QuAC: Quantum fluids of light for All-optical Computing

Project funded by Fundação para a Ciência e a Tecnologia, I.P. (FCT) (Concurso para Projetos de Investigação de caráter Exploratório em Todos os Domínios Científicos 2024), grant 2024.16086.PEX. Funded amount: EUR 59,543.10, as stated by the source. Runs 2026-01-01 to 2027-06-30. Keywords: Optical Computing, Extreme learning machines, Quantum Fluids, Photonics, Engineering and technology. Source: OpenAIRE Graph, from the funder's project data (collected by OpenAIRE on 2025-12-26); national funding data, not CORDIS. Abstract given by the source: The QuAC project targets the development and deployment of optical computing hardware based on the extreme learning machine(ELM) architecture . For that, our solution harnesses the potential of paraxial fluids of light to add a new layer of nonlinear dynamics, expanding beyond the state-of-the-art implementations based on wave interference and enabling truthful all-optical computing with enhanced performance due to the higher dimensionality of the hidden space. The ultimate goal is to demonstrate data processing in the optical domain towards light-speed and energy-efficient information processing while committing to the development of user-centric prototype hardware to be available to external researchers in a transparent manner. To put into context this ambitious project, the all-optical processor is one of the most elusive long-standing goals of the scientific community, promising massively parallel, energy-efficient, and light-speed information processing(McMahon2023;Caulfield et al.2010;Athale et al.2016). However, general-purpose optical computing was long held by obstacles such as light confinement and memory storage necessary under the von Neumann paradigm(Caulfield et al.2010;Athale et al.2016). The recent advent of neuromorphic computing paradigms paved a promising alternative path, by tackling optical computing from both algorithm and hardware level. Along this line, the major highlight is perhaps the deployment of deep neural networks(NN) in the form of the so-called diffractive deep NN(Lin et al.2018;Yan et al.2019;Qian et al.2020) and integrated photonics NN(Zhu et al.2022), proven to perform regression and image classification tasks with impressive throughputs. Yet, these pre-trained solutions usually require the fabrication of diffractive layers which limits versatility to distinct tasks (Zhou et al.2024), and lack nonlinear activation functions. Combined, these strongly bound the applicability as a general-purpose processor for the end-user. In this [shortened; the full text is on the OpenAIRE page]

Original source.

Source: OpenAIRE Graph: projects of national funders, OpenAIRE, from the project data of Fundação para a Ciência e a Tecnologia, I.P. (FCT). Retrieved 2026-10-28. Licence: Creative Commons Attribution 4.0 (OpenAIRE Graph). Selected metadata, adapted presentation; the publisher's record is authoritative. No endorsement by the source is implied. Licence: Creative Commons Attribution 4.0 (CC BY 4.0).

AI-assisted summary: The QuAC project, funded by FCT with a grant of EUR 59,543.10, aims to develop optical computing hardware using extreme learning machines from January 2026 to June 2027. It focuses on utilizing paraxial fluids of light for enhanced all-optical computing performance. Suggested technology pillar: computing. Generated by AI from the public source text; it may contain errors and the original source is authoritative.