Introducing an automated and effective cyber-threat detection technique remains a significant challenge in the field of cyber security. This paper aims to address this challenge by presenting an innovative approach that utilizes artificial neural networks and AI technology. Our proposed technique revolutionizes cyber-threat detection by converting a multitude of collected security events into individual event profiles and employing a deep learning-based detection method. To facilitate this research, we have developed an AI-SIEM system that combines event profiling for data pre-processing with various artificial neural network methods, including FCNN, CNN, and LSTM.The primary objective of our system is to distinguish between true positive and false positive alerts, enabling security analysts to respond promptly to cyber threats. To evaluate the effectiveness of our approach, we conducted extensive experiments using two benchmark datasets (NSLKDD and CICIDS2017) as well as two real-world datasets. Through these experiments, we were able to demonstrate the superior performance and accuracy of our AI technique for cyber-threat detection. By introducing this cutting-edge methodology, we aim to contribute to the advancement of cyber security and provide valuable insights for researchers and practitioners in the field. Our innovative approach not only enhances the overall efficiency of cyber-threat detection but also empowers security analysts to combat emerging threats more effectively.
