Patent · US Active

Cyber auto tactics techniques and procedures multiple hypothesis engine

US8583583B1 · kind B1 · utility

5Cited by
1References
3Claims
0Family size

Assignee

Inventors

Key dates

Filing dateSep 16, 2010
Grant dateNov 12, 2013
Priority date
Expiry dateDec 8, 2031

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG06F2221/2145
  • WIPO fieldDigital communication
  • WIPO sectorElectrical engineering

Abstract

Disclosed is an exemplary multiple hypothesis engine that provides situation assessment capabilities regarding cyber auto tactics techniques and procedures. Dynamic cyber adversarial operations are evaluated via a combination of techniques using a Bayesian multiple hypothesis tree, or graph, as a framework. A top-down probability propagation mechanism solves different aspects of the problem in a round-robin fashion. The top-down probability propagation mechanism comprises the Hypothesis Refinement Engine. A model-based abductive reasoner comprising The Hypothesis Validator is used to confirm or refute the refined hypothesis. A model-based learning engine comprising Behavior Model Trainer is used to incrementally augment the knowledge base of behavior models as new adversarial TTPs are discovered. These three techniques behave in a cooperative manner by operating upon the Bayesian multiple hypothesis tree framework.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.