Elena Solovyeva

Prof. Elena Solovyeva

Saint Petersburg Electrotechnical University "LETI"
orcid

Research Focus & Metrics

Academic Career Length

22Years (2005-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
26 January 2021
Verified
Conference Paper
Closed access
ESElena Solovyeva
Elena Solovyeva
VAVasiliy A. Azarov
Vasiliy A. Azarov
In recent decades, memristive systems, synthesized on the basis of the fourth passive electrical element, the memristor, are developed. A memristor is a controlled resistor with memory, its resistance is nonlinearly dependent on the history of the current change in it. The application of memristor systems is wide, for example, memory devices, neuromorphic systems, neural networks, nanoelectronics devices used for various purposes. This diversity is due to the advantages of memristive systems (compactness, energy efficiency, high performance). Memristors are created using new materials and technologies. As a result, there are difficulties in their analytical and numerical modeling. The main mathematical models of memristors, their advantages and disadvantages, the possibility of implementation in LTspice are represented.

13 September 2022 in Journal of Imaging
Verified
Journal Article
Open access
PDF file
ESElena Solovyeva
Elena Solovyeva
AAAli Abdullah
Ali Abdullah
A dual autoencoder employing separable convolutional layers for image denoising and deblurring is represented. Combining two autoencoders is presented to gain higher accuracy and simultaneously reduce the complexity of neural network parameters by using separable convolutional layers. In the proposed structure of the dual autoencoder, the first autoencoder aims to denoise the image, while the second one aims to enhance the quality of the denoised image. The research includes Gaussian noise (Gaussian blur), Poisson noise, speckle noise, and random impulse noise. The advantages of the proposed neural network are the number reduction in the trainable parameters and the increase in the similarity between the denoised or deblurred image and the original one. The similarity is increased by decreasing the main square error and increasing the structural similarity index. The advantages of a dual autoencoder network with separable convolutional layers are demonstrated by a comparison of the proposed network with a convolutional autoencoder and dual convolutional autoencoder.

1 May 2017
Verified
Conference Paper
Closed access
ESElena Solovyeva
Elena Solovyeva
Recurrent neural networks are represented as non-linear models of dynamic systems. This kind of neural networks is divided into two groups, which are globally and locally recurrent neural networks. Some types are distinguished among globally recurrent networks. The major approximation properties and features of every distinguished type are emphasized. The represented analysis is useful for choosing the neural network structure a priori (prior to its training or constructing the mathematical model of the nonlinear system).

21 December 2018 in Journal of Physics: Conference Series
Verified
Conference Paper
Open access
PDF file
ESElena Solovyeva
Elena Solovyeva
In cases that the mathematical model of a device is complicated or it can not be constructed because of the absence of sufficient information about a researched object, the approach based on the description of unique relationship between the sets of input and output signals is used. The point is the approximation of a non-linear operator, establishing the unique input-output mapping, by mathematical constructions (multidimensional polynomials, regression models, neural networks). Demand to reach the high accuracy of approximation often arises in practice. Recurrent neural networks possessing the properties of dynamics and nonlinearity are considered as mathematical models within the framework of the input-output approach. The non-linear filters synthesis is the constructing of mathematical models using the available sets of input and output signals. Non-linear filters serve for cancelling non-Gaussian noise from distorted signals. As an example, images distorted by the impulse noise is recovered with the help of non-linear filtering performed by different kinds of neural networks.

1 January 2017 in Journal of Physics: Conference Series
Verified
Conference Paper
Open access
PDF file
ESElena Solovyeva
Elena Solovyeva
The synthesis of the non-linear non-recursive digital filter of impulse noise on the basis of the splitting method in time domain is described. The filter recovers speech signals, distorted by impulse noise. The filter model is constructed as the splitting polynomial of an odd degree. The splitter is the time delay line, comprising the equal number of previous and subsequent samples with respect to the current time moment. The polynomial parameters result from solving an approximation problem in the mean-square norm. It is shown that the filter with the splitting model provides more precise speech signal recovery than the median and Volterra filters.

19 October 2021 in Inventions
Verified
Journal Article
Open access
PDF file
ESElena Solovyeva
Elena Solovyeva
AAAli Abdullah
Ali Abdullah
In this paper, the structure of a separable convolutional neural network that consists of an embedding layer, separable convolutional layers, convolutional layer and global average pooling is represented for binary and multiclass text classifications. The advantage of the proposed structure is the absence of multiple fully connected layers, which is used to increase the classification accuracy but raises the computational cost. The combination of low-cost separable convolutional layers and a convolutional layer is proposed to gain high accuracy and, simultaneously, to reduce the complexity of neural classifiers. Advantages are demonstrated at binary and multiclass classifications of written texts by means of the proposed networks under the sigmoid and Softmax activation functions in convolutional layer. At binary and multiclass classifications, the accuracy obtained by separable convolutional neural networks is higher in comparison with some investigated types of recurrent neural networks and fully connected networks.