PhoMemtor – Photonic Quantum Memristor Networks

In the past few decades, the field of computer science has witnessed two fundamental paradigm shifts. The first was brought about by artificial neural networks, which have proven extremely effective in tasks as diverse as language recognition, medical diagnosis, advanced automation and the most advanced artificial intelligence algorithms. The second is quantum computation, which harnesses unique quantum features such as superposition and entanglement to provide dramatic advantages for solving classically intractable problems.

Prof. Magdalena Stobińska, photo by Michał Łepecki This project aims to use miniature photonic quantum systems, consisting of laser micromachined circuits and optical elements, to combine both paradigms: a demonstration of quantum neural networks that rely on novel quantum memristor devices to introduce controllable nonlinear gate operations and short-time memory. The integrated photonic processor is going to be built on a glass substrate by femtosecond laser micromachining, a technique which provides outstanding advantages such as circuit reconfigurability, low insertion losses, rapid prototyping and three-dimensional circuit topology, all of which are critical for the success of the project. The quantum processor will be capable of executing programmable finite discrete mathematical transforms. Memristors, on the other hand, are elements of photonic or electronic systems that store information and have already been the subject of wide-ranging research. By combining complementary expertise in photonic quantum computing, integrated quantum photonics and quantum information theory, we will build a tuneable photonic quantum memristor network. The versatility of this nonlinear processor will be shown by demonstrating real-life quantum-enhanced applications, ranging from speech recognition to image identification, accelerated via quantum reservoir computing architectures.

The project will lay the groundwork for a new quantum technology tapping the special properties of quantum memristors. The interdisciplinary nature of our consortium, as well as our methodology, put us in the best position to meet the conceptual and technological challenges of the project, allowing us to create the first generation of quantum neuromorphic computing hardware based on a quantum photonic platform.