Daniel Varro will give a talk about his work on Automated Graph Generation for the Assurance of Software-intensive Cyber-Physical Systems. Details below.
Daniel Varro, Prof., MCGill.
Automated Graph Generation for the Assurance of Software-intensive Cyber-Physical Systems
When engineering critical software-intensive cyber-physical systems, system-level assurance often uses a combination of design-time and runtime verification and validation techniques to ensure compliance with related safety standards. For example, software tool qualification necessitated by safety standards requires that no tools used in the design of avionics systems are trusted without sufficient assurance ideally provided by a large set of consistent (well-formed or malformed) instance models specific to a domain. System-level testing of autonomous vehicles also frequently relies upon the automated synthesis of realistic test contexts and abstract test scenarios. The testing of machine learning based components used in may exploit consistent synthetic scene graphs. My talk will focus on how the automated generation of consistent, realistic and diverse domain-specific graph models can help in design-time assurance and runtime monitoring of critical cyber-physical systems. The talk will cover recent research results including conceptual foundations and open-source tool support in the Refinery framework.