Metaheuristic design of feedforward neural networks

In recent years, deep learning has had a profound impact on machine learning and artificial intelligence. At the same time, algorithms for quantum computers have been shown to efficiently solve some problems that are intractable on conventional, classical computers. We show that quantum computing not only reduces the time required to train a deep restricted Boltzmann machine, but also provides a richer and more comprehensive framework for deep learning than classical computing and leads to significant improvements in the optimization of the underlying objective function. Our quantum methods also permit efficient training of multilayer and fully connected models.

/pdf/quantum-deep-learning-22lskrh8pw.pdf

Quantum deep learning

Deep learning achieves unprecedented success involves many fields, whereas the high requirement of memory and time efficiency tolerance have been the intractable challenges for a long time On the other hand, quantum computing shows its superiorities in some computation problems owing to its intrinsic properties of superposition and entanglement, which may provide a new path to settle these issues In this paper, a quantum deep convolutional neural network (QDCNN) model based on the quantum parameterized circuit for image recognition is investigated In analogy to the classical deep convolutional neural network (DCNN), the architecture that a sequence of quantum convolutional layers followed by a quantum classified layer is illustrated Inspired by the variational quantum algorithms, a quantum–classical hybrid training scheme is demonstrated for the parameter updating in the QDCNN The network complexity analysis indicates the proposed model provides the exponential acceleration comparing with the classical counterpart Furthermore, the MNIST and GTSRB datasets are employed to numerical simulation and the quantitative experimental results verify the feasibility and validity

A quantum deep convolutional neural network for image recognition

Quantum perceptron over a field and neural network architecture selection in a quantum computer

This work aims to provide a broad overview of computational techniques belonging to the area of artificial intelligence tailored for identification of nonlinear dynamical systems. Both parametric and nonparametric identification problems are considered. The examined computational intelligence techniques for parametric identification deal with genetic algorithm, particle swarm optimization, and differential evolution. Special attention is paid to the parameters estimation for a rich class of nonlinear dynamical models, including the Bouc–Wen model, chaotic systems, the Jiles–Atherton model, the LuGre model, the Prandtl–Ishlinskii model, the Preisach model, and the Wiener–Hammerstein model. On the other hand, genetic programming and artificial neural networks are discussed for nonparametric identification applications. Once the identification problem is formulated, a detailed illustration of the considered computational intelligence techniques is provided, together with a comprehensive examination of relevant applications in the fields of structural mechanics and engineering. Possible directions for future research are also addressed.

A review on computational intelligence for identification of nonlinear dynamical systems

A hybrid quantum-inspired neural networks with sequence inputs

In order to solve the problem of embedding the watermark into the quantum color image, in this paper, an improved scheme of using small-scale quantum circuits and color scrambling is proposed. Both color carrier image and color watermark image are represented using novel enhanced quantum representation. The image sizes for carrier and watermark are assumed to be $$2^{n+1}\times 2^{n+2}$$2n+1×2n+2 and $$2^{n}\times 2^{n}$$2n×2n, respectively. At first, the color of pixels in watermark image is scrambled using the controlled rotation gates, and then, the scrambled watermark with $$2^n\times 2^n$$2n×2n image size and 24-qubit gray scale is expanded to an image with $$2^{n+1}\times 2^{n+2}$$2n+1×2n+2 image size and 3-qubit gray scale. Finally, the expanded watermark image is embedded into the carrier image by the controlled-NOT gates. The extraction of watermark is the reverse process of embedding it into carrier image, which is achieved by applying operations in the reverse order. Simulation-based experimental results show that the proposed scheme is superior to other similar algorithms in terms of three items, visual quality, scrambling effect of watermark image, and noise resistibility.

An improved quantum watermarking scheme using small-scale quantum circuits and color scrambling

Spatial filtering is one principal tool used in image processing for a broad spectrum of applications. Median filtering has become a prominent representation of spatial filtering because its performance in noise reduction is excellent. Although filtering of quantum images in the frequency domain has been described in the literature, and there is a one-to-one correspondence between linear spatial filters and filters in the frequency domain, median filtering is a nonlinear process that cannot be achieved in the frequency domain. We therefore investigated the spatial filtering of quantum image, focusing on the design method of the quantum median filter and applications in image de-noising. To this end, first, we presented the quantum circuits for three basic modules (i.e., Cycle Shift, Comparator, and Swap), and then, we design two composite modules (i.e., Sort and Median Calculation). We next constructed a complete quantum circuit that implements the median filtering task and present the results of several simulation experiments on some grayscale images with different noise patterns. Although experimental results show that the proposed scheme has almost the same noise suppression capacity as its classical counterpart, the complexity analysis shows that the proposed scheme can reduce the computational complexity of the classical median filter from the exponential function of image size n to the second-order polynomial function of image size n, so that the classical method can be speeded up.

Quantum image median filtering in the spatial domain

This paper presented a quantum representation and some basic operations for one-dimensional finite length digital signal. For the representation, similar to the existing flexible representation of quantum audio (FRQA), two sets of qubits are used to represent the amplitudes and positions of the signal, respectively, wherein the amplitude is represented by an (n + 1)-bit signed number in twos complement, including one sign bit, m integer bits and n − m fractional bits. The basic operations include the addition, subtraction, multiplication, division, and circular convolution of two digital signals. We designed the quantum circuits that implements these basic operations, analyzed the computational complexity, and verified the correctness of these circuits by simulations on classical computers.

Quantum Representation and Basic Operations of Digital Signals

Process neural networks (PNN) can only receive time-varying continuous functions, can not receive discrete samples. To solve this problem, a training algorithm of PNN based on piecewise linear interpolation function is proposed. First the discrete data of both sample functions and weight functions are transformed to piecewise linear functions,and then the integrals of product of two linear functions at a given sampling interval are computed. As a result of aggregation,these integrals are submitted to process neurons of PNN hide layer.Finally,the networks output is obtained in output layer. Some advantages of piecewise linear interpolation function,e.g. continuity, calculability, lower exponential and less parameters, simplifies aggregation operation of PNN in both space and time. The experimental results are illustrated the availability of the proposed method.

Hong Xiao

Papers

A hybrid quantum-inspired neural networks with sequence inputs

An improved quantum watermarking scheme using small-scale quantum circuits and color scrambling

Quantum image median filtering in the spatial domain

Quantum Representation and Basic Operations of Digital Signals

Process Neural Network Algorithm Based on Piecewise Linear Interpolation Function