qcbm

To train a quantum circuit Born machine, you optimize a parameterized variational quantum circuit by minimizing KL-divergence against a target discrete probability distribution using the parameter-shift gradient method.

A quantum circuit Born machine is used for generative modelling tasks to learn and sample from discrete probability distributions, such as the Bars-and-Stripes dataset, for quantum machine learning demonstrations.

The parameter-shift gradient optimizes variational quantum circuits by evaluating the circuit cost function at shifted parameter values, enabling gradient descent without direct analytical derivatives for the KL-divergence objective.

Yes, you can use a quantum circuit Born machine to generate Bars-and-Stripes data by training a layered parameterized quantum circuit on a 2×2 target distribution with 4 qubits and 4 layers.

Quantum circuit Born machines are limited to modeling small discrete probability distributions and are primarily suited for educational demonstrations of quantum machine learning rather than large-scale generative tasks.