This article analyzes two methods of adversarial attacks created against machine — training — based pest program detection systems-JSMA (Jacobian Saliency Map Attack) and Carlini & Wagner (C&W) - in a practical context and proposes methods to adapt them to convolutional (CNN) as well as sequence-based (RNN/LSTM) classifiers. The goal of the research is to identify the weaknesses of existing detectors, improve algorithms for generating attack scenarios, and develop more robust protection mechanisms based on this. The methodology section mathematically represents the process of creating adversarial samples, and the specific modifications of the JSMA, C&W algorithms are described in detail.
| Mualliflar | Haydarov, Elshod |
|---|---|
| Jurnal | Al-Farg'oniy avlodlari |
| Nashr sanasi | 2026-03-27 |
| Son | 1 |
| Betlar | 364-370 |
| Til | Rus |
состязательные атаки, JSMA, Carlini & Wagner, CNN, RNN/LSTM, состязательное обучение, обнаружение вредоносных программ., adversarial hujumlar, JSMA, Carlini & Wagner, CNN, RNN/LSTM, adversarial trening, zararkunanda dasturlar deteksiyasi., adversarial attacks, JSMA, Carlini & Wagner, CNN, RNN/LSTM, adversarial training, malware detection.
This article presents a study of a dictionary-based pre-filter for safe educational content generation in a school environment. An Android application based on the MVVM architecture was developed as an experimental…
This paper proposes a speech recognition system for automatically recognizing separately pronounced Uzbek words. The system uses MFCC for acoustic feature extraction and a Gaussian Hidden Markov Model for word modeling…
This article analyzes current trends in the field of cybersecurity. Today, the stable and protected operation of data transmission networks, computer systems, and mobile devices has become a fundamental condition…
In this work, methods and algorithms for improving the quality of a given image are studied. Methods for reducing noise, improving contrast, increasing sharpness, and clearly displaying image elements during image…
The scarcity of labeled data remains a fundamental limitation of supervised learning models. This study proposes an autoencoder-based unsupervised approach for detecting anomalies in network traffic. Experiments…
This article describes the research conducted by training the neural network model proposed by NVIDIA in speech recognition with a speech dataset of the Uzbek language. Currently, various models of neural networks are…
This article analyzes the effectiveness of malware detection methods using artificial intelligence technologies. With the rapid development of modern information and communication technologies, the number of malicious…
While large-scale pre-trained models have significantly advanced multilingual Automatic Speech Recognition (ASR), many low-resource languages remain under-served due to the scarcity of high-quality annotated speech…
This study presents a comprehensive simulation analysis of germanium (Ge) diffusion into silicon (Si) substrates with a resistivity of 100 Ω·cm (n-type) in the temperature range of 800–1200 °C using the Synopsys…
This article studies the effectiveness of modern deep learning models for the task of deep contextual analysis of text data in social networks. As part of the study, the ability of RNN, LSTM and DistilBERT models based…