logo
calendar30 Dekabr 2025
view12
Main language:Uzbek

MATHEMATICAL MODELING BASED ON ARTIFICIAL INTELLIGENCE: A COMPLETE ANALYSIS OF METHODS OF FORWARD PROPAGATION, LOSS FUNCTION, BACKPROPAGATION AND GRADIENT DESCENT

Field of Science:Artificial Intelligence
pdf

36._Narmanov_O.__Azimov....pdf

PDF

Scientific electronic journal "Digital Economy"

Scientific electronic journal "Digital Economy"

Publish Articlesee more

ARTICLE ANNOTATION

quote
In the field of artificial intelligence and deep learning, neural network training consists of four main stages: forward propagation, calculation of the loss function, backpropagation, and the gradient descent method. This article discusses in detail the mathematical foundations of neural networks, the theoretical and practical aspects of each stage, formulas and algorithms. Mathematical formulas illustrate the processes of data movement from the input layer to the output layer in the process of forward propagation, error estimation using the Loss function, gradient calculation using the backpropagation algorithm and updating weights using the gradient descent method. The article also provides a comparative analysis of various activation functions, Loss functions, and optimization methods.

AUTHORS

N.Otabek

"MUHAMMAD AL-XORAZMIY NOMIDAGI TOSHKENT AXBOROT TEXNOLOGIYALARI UNIVERSITETI" DAVLAT MUASSASASI

A.Murodjon

"MUHAMMAD AL-XORAZMIY NOMIDAGI TOSHKENT AXBOROT TEXNOLOGIYALARI UNIVERSITETI" DAVLAT MUASSASASI

A.Guzal

MIRZO ULUGʻBEK NOMIDAGI OʻZBEKISTON MILLIY UNIVERSITETI

Tags

# artificial intelligence# neural networks# deep learning# mathematical modeling# forward propagation# backpropagation# gradient descent# Loss function# optimization methods

SIMILAR ARTICLES

OTHER ARTICLES IN THIS JOURNAL

Rate Article

0
0 ratings
5
4
3
2
1

Article Identifiers

ROI:

https://eroi.uz/11.142/A026-667343

DOI:

Not Available

References

Burden R. L., Faires J. D., Numerical Analysis, 10th Edition, Brooks/Cole, 2015.

Karimov F. T., “Differensial tenglamalar va ularni hisoblash usullari”, Toshkent: Universitet, 2015.

J. R. Chasnov, “Numerical Methods,” The Hong Kong University of Science and Technology. Hong Kong. pp. 60, 2021.

Alex Lewandowski, Varun Ranganathan ZORB: A Derivative-Free Backpropagation Algorithm for Neural Networks. November 2020. https://www.researchgate.net/publication/346015130_ZORB_A_Derivative-Free_Backpropagation_Algorithm_for_Neural_Networks

Chirag Agarwal, Joe Klobusicky, Don Schonfeld Convergence of backpropagation with momentum for network architectures with skip connections. https://arxiv.org/pdf/1705.07404

A.G. Baydin, B.A. Pearlmutter, D. Syme, F. Wood, P. Torr. Gradients without Backpropagation. https://arxiv.org/abs/2202.08587

G. Kim, Y. Bak. Forward and Backpropagation-Based Artificial Neural Network Modeling Method for Power Conversion System. https://www.mdpi.com/2079-9292/14/23/4718

A.B. Yemberdiyeva A.B. Yemberdiyeva, I.C. Young, M. Symbat, B.M. Samat “Mathematics approach of the backprogation method for constructing artificial neural networks. https://journal.iitu.edu.kz/index.php/ijict/article/view/442

Annette Lopez. Neural Networks: The Backpropagation Algorithm. https://cklixx.people.wm.edu/teaching/math400/Annette-paper.pdf