Create and release your Profile on Zintellect – Postdoctoral applicants must create an account and complete a profile in the on-line application system. Please note: your resume/CV may not exceed 2 pages.
Complete your application – Enter the rest of the information required for the IC Postdoc Program Research Opportunity. The application itself contains detailed instructions for each one of these components: availability, citizenship, transcripts, dissertation abstract, publication and presentation plan, and information about your Research Advisor co-applicant.
Research Topic Description, including Problem Statement:
Computer network defense relies heavily on human operators and analysts but we have failed to monitor, analyze, optimize, or mitigate issues related to physical and cognitive limits that may threaten human and mission safety and productivity. This is a research project to study the potential for increased human resiliency using risk-aware recommender systems that leverage machine power to aid human computer defense activities.
Recommender systems have become increasingly commonplace in everyday life, from movie and music to online dating. These systems use past behavior and other data sources to help the user make more informed, relevant, and timely decisions. Despite large quantities of similar data, however, cyber security and computer network defense have underutilized historical data to inform human decisions.
Consider this example. An analyst in the security operations center is investigating suspicious web requests that may indicate an attack against his corporate webserver. The analyst suspects that blocking the machine generating the suspicious traffic may mitigate the problem. Should he take the action? A recommender system might recommend a different course of action based on historical knowledge about the webserver, suspected attacker, or even world events.
Recommender systems for cyber defense are likely to differ from existing solutions and approaches. Unlike stable and predictable recommender systems like Netflix, cyber defense is incredibly dynamic and must consider both historic and real-time information when recommending a course of action. Another difference is that Netflix is primarily concerned with optimizing the accuracy and relevance of its recommendations. Recommendations in cyber defense must be not only accurate and relevant, but must consider a variety of user-defined or machine-inferred risks of accepting (or rejecting) a recommendation.
Even if new technology can produce a recommender system for network defense, there may still be challenges related to human usability and effectiveness. Research suggests that algorithm avoidance is a powerful detractor, and future work remains which considers human trust and acceptance of recommender systems in cyber security.
Unclassified Example Approaches: