New Approaches in the Learning of Complex-Valued Neural Networks

Kotsovsky, Vladyslav; Batyuk, Anatoliy; Yurchenko, Maksym

doi:10.1109/dsmp47368.2020.9204332

Cited by 13 publications

(5 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The General Regression Neural Network (GRNN) is currently one of the best ANNs for working with short tabular datasets. This is explained primarily by the highest generalization properties of this neural network type compared to all other existing architectures [23]. In addition, due to the lack of a training algorithm, as well as the ease of implementation of this ANN, it is often used to solve various applied tasks [24].…”

Section: Grnnmentioning

confidence: 99%

An improved ANN-based global-local approximation for small medical data analysis

Izonin,

Tkachenko,

Bliakhar

et al. 2023

EAI Endorsed Trans Perv Health Tech

View full text Add to dashboard Cite

INTRODUCTION: The task of approximation of complex nonlinear dependencies, especially in the case of short datasets, is important in various applied fields of medicine. Global approximation methods describe the generalized behavior of the model, while local methods explain the behavior of the model at specific data points. Global-local approximation combines both approaches, which makes such methods a powerful tool for processing short sets of medical data that can have both broad trends and local variations.OBJECTIVES: This paper aims to improve the method of sequential obtaining global and local components of the response surface to increase the accuracy of prediction in the case of short sets of medical data.METHODS: In this paper, the authors developed a new method that combined two ANNs: a non-iterative SGTM neural-like structure for obtaining the global component and GRNN as a powerful tool of local approximation in the case of short datasets.RESULTS: The authors have improved the method of global-local approximation due to the use of a General Regression Neural Network instead of RBF ANN for obtaining the local component, which ensured an increase in the accuracy of the body fat prediction task. The authors optimized the operation of the method and investigated the efficiency of the sequential obtaining global and local components of the response surface in comparison with the efficiency using a number of existing methods.CONCLUSION: The conducted experimental studies for solving the body fat prediction task showed the high efficiency of using the improved method in comparison with a number of existing methods, including ensemble methods.

show abstract

Section: Grnnmentioning

confidence: 99%

An improved ANN-based global-local approximation for small medical data analysis

Izonin,

Tkachenko,

Bliakhar

et al. 2023

EAI Endorsed Trans Perv Health Tech

View full text Add to dashboard Cite

show abstract

“…The zoom level is defined by the number of items which can be displayed on the view without scrolling. The software developed can be used during various data mining tasks [1,3,14,15,19].…”

Section: Principal Component Analysismentioning

confidence: 99%

An Integral Software Solution of the SGTM Neural-Like Structures Implementation for Solving Different Data Mining Tasks

Tkachenko

2021

Lecture Notes on Data Engineering and Communications Technologies

View full text Add to dashboard Cite

The paper presents a developed software solution that implements a new learning model and application of artificial neural networks, i.e. the Successive Geometric Transformations Model, to solve various applied data mining tasks. This model is applied to construct Feed Forward Neural Networks, which anticipates the rejection of training goal interpretation as a multi-extremal optimization procedure. The modes of use of the system developed (prediction, time series, and cascade modes) are described in detail and illustrated, as well as the structure and functions of separate constituents of the graphical user interface; data selection and download procedures; neuro-like structure parameterization capabilities for each different mode of operation of the program. Simulation of the developed integral software solution was performed while fulfilling prediction, classification, and time series forecasting. Besides, the feasibility of performing nonlinear PCA analysis at high speed is shown. When performing all the above tasks, the program outcomes are presented in both numerical form and using the appropriate one for a particular mode of visualization in 2D or 3D formats.

show abstract

“…In order to address these typical problems on nonlinear activations in the complex-valued domain, many scholars started to implement partial derivatives on the real and imaginary parts respectively considering the Cauchy-Riemann condition but not to implement on these fundamental activations allowing for the complex-valued entities [23], such as Sigmoid, Tanh, Swish, and also the amplitude-phase split-type activations were applied to produce a very complicate format [24], these works do not use the Cauchy-Riemann condition. Recently, holomorphic functions were implemented to induce a formulation of CVBP [25], and also we unveiled a compatible condition [13] to encapsulate the fundamental activations available in the CVBP.…”

Section: Introductionmentioning

confidence: 99%

Unitary learning in conditional modes for deep optics neural networks

Zhang

Xiao

Zhong

2023

Conference on Infrared, Millimeter, Terahertz Waves and Applications (IMT2022)

View full text Add to dashboard Cite

Complex-valued backprogation (CVBP) is an essential algorithm parsing a large amount of complex data for statistical inferences to deep optics neural networks. However, the operating optical transformations are almost unitary, which require the unitary backpropagation algorithms to meet the hetero-equivalence from their mathematical to optical mechanisms. This paper mainly compares two compact formulations for a common CVBP and its unitary-variants under holomorphic and compatible conditions, for which composite mechanisms are invented. For unitary weights-update, essentially, the Riemannian gradient of a CVBP is expanded from Euclidean gradient by executing an exponential mapping onto its unitary manifold. The convergent speed and accuracy, as well as the evolutions of their energies and averaging-phases at each layer are progressively investigated and compared, which reflect the dynamic behaviors on the network convergences. We find that not all the nonlinear activations under the respective conditions are always ensured to convergences, and the averaging-phases still remain several jumps although the energies at current layer situate at the convergent districts.

show abstract

New Approaches in the Learning of Complex-Valued Neural Networks

Cited by 13 publications

References 19 publications

An improved ANN-based global-local approximation for small medical data analysis

An improved ANN-based global-local approximation for small medical data analysis

An Integral Software Solution of the SGTM Neural-Like Structures Implementation for Solving Different Data Mining Tasks

Unitary learning in conditional modes for deep optics neural networks

Contact Info

Product

Resources

About