Training robust anomaly detection using ML-Enhanced simulations

TL;DR

This paper proposes using neural networks to improve simulation realism, enabling more effective training of anomaly detection systems that can better handle real-world variability and edge cases.

Contribution

The paper introduces a neural network-based method to enhance simulations, making them more realistic for training anomaly detection systems, bridging the gap between simulation and real-world data.

Findings

Enhanced simulations produce more realistic data.

Anomaly detection systems trained on improved simulations perform better in real-world scenarios.

Neural network enhancement increases simulation variability and edge condition coverage.

Abstract

This paper describes the use of neural networks to enhance simulations for subsequent training of anomaly-detection systems. Simulations can provide edge conditions for anomaly detection which may be sparse or non-existent in real-world data. Simulations suffer, however, by producing data that is "too clean" resulting in anomaly detection systems that cannot transition from simulated data to actual conditions. Our approach enhances simulations using neural networks trained on real-world data to create outputs that are more realistic and variable than traditional simulations.