PHOTONICS QUANTUM SAMPLING MACHINE

Informations projet

PHOQUSING

N° de convention de subvention: 899544

DOI

10.3030/899544

Projet clôturé

Date de signature de la CE 26 Mai 2020

Date de début 1 Septembre 2020

Date de fin 31 Août 2024

Financé au titre de

• EXCELLENT SCIENCE - Future and Emerging Technologies (FET)

Coût total

€ 3 305 955,00

Contribution de l’UE

€ 3 305 955,00

Coordonné par

UNIVERSITA DEGLI STUDI DI ROMA LA SAPIENZA
Italy

Periodic Reporting for period 3 - PHOQUSING (PHOTONICS QUANTUM SAMPLING MACHINE)

Période du rapport: 2023-03-01 au 2024-08-31

Résumé du contexte et des objectifs généraux du projet

Quantum theory provides a probabilistic description of nature, marking a departure from a classical picture. The intrinsic randomness of quantum processes has found applications in efficient probabilistic algorithms for simulation, integration, machine learning, and financial pricing, among other applications. Recently, a large research effort has been devoted to exploit quantum systems to generate probability distributions that are computationally hard for classical computers to obtain. In particular, several works reported the insight that devices exploiting quantum resources can generate probability distributions that are inaccessible with classical means. Hybrid Quantum Computational models combine classical processing with these quantum sampling machines to obtain computational advantage over standard classical means in some tasks. Moreover, NISQ (Noisy, Intermediate-Scale Quantum) technology may contribute to obtaining this advantage in the near term. The aim of this project is to implement PHOtonic Quantum SamplING (PHOQUSING) machines based on large, reconfigurable interferometers with active feedback.
Among the several objectives of this project, PHOQUSING define the most suitable architectures enabling the generation of these hard-to-sample distributions using integrated photonics, optimizing the designs and studying the tolerance to errors. Within the project, the consortium aimed at developing novel advanced components, such as solid-state single-photon sources, or multimode reconfigurable integrated processors. PHOQUSING then has foreseen the assembly two fully operating photonic quantum sampling machines, with first demonstrations of their algorithmic applications. This allowed to aim the implementation of Hybrid Quantum Computing (HQC) models that include quantum and classical elements for applications in machine learning and optimization. Prominent examples included development and implementation of randomness manipulation protocols, variational quantum algorithms, or quantum machine learning approaches. The expected results of PHOQUSING will also have applications outside the scientific community. Overall, the PHOQUSING project aimed at developing different innovative products, ranging from software to hardware for photonic quantum computing.

Travail effectué depuis le début du projet jusqu’à la fin de la période considérée dans le rapport et principaux résultats atteints jusqu’à présent

The PHOQUSING project has provided several contributions both in terms of software, hardware and applications.
Regarding the software side, theoretical work has been performed by implementing simulation software for processes based on a Boson Sampling architecture, that enabled testing different protocols in small-instance size classical simulations, with particular focus on generalized Boson Sampling and on sampling processes with nonlinearities. Furthermore, novel protocols have been defined, including an innovative approach for randomness manipulation with a modular architecture suitable for integrated devices. The consortium also defined optimized reconfigurable circuit designs. In particular, it has identified different architectures for the integrated platforms employed within the project to fully exploit the advantages of each approach. Extensive work has been carried out regarding the calibration and certification of the photonic hardware, identifying novel methodologies that has been merged in a unique and advanced toolbox.
PHOQUSING also provided several results on the hardware development and integration. More specifically, the consortium has developed different and innovative components. These innovations include significant technological advances in the efficiency of single-photon sources, based both on integrated waveguides and solid-state emitters. Furthermore, novel functionalities have been disclosed, such as the experimental generation of quantum superposition of single-photon states encoded in two frequencies. Other results involved the fabrication of advanced integrated reconfigurable photonic processors, using two of the most promising technologies (femtosecond laser-written circuits in glass and lithographic circuits in silicon nitride), leading to the fabrication of devices with unprecedented number of modes and thermo-optic phase shifters. These innovative components have been at the basis of QOLOSSUS and QALCULUS, ìsampling machine that have been developed and assembled within the project as its main objective.
Finally, new protocols oriented to applications have been devised and tested within the project on the assembled sampling machines. Prominent examples have been the implementation of the aforementioned randomness manipulation approach, called Quantum-to-Quantum Bernoulli factories, variational quantum algorithms at the interface between quantum technologies and machine learning, or novel methods for secure delegated quantum computing.
Overall, the PHOQUSING project has led to 21 publications in high impact scientific journal, 16 publications on ArXiv, and 2 patents with joint ownership between multiple partners, as well as a large activity for the dissemination of the results to the general public.

Progrès au-delà de l’état des connaissances et impact potentiel prévu (y compris l’impact socio-économique et les conséquences sociétales plus larges du projet jusqu’à présent)

Several results beyond state-of-the-art have been obtained within PHOQUSING project.
The implementation of software codes for computational analysis of different models and Boson sampling variations have been addressed, representing an important stepping stone through the final vision of PHOQUSING. Further analyses have been reported on the Boson Sampling paradigms in terms of classification of the different variants. A complete Boson sampling experiment on a 3D integrated/reconfigurable photonic chip was successfully performed, making a step forward in the complexity of reconfigurable circuits employed in this task. Furthermore, we have devised and tested several methods for the calibration of complex integrated circuits, error modeling, and for the assessment of partial photon indistinguishability of multiphoton states. This led to the development of a complete toolbox for the validation of the sampling machines operation.
From the hardware perspective, strong improvements in the complexity of reconfigurable circuits have been achieved (both for femtosecond laser-writing technology and silicon nitride platform), up to fabrication of a 128-mode device with 3-dimensional architecture. Another important result obtained is the improvement of the brightness of single-photon sources (first lens brightness >60%). This result represented an important progress beyond current-state-of-the-art in single-photon sources. Additionally, steps ahead have been done in the experimental demonstration of quantum superposition of single-photon states encoded in two frequencies. The assembling of both machines QOLOSSUS and QALCULUS have been completed. These machines simultaneously rely on interfacing different components: single-photon sources based on quantum dots, time-to-spatial demultiplexing systems and integrated circuits. These machines also included innovative functionalities, such as the capability of supporting the polarization degree of freedom.
Finally, the project has also contributed to the implementation of novel protocols on the developed sampling machines, including randomness manipulation, quantum machine learning algorithms, or innovative approaches for secure cloud quantum computing in the multiparty scenario.

figuraleaflet.png