MonkeyGPT: Generative AI in Network Anomaly Detection of Video Conference Applications

• Published in: Proceedings of the ... International Symposium on Parallel and Distributed Processing with Applications (Print) pp. 1273 - 1280
• Main Authors: Xu, Junhao, He, Dongbiao, Zhang, Chen, Zhou, Xu, Ming, Zhongxing, Cui, Laizhong
• Format: Conference Proceeding
• Language: English
• Published: IEEE 30.10.2024
• Subjects: Accuracy; Anomaly detection; Data models; Generative AI; Large language models; Network anomaly detection; Real-time systems; Streaming media; Telecommunication traffic; Text processing; Video Conference Application; Videoconferences
• ISSN: 2158-9208
• DOI: 10.1109/ISPA63168.2024.00171

The rapid advancement of generative artificial intelligence (GAI) has led to the creation of transformative applications such as ChatGPT, which significantly boosts text processing efficiency and diversifies audio, image, and video content. Beyond digital content creation, GAI's capability to analyze complex data distributions holds immense potential for next-generation networks and communications, especially given the swift rise of video conferencing applications (VCAs). This paper presents a dynamic, real-time method for detecting anomalous network links in video conferencing applications. The proposed tool, MonkeyGPT, generates tracing representations of network activity and trains a large language model from scratch to serve as a detection system based on network traffic data. Unlike traditional methods, MonkeyGPT provides an unrestricted search space and does not rely on predefined rules or patterns, enabling it to detect a wider range of anomalies. We demonstrate the effectiveness of MonkeyGPT as an anomaly detection tool in real-world VCAs. The results indicate that the model possesses strong detection capabilities, achieving an accuracy rate of over 97%. It is applicable to various platforms, including Zoom, Microsoft Teams, Tencent Meeting, and Feishu, showcasing its robust adaptability.