Detection of Dataset Shifts in Learning-Enabled Cyber-Physical Systems using Variational Autoencoder for Regression

📖 Description

Detection of Dataset Shifts in Learning-Enabled Cyber-Physical Systems using Variational Autoencoder for Regression Feiyang Cai Vanderbilt University Nashville, TN [email removed]Ali I. Ozdagli Vanderbilt University Nashville, TN [email removed]Xenofon Koutsoukos Vanderbilt University Nashville, TN [email removed] Abstract ?Cyber-physical systems (CPSs) use learning-enabled components (LECs) extensively to cope with various complex tasks under high-uncertainty environments. However, the dataset shifts between the training and testing phase may lead the LECs to become ineffective to make large-error predictions, and fur- ther, compromise the safety of the overall system. In our paper, we ?rst provide the formal de?nitions for different types of dataset shifts in learning-enabled CPS. Then, we propose an approach to detect the dataset shifts effectively for regression problems. Our approach is based on the inductive conformal anomaly detection...

📊 Game Impacts

💭 Reactions

🔗 Sources

• [1] https://arxiv.org/abs/2104.06613