AI Techniques for Context-Sensitive Tactical Decision Aids

Not Available Available commercially off the shelf software (COTS) has the potential to address the requirements for an intelligent security agent module. Agent software is designed to be deployed in remote computer systems in order to process at the desired location, or take advantage of available computing resources. Security of agent software is a primary concern for U.S. Government agencies and commercial businesses, including those that are planning to deploy agent-based systems on the Internet. This proposal presents a Phase I project for the implementation of an intelligent agent security module. It describes an approach that utilizes the distributed object middleware of the Object Management Group's (OMG's)Common Object Request Broker Architecture (CORBA) as the basis for secure agent messaging, coupled with available Neural Network agent software from Computer Associates. The goals of this project are; to develop an intelligent security agent module which integrates adaptable wrappers through CORBA facilities, use neural networks for security intrusion detection collection, interpretation, pattern recognition, adaptive learning capabilities and demonstrate an intelligent agent security module to the Navy. The resultant system can provide integrated network administration and security management reducing workload to manage an intelligent agent-based system.