Time Series Quantum Reservoir Computing with Weak and Projective Measurements

Mujal, Pere; Martínez-Peña, Rodrigo; Giorgi, Gian Luca; Soriano, Miguel C.; Zambrini, Roberta

doi:10.48550/arxiv.2205.06809

Cited by 3 publications

(8 citation statements)

References 78 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[15], were also studied in refs. [16][17][18][19][20][21][22][23][24] under different lenses, while bosonic networks were proposed in ref. [25] in the context of continuous-variable Gaussian states and in refs.…”

Section: Introductionmentioning

confidence: 99%

Benchmarking the Role of Particle Statistics in Quantum Reservoir Computing

et al. 2022

View full text Add to dashboard Cite

Quantum reservoir computing is a neuro-inspired machine learning approach harnessing the rich dynamics of quantum systems to solve temporal tasks. It has gathered attention for its suitability for NISQ devices, for easy and fast trainability, and for potential quantum advantage. Although several types of systems have been proposed as quantum reservoirs, differences arising from particle statistics have not been established yet. In this work, the ability of bosons, fermions, and qubits are assessed and compared to store information from past inputs by measuring linear and nonlinear memory capacity. While, in general, the performance of the system improves with the Hilbert space size, it is shown that also the information spreading capability is a key factor. For the simplest reservoir Hamiltonian choice, and for each boson limited to at most one excitation, fermions provide the best reservoir due to their intrinsic nonlocal properties. On the other hand, a tailored input injection strategy allows the exploitation of the abundance of degrees of freedom of the Hilbert space for bosonic quantum reservoir computing and enhances the computational power compared to both qubits and fermions.

show abstract

Section: Introductionmentioning

confidence: 99%

Benchmarking the Role of Particle Statistics in Quantum Reservoir Computing

et al. 2022

View full text Add to dashboard Cite

show abstract

“…In ideal conditions, photonic platforms for QRC have been predicted to achieve optimal performance, displaying a quantum advantage, in particular the access to an enlarged Hilbert space and operation with low signals [31,50]. Still, to implement successfully and in real-time temporal tasks with a QRC advantage, open challenges need to be overcome [28,29,50,51], namely the limited experimental precision when estimating the output layer, as well as reducing the needed resources, also avoiding the use of external memories. Here, we have proposed an optical platform suitable for real-time QRC based on a physical ensemble of reservoirs, as independent pulses recirculating inside an optical fiber at each input injection.…”

Section: Discussion and Outlookmentioning

confidence: 99%

“…Back-action effects in other quantum substrates would generally affect averages performed over conditional ensembles, so the unconditional evolution is not completely obtained through averaging. In the context of QRC this can negatively affect the performance, as it was shown in the case of qubits [28].…”

Section: Gaussian Measurements On Multipartite Systemsmentioning

confidence: 96%

“…Two strategies to overcome this last issue in qubits platforms have been recently proposed in Ref. [28] using weak measurements and partial sequence repetition (rewinding). Otherwise one needs to restart the protocol for each input injection, buffering the input sequence as in Refs.…”

Section: Photonic Platformmentioning

confidence: 99%

“…Ultimately, RC has been generalized to the quantum regime in order to benefit from the large number of degrees of freedom available in quantum systems [25,26]. In order to experimentally achieve time series processing with superior performance in quantum reservoir computing with respect to classical approaches, several challenges need to be addressed, identifying the most promising applications, efficient platform designs and dealing with quantum measurement retaining quantum advantage [25,27,28].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Scalable photonic platform for real-time quantum reservoir computing

García-Beni¹,

Giorgi²,

Soriano³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Quantum Reservoir Computing (QRC) exploits the information processing capabilities of quantum systems to solve non-trivial temporal tasks, improving over their classical counterparts. Recent progress has shown the potential of QRC exploiting the enlarged Hilbert space, but real-time processing and the achievement of a quantum advantage with efficient use of resources are prominent challenges towards viable experimental realizations. In this work, we propose a photonic platform suitable for real-time QRC based on a physical ensemble of reservoirs in the form of identical optical pulses recirculating through a closed loop. While ideal operation achieves maximum performance, statistical noise is shown to undermine a quantum advantage. We propose a strategy to overcome this limitation and sustain the QRC performance when the size of the system is scaled up. The platform is conceived for experimental implementations to be viable with current technology.

show abstract

Benchmarking the role of particle statistics in Quantum Reservoir Computing

Llodrà,

Charalambous,

Giorgi

et al. 2023

Preprint

View full text Add to dashboard Cite

Quantum reservoir computing is a neuro-inspired machine learning approach harnessing the rich dynamics of quantum systems to solve temporal tasks. It has gathered attention for its suitability for NISQ devices, for easy and fast trainability, and for potential quantum advantage. Although several types of systems have been proposed as quantum reservoirs, differences arising from particle statistics have not been established yet. In this work, we assess and compare the ability of bosons, fermions, and qubits to store information from past inputs by measuring linear and nonlinear memory capacity. While, in general, the performance of the system improves with the Hilbert space size, we show that also the information spreading capability is a key factor. For the simplest reservoir Hamiltonian choice, and for each boson limited to at most one excitation, fermions provide the best reservoir due to their intrinsic nonlocal properties. On the other hand, a tailored input injection strategy allows the exploitation of the abundance of degrees of freedom of the Hilbert space for bosonic quantum reservoir computing and enhances the computational power compared to both qubits and fermions.

show abstract

Time Series Quantum Reservoir Computing with Weak and Projective Measurements

Cited by 3 publications

References 78 publications

Benchmarking the Role of Particle Statistics in Quantum Reservoir Computing

Benchmarking the Role of Particle Statistics in Quantum Reservoir Computing

Scalable photonic platform for real-time quantum reservoir computing

Benchmarking the role of particle statistics in Quantum Reservoir Computing

Contact Info

Product

Resources

About