*RE-OPEN* AI for Software System Security

Description

Investiage use of AI planning (and model checking) technologies to find and repair bugs in software/protocols.

Technologies in scope include one of more of:

  • Planning Heuristics:
    • Landmarks and orderings
    • Combine and Abstract (aka Merge and Shrink) 
  • Search Control:
    • Preferred Operators
    • Partial Order Reduction
    • Sequential nogood learning (via Machine Learning if that is consistent with student's interest)
  • Problem Reformulation/Translation:
    • to SAT/CSP
    • Descriptive Quotients
    • Bound-derived Subproblems

 Important to note that the student undertaking this project shall investigate a highly limited subset of the above.

Goals

Learn about advanced topics in AI planning and cyber security.

Use AI planning search to synthesize an exploit.

Stretch goal is to identify a previously unknown exploit. 

Requirements

passion for artificial intelligence, programming, and interest in research.

 

Gain

research experience, and knowledge of advanced AI and cyber topics

Keywords

Planning, Heursitics, Control Knowledge, Cyber Security

Updated:  1 November 2018/Responsible Officer:  Dean, CECS/Page Contact:  CECS Marketing