Dariusz Lasecki

The goal of this epic is to rewrite code in Qiskit Terra Algorithms using primitives - Sampler and Estimator.

TODO

Includes calculating a probability distribution from input data and from a Gibbs state for a Quantum Boltzmann Machine.

The implementation of the Quantum Boltzmann Machine should follow the `NeuralNetwork` interface.

The implementation should make us of gradients calculated from a Gibbs state to calculate gradients of a loss function for a Quantum Boltzmann Machine.

Integrate convolutional neural network quantum circuits with exisiting interfaces.

1

The goal of this issue is to make use of existing interfaces for neural networks and reuse loss functions, optimizers etc. for convolutional quantum neural networks.