantum_instance])	Execute the algorithm with selected backend.
set_backend(backend, **kwargs)	Sets backend with configuration.
set_discriminator([discriminator])	Initialize discriminator.
set_generator([generator_circuit, …])	Initialize generator.
train()	Train the qGAN

Attributes

backend	Returns backend.
d_loss	Returns discriminator loss
discriminator	Returns discriminator
g_loss	Returns generator loss
generator	Returns generator
quantum_instance	Returns quantum instance.
random	Return a numpy random.
rel_entr	Returns relative entropy between target and trained distribution
seed	Returns random seed
tol_rel_ent	Returns tolerance for relative entropy

property backend

Returns backend.

Return type

Union[Backend, BaseBackend]

property d_loss

Returns discriminator loss

Return type

List[float]

property discriminator

Returns discriminator

property g_loss

Returns generator loss

Return type

List[float]

property generator

Returns generator

get_rel_entr()

Get relative entropy between target and trained distribution

Return type

float

property quantum_instance

Returns quantum instance.

Return type

Optional[QuantumInstance]

property random

Return a numpy random.

property rel_entr

Returns relative entropy between target and trained distribution

Return type

List[float]

run(quantum_instance=None, **kwargs)

Execute the algorithm with selected backend.

Parameters

• quantum_instance (Union[QuantumInstance, Backend, BaseBackend, None]) – the experimental setting.

• kwargs (dict) – kwargs

Returns

results of an algorithm.

Return type

dict

Raises

AquaError – If a quantum instance or backend has not been provided

property seed

Returns random seed

set_backend(backend, **kwargs)

Sets backend with configuration.

Return type

None

set_discriminator(discriminator=None)

Initialize discriminator.

Parameters

discriminator (Discriminator) – discriminator

set_generator(generator_circuit=None, generator_init_params=None, generator_optimizer=None)

Initialize generator.

Parameters

• generator_circuit (Union[UnivariateVariationalDistribution, MultivariateVariationalDistribution, QuantumCircuit, None]) – parameterized quantum circuit which sets the structure of the quantum generator

• generator_init_params (Optional[ndarray]) – initial parameters for the generator circuit

• generator_optimizer (Optional[Optimizer]) – optimizer to be used for the training of the generator

property tol_rel_ent

Returns tolerance for relative entropy

train()

Train the qGAN

Raises

AquaError – Batch size bigger than the number of items in the truncated data set