Techniques described herein provide for a Dynamic Latent Space Adaptation model that allows a Variational Autoencoder (VAE) to adjust its latent space dimensions in response to evolving complexities in input data. The model is initially trained on a comprehensive dataset to establish a baseline, continuously monitors incoming syslog data, and adjusts its latent space dimensions based on a complexity analysis within static and variable parameters. The real-time adaptation enables the model to maintain acute sensitivity to new and unusual data patterns, thereby improving anomaly detection.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.