research
CREF promotes original and high-impact lines of research, based on physical methods, but with a strong interdisciplinary character and in relation to the main problems of the modern knowledge society.
Principal Research Lines
Individual research areas
History of Physics
third mission
The CREF was born with a dual soul: a research centre and a historical museum. Its aim is to preserve and disseminate the memory of Enrico Fermi and to promote the dissemination and communication of scientific culture.
next EVENT
Young researchers
Higher education and projects for young researchers
Open positions...
Il CREF indice una selezione pubblica, per titoli e colloquio, per il conferimento di n. 1 assegno di...
Title: Spectroscopic Benchmarks by Machine Learning as Discriminant Analysis for Unconventional Italian...
la ricerca
Photonic Technologies and Artificial Intelligence
Head of research
Claudio Conti
Overview
Enrico Fermi pioneered the creation and use of calculating machines. In recent years, the concept of the computational machine has broadened considerably and there is intense activity for designing new computers that include quantum and photonic technologies. The major industrial companies worldwide are largely involved in this direction, including IBM, NTT, HUAWEI and numerous startups in various technologically advanced nations. The lever for new computational technologies is given by the so-called end of Moore’s law, the empirical rule that has described the rapid growth of computing power in traditional semiconductor systems over the last twenty years. In recent years, we have observed the end of this growth, highlighting a stall in performance mainly due to physical limitations.
This circumstance has motivated numerous searches for new computing technologies. Recent results on the environmental impact of new artificial intelligence methodologies further fuel this research. The algorithms that are changing society today require computing resources that grow exponentially over the years and whose environmental impact can be understood by observing that the training of a single artificial intelligence produces carbon emissions comparable to that of dozens of intercontinental flights.
It is therefore necessary to develop new technologies that perform better than conventional processors and consume less energy, such as operating at room temperature without cooling systems.
Photonics is considered the most promising technology in this context. Studies demonstrate the parallel processing of enormous amounts of data using laser beams that encode the information using advanced modulation techniques. In the long term, including quantum algorithms can radically accelerate the speed of calculation and innovative cryptography methods.
Photonic and quantum systems can solve optimisation problems in polynomial time with the size of the system, a possibility often referred to as “Quantum Advantage.” What CREF wants to pursue is the development of photonic quantum systems for the acceleration of computation, which provides the result of the calculation in a robust classical form, which is not subject to decoherence and, therefore can be immediately interfaced with traditional computers
CREF researchers reported the first experimental evidence of this possibility, demonstrating optical calculations with hundreds of thousands of variables, a scale never achieved before. These proof-of-concepts will be developed extensively towards a new generation of photonic computers. Fundamental physical problems related to the role of entanglement, nonlinear effects and collective modes will also be addressed.
Purpose and Goals
The project on photonic technologies and artificial intelligence aims to experimentally demonstrate new computing machines that use light to accelerate the solution of combinatorial optimisation problems and for hybrid electronic and photonic neural networks. A further objective is to develop knowledge about these devices’ basic classical and quantum physics, through theories and numerical simulations.
The aims include the implementation of a new laboratory with optical instrumentation and creating a group of 2 researchers, one technologist, 2 research fellows, and some associates in the context of collaborations with institutions and universities.
The laboratory is under construction. Three equipped optical benches have been installed, including spatial light modulators and the first low-power laser sources, as described in the infrastructure section. A permanent researcher is currently working on the design and simulation of photonic machines. A second fixed-term researcher has been hired on a project with external PNRR funds for the young researchers initiative and is responsible for creating the prototypes and related experiments. Furthermore, a technologist is dedicated to maintaining the laboratory and the existing equipment.
The construction of the first computational machines is expected during 2023 and 2024.
Contents and Methods
The project is divided into three main directions
- Design, engineering, and computer simulation of new photonic computational machines. Theory of the classical and quantum systems that make up such machines.
- Implementation and development of a computational photonics laboratory with the aim of experimentally testing computational machines. Creation of an infrastructure equipped with advanced laser systems integrated with computational systems.
- Experimental demonstration of prototypes and new experiments in interdisciplinary, classical and quantum physics, based on photonic computational machines
Theoretical analysis is divided into two main directions: classical and quantum. With the study of classical systems, the project involves on the one hand the formulation and analysis of new mathematical models based on the nonlinear dynamics of coupled parametric oscillators, with the aim of simulating spin models in arbitrary dimensions; on the other, the simulation of such models through the formalism of discrete nonlinear maps with the aim of engineering experiments aimed at implementing new calculation paradigms. With the analysis of quantum systems, the project aims to study quantum-mechanical models of oscillators whose mathematical description is mainly based on the formalism of dissipative open systems (Lindblad formalism). In this direction, the aim of the project is mainly to investigate how the presence of quantum mechanical correlations between oscillators influences the computational efficiency of the machine, thus showing a “quantum advantage” compared to the classical counterpart.
The experimental methods are based on the modulation of laser beams using spatial light modulators and on the interaction of light with complex photonic systems. The spatial modulation of the phase of the optical field allows millions of variables to be encoded in a single light spot of a few millimetres. Passing light through photonic materials with controlled properties, such as disorder and nonlinearity, enables parallel processing of this enormous amount of data. The lasers used include continuous and pulsed sources. The detection takes place using high resolution cameras. The photonic materials that support the calculation are polycrystalline structures obtained through bottom-up assembly techniques starting from nanoparticles. The experimental equipment is controlled by classical computers that allow the programming and engineering of photonic machines. In particular, equipment capable of implementing the models and methods developed in the theoretical activity is developed. Particular innovative setups for measuring the phase and amplitude profile allow the effective processing of large datasets.
- Weizmann Institute for Science, Israel (Prof. Nir Davidson)
- Ecole Normale de Paris (Prof. Sylvain Gigan)
- Universitè de Paris (Prof. Cristiano Ciuti)
- University of St. Adrews (Prof. Andrea Di Falco)
- ETH Zurich (Prof. Rachel Grange)
- Institute for Complex Systems, CNR-Italy (Dr.ssa Pilozzi, Dr.ssa Brosco, Dr.ssa Gentilini, Dr. Pierangeli, Dr.ssa Ghofraniha, Dr. Felicetti)
- Claudio Conti, Università Sapienza, 50%,
- Marcello Calvanese Strinati, Ricercatore, CREF, 90%
- Romolo Savo, Ricercatore, CREF, 90%
- Giovanni De Angelis, Tecnologo, CREF, 50 %.
- M. Calvanese Strinati, D. Pierangeli, and C. Conti, “All-optical scalable spatial coherent Ising machine”, Phys. Rev. Applied 16, 054022 (2021)
- M. Calvanese Strinati, F. Vewinger, and C. Conti, “Nonlocality-induced surface localization in Bose-Einstein condensates of light”, Phys. Rev. A 105, 043318 (2022)
- M. Calvanese Strinati and C. Conti, “Multidimensional hyperspin machine”, Nature Commun. 13, 7248 (2022)
- M. Calvanese Strinati and C. Conti, “Hyperscaling in the coherent hyperspin machine”, arXiv:2308.02329 (2023)
- Wang, X. S., Savo, R., Maeder, A., Kaufmann, F., Kellner, J., Morandi, A., Rotter, S. Sapienza, R. and Grange, R. (2023). “Graph model for multiple scattering in lithium niobate on insulator integrated photonic networks”. ArXiv:2306.15483 (2023).
- M. Missori, L. Pilozzi, C. Conti, “Observation of terahertz transition from Fano resonances to bound states in the continuum,” Optics Letters 48, 2381 (2023)
- R. Falcone, C. Conti, “Frame dependence of the nonrelativistic limit of quantum fields,” Phys. Rev. D 107, 085016 (2023)
- D. Pierangeli, C. Conti, “Single-shot polarimetry of vector beams by supervised learning,” Nature Communi. 14. 1831 (2023)
- R. Falcone, C. Conti, “Observing single particles beyond the Rindler horizon,” Phys. Rev. A 107, L030203 (2023)
- R. Falcone, C. Conti, “Nonrelativistic limit of scalar and Dirac fields in curved spacetime,” Phys. Rev. D 107, 045012 (2023)
- C. Conti, “Random walk and non-Gaussianity of the 3D second-quantized Schrödinger–Newton nonlocal soliton,” New Journal of Physics 25, 023026 (2023)
- Carlo M. Valensise, Ivana Grecco, Davide Pierangeli, and Claudio Conti, “Large-scale photonic natural language processing,” Photonics Research 10, 2846 (2022)
- R. Falcone, C. Conti, “Minkowski-Fock states in accelerated frames,” Phys. Rev. D106, 045013 (2022)
- C. Conti, “Variational quantum algorithm for Gaussian discrete solitons and their boson sampling,” Phys. Rev. A 106, 013518 (2022)
- M. Missori, L. Pilozzi, C. Conti, “Terahertz waves dynamic diffusion in 3D printed structures,” Scientific Reports 12, 8613 (2022)
- Zhen-Ting Huang, Kuo-Bin Hong, Ray-Kuang Lee, Laura Pilozzi, Claudio Conti, Jhih-Sheng Wu, Tien-Chang Lu, “Pattern-tunable synthetic gauge fields in topological photonic graphene,” Nanophotonics 11, 1297 (2022)
- C. Conti, E. Del Re, “Photonics and the Nobel Prize in Physics,” Nature Photonics 16, 6 (2022)
- C. Conti, “Training Gaussian boson sampling by quantum machine learning,” Quantum Machine Intelligence 3, 26 (2021)
Via Panisperna 89 A – 00184 Roma
PEC: centrofermi@pec.centrofermi.it
CUU: UF5JTW
Phone: +39 06 4550 2901
VAT: 06431991006
CF: 97214300580
© Centro Studi e Ricerche Enrico Fermi. All rights reserved | Cookie Policy