Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 2
Encontrada(s) 1 página(s)

DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Quantum active learning

  • Ding, Y.
  • Ban, Yue
  • Sanz, M.
  • Martín-Guerrero, José D.
  • Chen, Xi
[Data Availability] We utilized the quantum platform mindspore quantum [64] in our experiments. All codes, data sets, and results are provided in this repository [65] for reproducibility of this study. Credit for the tic-tac-toe data set generation code is attributed to pennylane on their web page [66]., Quantum machine learning (QML), as an extension of classical machine learning that harnesses quantum mechanics, facilitates efficient learning from data encoded in quantum states. Training a quantum neural network (QNN) typically demands a substantial labeled training set for supervised learning. Human annotators, often experts, provide labels for samples through additional experiments, adding to the training cost. To mitigate this expense, there is a quest for methods that maintain model performance over fully labeled data sets while requiring fewer labeled samples in practice, thereby extending few-shot learning to the quantum realm. Quantum active learning (QAL) estimates the uncertainty of quantum data to select the most informative samples from a pool for labeling. Consequently, a QML model is supposed to accumulate maximal knowledge as the training set comprises labeled samples selected via sampling strategies. Notably, the QML models trained within the QAL framework are not restricted to specific types, allowing performance enhancement from the model architecture's perspective towards few-shot learning. Recognizing symmetry as a fundamental concept in physics ubiquitous across various domains, we leverage the symmetry inherent in quantum states induced by the embedding of classical data for model design. We employ an equivariant QNN capable of generalizing from fewer data with geometric priors. We benchmark the performance of QAL on two classification problems, observing both positive and negative results. QAL effectively trains the model, achieving performance comparable to that on fully labeled data sets by labeling less than 7% of the samples in the pool with unbiased sampling behavior. Furthermore, we elucidate the negative result of QAL being overtaken by random sampling baseline through miscellaneous numerical experiments. Our work lays the groundwork for real-world applications of QML by addressing realistic experimental costs., This work is supported by NSFC (12075145 and 12211540002), STCSM (2019SHZDZX01-ZX04), the Innovation Program for Quantum Science and Technology (2021ZD0302302), EU FET Open Grant EPIQUS (899368), HORIZON-CL4-2022-QUANTUM-01-SGA Project No. 101113946 OpenSuperQPlus100 of the EU Flagship on Quantum Technologies, the Basque Government through Grant No. IT1470-22, the Project Grants No. PID2021-126273NB-I00 and No. PID2021-125823NA-I00 funded by MCIN/AEI/10.13039/501100011033, by “ERDFA way of making Europe,” “ERDF Invest in your Future,” Nanoscale NMR and complex systems (PID2021-126694NB-C21, PID2021-126694NA-C22), the Valencian Government Grant with Reference No. CIAICO/2021/184, the Spanish Ministry of Economic Affairs and Digital Transformation through the QUANTUM ENIA project call–Quantum Spain project, and the European Union through the Recovery, Transformation and Resilience Plan–NextGenerationEU within the framework of the Digital Spain 2026 Agenda. The authors also acknowledge the financial support received from the IKUR Strategy under the collaboration agreement between Ikerbasque Foundation and BCAM on behalf of the Department of Education of the Basque Government. M.S. acknowledges “Ayudas para contratos Ramón y Cajal” 2015-2020 (RYC-2020-030503-I). Y.B. acknowledges the Ramón y Cajal (RYC2023-042699-I) research fellowship. Y.D. thanks the European Commission for a Marie Curie PF Grant (No. 101204580 FELQO). Y.D. acknowledges “CPS-MindSpore funding for quantum machine learning” with the support from the technical team of MindSpore Quantum., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Superoscillating quantum control induced by sequential selections

  • Ding, Y.
  • Pan, Yiming
  • Chen, Xi
Superoscillation is a counterintuitive phenomenon for its mathematical feature of faster than Fourier, which has allowed novel optical imaging beyond the diffraction limit. In this article we introduce a superoscillating quantum control protocol realized by sequential selections within the framework of weak measurement, which steers the apparatus (target) by repeatedly applying optimal pre- and postselections to the system (controller). Our protocol accelerates the adiabatic transport of trapped ions and the adiabatic quantum search algorithm at a finite energy cost. We demonstrate the accuracy and robustness of the protocol in the presence of decoherence and fluctuating noise and elucidate the trade-off between fidelity and rounds of selections. Our findings open avenues for quantum-state manipulation and wave-packet control using superoscillation across a range of quantum platforms., This work was supported by the Basque Government through Grant No. IT1470-22, Project Grants No. PID2021-126273NBI00 and No. PID2021-123131NA-I00 funded by MCIN/AEI/10.13039/501100011033, by the ERDF A way of making Europe, and by the European Union Next Generation EU/PRTR. Y.D. thanks the European Commission for Marie Curie postdoctoral Grant No. 101204580 (FELQO)., Peer reviewed