Autonomous Trustworthy Computing Platforms and Devices
The goal of this project is to develop novel, autonomous system health management techniques for Trustworthy Computing Platforms and Devices. The goal of these techniques is automatic diagnosis and re-configuration when hardware or software fails to work as anticipated. Bayesian networks, an approach to model multi-variate probability distributions, will be a key component in our approach, where initially we plan to investigate software bugs. The expected outcome of the research is improved self-diagnosing, self-configuring, and self-managing capabilities of Trustworthy Computing Platforms and Devices. The work will be based on our previous research on Bayesian networks, in particular the ProADAPT diagnostic system for an Electrical Power System, which recently had the highest scores the Industrial Track of DXC-09, visit the DXC-09 website for details. The project is supported by Carnegie Mellon CyLab.
Contact: Ole Mengshoel