Speeding up quantum perceptron via shortcuts to adiabaticity

Ban, Yue; Chen, Xi; Torrontegui, E.; Solano, E.; Casanova, Jorge

doi:10.1038/s41598-021-85208-3

Cited by 19 publications

(20 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This concept has been proved with implementation on a real quantum computer [27]. Furthermore, the complexity of quantum artificial neurons can be reduced by techniques such as shortcuts to adiabaticity [28]. However, this approach is still in its early stage and current QNNs based on quantum artificial neurons are very difficult to be trained [5], so our evaluation will only consider hardware-efficient QNNs.…”

Section: Related Workmentioning

confidence: 99%

An Evaluation of Hardware-Efficient Quantum Neural Networks for Image Data Classification

et al. 2022

View full text Add to dashboard Cite

Quantum computing is expected to fundamentally change computer systems in the future. Recently, a new research topic of quantum computing is the hybrid quantum–classical approach for machine learning, in which a parameterized quantum circuit, also called quantum neural network (QNN), is optimized by a classical computer. This hybrid approach can have the benefits of both quantum computing and classical machine learning methods. In this early stage, it is of crucial importance to understand the new characteristics of quantum neural networks for different machine learning tasks. In this paper, we will study quantum neural networks for the task of classifying images, which are high-dimensional spatial data. In contrast to previous evaluations of low-dimensional or scalar data, we will investigate the impacts of practical encoding types, circuit depth, bias term, and readout on classification performance on the popular MNIST image dataset. Various interesting findings on learning behaviors of different QNNs are obtained through experimental results. To the best of our knowledge, this is the first work that considers various QNN aspects for image data.

show abstract

Section: Related Workmentioning

confidence: 99%

An Evaluation of Hardware-Efficient Quantum Neural Networks for Image Data Classification

et al. 2022

View full text Add to dashboard Cite

show abstract

“…It can be constructed as a qubit that presents a nonlinear response to an input potential x j in the excitation probability. This can be written as the following quantum gate acting on a jth qubit that encodes the quantum perceptron [44,45]:…”

Section: Quantum Perceptrons With Multi-qubit Potentialsmentioning

confidence: 99%

“…where x j is the potential exerted by other neurons on the perceptron, and the applied external field Ω(t) leads to a tunable energy gap in the dressed-state qubit basis |± , with σx j |± = ±|± . Typically, x j = k i=1 (w ji σz i ) − b j [44,45] which implies that the jth perceptron is coupled to a number k of neurons (labelled with i) in the previous/input layer via standard spin-spin interactions. The Hamiltonian in Eq.…”

Section: Quantum Perceptrons With Multi-qubit Potentialsmentioning

confidence: 99%

See 1 more Smart Citation

Quantum neural networks with multi-qubit potentials

Ban¹,

Torrontegui²,

Casanova³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

We propose quantum neural networks that include multi-qubit interactions in the neural potential leading to a reduction of the network depth without losing approximative power. We show that the presence of multiqubit potentials in the quantum perceptrons enables more efficient information processing tasks such as XOR gate implementation and prime numbers search, while it also provides a depth reduction to construct distinct entangling quantum gates like CNOT, Toffoli, and Fredkin. This simplification in the network architecture paves the way to address the connectivity challenge to scale up a quantum neural network while facilitates its training.

show abstract

“…In this letter, we implement a fully tunable quantum perceptron gate, the fundamental unit for the design of artificial quantum neural networks, in which quantum interactions between qubits give rise to a sigmoidal behaviour. There are various theoretical designs of quantum perceptrons [16,23,24,31,32], of which we follow Ref. [24].…”

mentioning

confidence: 99%

Realization of a quantum perceptron gate with trapped ions

Huber¹,

Haber²,

Barthel³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

We report the implementation of a perceptron quantum gate in an ion-trap quantum computer. In this scheme, a perceptron's target qubit changes its state depending on the interactions with several qubits. The target qubit displays a tunable sigmoid switching behaviour becoming a universal approximator when nested with other percetrons. The procedure consists on the adiabatic rampdown of a dressing-field applied to the target qubit. We also use two successive perceptron quantum gates to implement a XNOR-gate, where the perceptron qubit changes its state only when the parity of two input qubits is even. The applicability can be generalized to higher-dimensional gates as well as the reconstruction of arbitrary bounded continuous functions of the perceptron observables.

show abstract

Speeding up quantum perceptron via shortcuts to adiabaticity

Cited by 19 publications

References 51 publications

An Evaluation of Hardware-Efficient Quantum Neural Networks for Image Data Classification

An Evaluation of Hardware-Efficient Quantum Neural Networks for Image Data Classification

Quantum neural networks with multi-qubit potentials

Realization of a quantum perceptron gate with trapped ions

Contact Info

Product

Resources

About