Deep CNN Based Anomaly Detection in Centralized Metaverse Environment

In the continually expanding area of centralized metaverse environments, safeguarding against digital threats remains a top priority. This paper presents a model based on advanced deep learning techniques, specifically a Convolutional Neural Network (CNN), designed for the purpose of identifying unusual patterns. Our model showcases remarkable performance, achieving a final accuracy of around 94.73% and minimizing the test loss to 0.206631 throughout ten training sessions. In a comparative examination, our deep CNN model surpasses traditional Logistic Regression and a Feedforward Neural Network, underscoring its ability to discern intricate patterns and adapt to the complex dynamics of metaverse data. This research contributes to bolstering security in metaverse platforms, underscoring the pivotal role of deep CNN models in confronting the complexities tied to high-dimensional data.