neural networks

Recognition of Inclusion Characteristics Using Neural Network Methods in Stationary Process Modeling

Detection and identification of inclusions in the modeling of stationary processes is a crucial task in many technical fields, including materials science, electronics, and non-destructive testing. The presence of inclusions can affect the mechanical, thermal, and electrical properties of a material, making the accurate determination of their geometric and physical characteristics essential. The use of modern numerical methods and deep learning techniques opens new opportunities for improving the efficiency and accuracy of prediction results.

MODELS FOR TIME SERIES FORECASTING USING ARIMA AND LSTM IN ECONOMICS AND FINANCE

Time series forecasting is a crucial task in economics, business, and finance. Traditionally, forecasting methods such as autoregression (AR), moving average (MA), exponential smoothing (SES), and, most commonly, the autoregressive integrated moving average (ARIMA) model are used. The ARIMA model has demonstrated high accuracy in predicting future time series values. With the advancement of computational power and deep learning algorithms, new approaches to forecasting have emerged.

The role of functional activation in neural networks in the context of financial time series analysis

Nowadays, neural networks are among the most popular analysis tools.  They are effective in solving classification, pattern recognition, and clustering problems.  This paper provides a detailed description and analysis of the operational principles of two neural networks, namely a Siamese network and a multilayer perceptron.  A model for using these neural networks in time series forecasting is proposed.  As an example, a web application was created in which the described neural networks were used to analyze the correlation between pairs of financial assets and assess t

Capabilities and Limitations of Large Language Models

The work is dedicated to the study of large language models (LLMs) and approaches to improving their efficiency in a new service. The rapid development of LLMs based on transformer architecture has opened up new possibilities in natural language processing and the automation of various tasks. However, fully utilizing the potential of these models requires a thorough approach and consideration of numerous factors.

COMPUTER DIAGNOSTIC SYSTEMS: METHODS AND TOOLS

The paper investigates computer diagnostic systems, their architectures, methods, and algorithms used in their work to diagnose cancer, including breast, lung, brain, and other tumors.

Traditional and neural network methods for image segmentation and classification are analyzed and compared, and diagnostic tools in medicine are analyzed.

The key approaches to medical image processing are investigated, in particular, the analysis of segmentation methods based on U-Net networks and classification using convolutional neural networks.

FORECASTING THE ELECTRICITY CONSUMPTION USING AN ENSEMBLE OF MACHINE LEARNING MODELS

The use of machine learning models for electricity consumption prediction for smart grid has been investigated. It was found that data pre-processing can improve the performance of the energy consumption prediction model, while machine learning algorithms can improve model prediction accuracy through the integration of multiple algorithms and hyperparameter optimization. It was found that the ensemble learning method can provide better prediction accuracy than each individual method by combining the strong features of different methods that have different structural characteristics.

Models and Methods for Speech Separation in Digital Systems

The main purpose of the article is to describe state-of-the-art approaches to speech separation and de- monstrate the structures and challenges of building and training such systems. Designing efficient optimized neural network model for speech recognition requires using encoder-decoder model structure with masks estimation flow. The fully-convolutinoal SuDoRM-Rf model demonst- rates the high efficiency with relatively small number of parameters and can be boosted with accelerators, that supports convolutional operations.

Conception of a new quality control method based on neural networks

The prediction of failures in a factory is now an important area of industry that helps to reduce time and cost of non-quality from the data generated from the sensors installed on production lines, this data is used to detect anomalies and predict defects before they occur.  The purpose of this article is to model an intelligent production line capable of predicting various types of non-conforming products.  For that, we will utilize the neural network methodology within the specific context of a production line specialized in juice manufacturing.  Firstly, we introduc

Review of disease identification methods based on computed tomography imagery

Methods and approaches to computational diagnosis of various pulmonary diseases via automated analysis of chest images performed with computed tomography were reviewed. Google Scholar database was searched with several queries focused on deep learning and machine learning chest computed tomography imagery analysis studies published during or after 2017. A collection of 39 papers was collected after screening the search results. The collection was split by publication date into two separate sets based on the date being prior to or after the start of the COVID-19 pandemic.

Алгоритмічна складність задачі навчання двопорогових нейронів

Розглядаються питання, пов’язані з розпізнаванням скінченних множин за допомогою двопорогових нейронних елементів. Показано, що задача навчання ДНЕ є NP-повною. Також наведено умови, виконання яких забезпечує двопороговість булевих функцій, які задаються за допомогою списків рішень.

We study finite set dichotomies on bithreshold neurons. We prove that training a BN is NP-complete task. We also give sufficient conditions ensuring that decision list represents a bithreshold function.