Talks and Poster Presentations (with Proceedings-Entry):
M. Meiyi, E. Bartocci, J. Stankovic, L. Feng:
"Predictive monitoring with uncertainty for deep learning enabled smart cities: poster abstract";
Talk: SenSys '20: Proceedings of the 18th Conference on Embedded Networked Sensor Systems,
Virtual Event Japan;
- 2020-11-19; in: "SenSys '20: Proceedings of the 18th Conference on Embedded Networked Sensor Systems",
In order to prevent safety violations, predictive monitoring with uncertainty is crucial for deep learning-enabled services in smart cities. We develop a novel predictive monitoring system for smart city applications, which consists of an RNN-based predictor with uncertainty estimation and a new specification language, named Signal Temporal Logic with Uncertainty. The solution first predicts a sequence of distributions representing city's future states with uncertainty estimation and then checks the predicted results against STL-U specified safety and performance requirements. The system supports decision making by providing a quantitative satisfaction degree with confidence guarantees. We receive promising results from evaluations on two large-scale city datasets, and on a case study on real-time predictive monitoring in a simulated smart city.
"Official" electronic version of the publication (accessed through its Digital Object Identifier - DOI)
Created from the Publication Database of the Vienna University of Technology.