SentraAI: AI-Based Anomaly Detection with LogBERT

Created using ChatSlide

The SentraAI presentation outlines a robust approach to cyber attack mitigation using the LogBERT model, emphasizing unsupervised learning efficiency. Research in anomaly detection highlights comparisons between classical and AI-driven methodologies, showcasing technological advancements such as Kitsune and DeepLog. Methodologically, the preprocessing of HDFS logs and strategic use of LogBERT's pretraining and LoRA integration are key, with a focus on optimizing the F1 threshold. The dataset...