Alexia Salavrakos, Tigran Sedrakyan, James Mills, Rawad Mezher (May 06 2024).
Abstract: Generative machine learning models aim to learn the underlying distribution of the data in order to generate new samples. Quantum circuit Born machines (QCBMs) are a popular choice of quantum generative models, which are particularly well suited to near-term devices since they can be implemented on shallow circuits. Within the framework of photonic quantum computing, we design and simulate a QCBM that can be implemented with linear optics. We show that a newly developed error mitigation technique called recycling mitigation greatly improves the training of QCBMs in realistic scenarios with photon loss.