Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userMultivariable functions approximation using a single quantum neuron
54
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
We describe a model element able to perform universal stochastic approximations of continuous multivariable functions in both neuron-like and quantum form. The implementation of this model in the form of a multi-barrier, multiple-slit system is proposed and it is demonstrated that this single neuron-like model is able to perform the XOR function unrealizable with single classical neuron. For the simplified waveguide variant of this model it is proved for different interfering quantum alternatives with no correlated adjustable parameters, that the system can approximate any continuous function of many variables. This theorem is applied to the 2-input quantum neural model based on the use of the schemes developed for controlled nonlinear multiphoton absorption of light by quantum systems. The relation between the field of quantum neural computing and quantum control is discussed.
rate research
Read More
We present a method to reduce the variance of stochastic trace estimators used in quantum typicality (QT) methods via a randomized low-rank approximation of the finite-temperature density matrix $e^{-beta H}$. The trace can be evaluated with higher accuracy in the low-rank subspace while using the QT estimator to approximate the trace in the complementary subspace. We present two variants of the trace estimator and demonstrate their efficacy using numerical experiments. The experiments show that the low-rank approximation outperforms the standard QT trace estimator for moderate- to low-temperature. We argue this is due to the low-rank approximation accurately represent the density matrix at low temperatures, allowing for accurate results for the trace.
We study reduction schemes for functions of many variables into system of functions in one variable. Our setting includes infinite-dimensions. Following Cybenko-Kolmogorov, the outline for our results is as follows: We present explicit reductions schemes for multivariable problems, covering both a finite, and an infinite, number of variables. Starting with functions in many variables, we offer constructive reductions into superposition, with component terms, that make use of only functions in one variable, and specified choices of coordinate directions. Our proofs are transform based, using explicit transforms, Fourier and Radon; as well as multivariable Shannon interpolation.
With the constant increase of the number of quantum bits (qubits) in the actual quantum computers, implementing and accelerating the prevalent deep learning on quantum computers are becoming possible. Along with this trend, there emerge quantum neural architectures based on different designs of quantum neurons. A fundamental question in quantum deep learning arises: what is the best quantum neural architecture? Inspired by the design of neural architectures for classical computing which typically employs multiple types of neurons, this paper makes the very first attempt to mix quantum neuron designs to build quantum neural architectures. We observe that the existing quantum neuron designs may be quite different but complementary, such as neurons from variation quantum circuits (VQC) and Quantumflow. More specifically, VQC can apply real-valued weights but suffer from being extended to multiple layers, while QuantumFlow can build a multi-layer network efficiently, but is limited to use binary weights. To take their respective advantages, we propose to mix them together and figure out a way to connect them seamlessly without additional costly measurement. We further investigate the design principles to mix quantum neurons, which can provide guidance for quantum neural architecture exploration in the future. Experimental results demonstrate that the identified quantum neural architectures with mixed quantum neurons can achieve 90.62% of accuracy on the MNIST dataset, compared with 52.77% and 69.92% on the VQC and QuantumFlow, respectively.
We address the validity of the single-mode approximation that is commonly invoked in the analysis of entanglement in non-inertial frames and in other relativistic quantum information scenarios. We show that the single-mode approximation is not valid for arbitrary states, finding corrections to previous studies beyond such approximation in the bosonic and fermionic cases. We also exhibit a class of wave packets for which the single-mode approximation is justified subject to the peaking constraints set by an appropriate Fourier transform.
We perform experimental quantum polarimetry using a heralded single photon to analyze the optical activity of linearly polarized light traversing a chiral medium. Three kinds of estimators are considered to estimate the concentrations of sucrose solutions from measuring the rotation angle of the linear polarization of the output photons. Through repetition of independent and identical measurements performed for each individual scheme and different concentration sucrose solutions, we compare the estimation uncertainty among the three schemes. The results are also compared to classical benchmarks for which a coherent state of light is taken into account. The quantum enhancement in the estimation uncertainty is evaluated and the impact of experimental and technical imperfections is discussed. In this work, we lay out a route for future applications relying on quantum polarimetry.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
