English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Cyber-physical systems (CPSs) use learning-enabled components (LECs)extensively to cope with various complex tasks under high-uncertaintyenvironments. However, the dataset shifts between the training and testingphase may lead the LECs to become ineffective to make large-error predictions,and further, compromise the safety of the overall system. In our paper, wefirst provide the formal definitions for different types of dataset shifts inlearning-enabled CPS. Then, we propose an approach to detect the dataset shiftseffectively for regression problems. Our approach is based on the inductiveconformal anomaly detection and utilizes a variational autoencoder forregression model which enables the approach to take into consideration both LECinput and output for detecting dataset shifts. Additionally, in order toimprove the robustness of detection, layer-wise relevance propagation (LRP) isincorporated into our approach. We demonstrate our approach by using anadvanced emergency braking system implemented in an open-source simulator forself-driving cars. The evaluation results show that our approach can detectdifferent types of dataset shifts with a small number of false alarms while theexecution time is smaller than the sampling period of the system.

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