МЕТОД ВИЯВЛЕННЯ ОБ’ЄКТІВ ПРОПАГАНДИ НЕЙРОМЕРЕЖЕВИМИ МОДЕЛЯМИ ГЛИБОКОГО НАВЧАННЯ З ВІЗУАЛЬНОЮ ІНТЕРПРЕТАЦІЄЮ ПРИЙНЯТИХ РІШЕНЬ

The present work proposes the method of identifying propaganda objects by neural network models of deep learning with visual interpretation of the decisions, which differs from the existing ones by expanding the plural of propaganda objects by adding variants of their verbal submissions and using context windows to identify connections between the used methods and objects of propaganda, which made it possible to improve the results of detecting objects of propaganda and visually interpret them. The main stages of implementing the method of identifying propaganda objects by neural network models of deep learning with visual interpretation of the decisions made are the formation of a plural of propaganda objects by recognizing named entities, pre-processing of the text and expansion of the plural of propaganda objects by adding variants of their verbal representations, building contextual windows for verbal presentations of each object of propaganda, taking into account the threshold value of the minimum window size, assessment of the level of used propaganda methods within context windows according to neural network models, and building the plural of important connections between methods and objects of propaganda, taking into account the threshold value of the minimum level of ostent of propaganda. The effectiveness of using the proposed the method of identifying propaganda objects by neural network models of deep learning with visual interpretation of the decisions has been experimentally proven, which allows, unlike existing analogues, in addition to searching for NER using the "STANZA" neural network library, also expanding the plural of available propaganda objects in texts using the "FastText" machine learning library, as well as issuing an assessment by which the objects found are correlated with the propaganda methods used. For ease of use, the visual interpretation of the found objects of propaganda is also provided, which allows to visually observe the objects of influence within the framework of the used propaganda methods.

Authors

Publication Details

Journal
Herald of Khmelnytskyi National University Technical sciences
Published
2024-11-28
DOI
https://doi.org/10.31891/2307-5732-2024-343-6-27
Primary Topic
Scientific Research and Philosophical Inquiry
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

МЕТОД ВИЯВЛЕННЯ ОБ’ЄКТІВ ПРОПАГАНДИ НЕЙРОМЕРЕЖЕВИМИ МОДЕЛЯМИ ГЛИБОКОГО НАВЧАННЯ З ВІЗУАЛЬНОЮ ІНТЕРПРЕТАЦІЄЮ ПРИЙНЯТИХ РІШЕНЬ

МАРИНА МОЛЧАНОВА
Herald of Khmelnytskyi National University Technical sciences
Scientific Research and Philosophical Inquiry
article

МЕТОД ВИЯВЛЕННЯ ОБ’ЄКТІВ ПРОПАГАНДИ НЕЙРОМЕРЕЖЕВИМИ МОДЕЛЯМИ ГЛИБОКОГО НАВЧАННЯ З ВІЗУАЛЬНОЮ ІНТЕРПРЕТАЦІЄЮ ПРИЙНЯТИХ РІШЕНЬ

МАРИНА МОЛЧАНОВА
article en

Abstract

The present work proposes the method of identifying propaganda objects by neural network models of deep learning with visual interpretation of the decisions, which differs from the existing ones by expanding the plural of propaganda objects by adding variants of their verbal submissions and using context windows to identify connections between the used methods and objects of propaganda, which made it possible to improve the results of detecting objects of propaganda and visually interpret them. The main stages of implementing the method of identifying propaganda objects by neural network models of deep learning with visual interpretation of the decisions made are the formation of a plural of propaganda objects by recognizing named entities, pre-processing of the text and expansion of the plural of propaganda objects by adding variants of their verbal representations, building contextual windows for verbal presentations of each object of propaganda, taking into account the threshold value of the minimum window size, assessment of the level of used propaganda methods within context windows according to neural network models, and building the plural of important connections between methods and objects of propaganda, taking into account the threshold value of the minimum level of ostent of propaganda. The effectiveness of using the proposed the method of identifying propaganda objects by neural network models of deep learning with visual interpretation of the decisions has been experimentally proven, which allows, unlike existing analogues, in addition to searching for NER using the "STANZA" neural network library, also expanding the plural of available propaganda objects in texts using the "FastText" machine learning library, as well as issuing an assessment by which the objects found are correlated with the propaganda methods used. For ease of use, the visual interpretation of the found objects of propaganda is also provided, which allows to visually observe the objects of influence within the framework of the used propaganda methods.

Herald of Khmelnytskyi National University Technical sciencesVol. 343(6(1))
Peace, Justice and strong institutions
Openalex Percentile: Top 61%
Scientific Research and Philosophical Inquiry
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

МЕТОД ВИЯВЛЕННЯ ОБ’ЄКТІВ ПРОПАГАНДИ НЕЙРОМЕРЕЖЕВИМИ МОДЕЛЯМИ ГЛИБОКОГО НАВЧАННЯ З ВІЗУАЛЬНОЮ ІНТЕРПРЕТАЦІЄЮ ПРИЙНЯТИХ РІШЕНЬ — МАРИНА МОЛЧАНОВА · Herald of Khmelnytskyi National University Technical sciences (2024) | TGRS Research Map | TGRS