Obfuscated Malware Detection and Classification in Network Traffic Leveraging Hybrid Large Language Models and Synthetic Data

Sensors, Journal Year: 2025, Volume and Issue: 25(1), P. 202 - 202

Published: Jan. 1, 2025

Android malware detection remains a critical issue for mobile security. Cybercriminals target since it is the most popular smartphone operating system (OS). Malware detection, analysis, and classification have become diverse research areas. This paper presents smart sensing model based on large language models (LLMs) developing classifying network traffic-based malware. The traffic that constantly connects apps may contain harmful components damage these apps. However, one of main challenges in systems analysis scarcity data due to privacy concerns. To overcome this, two-step Syn-detect proposed. first step involves generating synthetic TCP with malicious content using GPT-2. These are then preprocessed used second step, which focuses classification. phase leverages fine-tuned LLM, Bidirectional Encoder Representations from Transformers (BERT), layers. BERT responsible tokenization, word embeddings, was tested two datasets: CIC-AndMal2017 CIC-AAGM2017. achieved an accuracy 99.8% 99.3% Matthew’s Correlation Coefficient (MCC) values predictions were 99% 98% results demonstrate strong performance model. Compared latest classification, outperformed other approaches, delivering promising results.

Language: Английский

A novel metaheuristic-based approach for prediction of corrosion characteristics in offshore pipelines

Engineering Failure Analysis, Journal Year: 2025, Volume and Issue: 170, P. 109231 - 109231

Published: Jan. 5, 2025

Language: Английский

BinOpLeR: Optimization level recovery from binaries based on rich semantic instruction image and weighted voting

Information and Software Technology, Journal Year: 2025, Volume and Issue: unknown, P. 107683 - 107683

Published: Feb. 1, 2025

Language: Английский