Case Analysis Software Agent

The Problem
Case analysis tools used by law enforcement for matching investigative leads to databases have limitations that reduce their usefulness. Current tools have inflexible data formats that do not permit the recording of much of the data that is accumulated, and many of the data fields provided are not relevant to any particular case. A typical police intelligence database might have one field for the name of a suspect and two fields for aliases. For many suspects, the alias fields will be blank while no fields exist to record possibly critical information such as a set of Internet handles, gang nicknames, and names typically used by the suspect in committing fraud. Queries against intelligence databases tend to be slow and are limited in the information that can be requested. That is, they permit matching data for one type of record with limited search parameters and can retrieve only information specifically stored in a database.

The Solution: CASA
The ANSER team, funded under National Institute of Justice Cooperative Agreement 98-LB-VX-K021, is developing the Case Analysis Software Agent (CASA), designed to overcome the limitations of current case analysis tools. CASA is based on the Software Agent Development and Integration Environment (Sadie). Sadie's knowledge base uses a semantic network model for storing information. This permits the user to dynamically define new data fields, provide any number of values for a field, and define new data fields, provide any number values for a field, and define the limitations of a field, such as the lowest and highest numbers permitted. To permit more powerful queries, CASA uses Sadie's artificial intelligence capabilities, including commonsense reasoning, problem solving, inheritance of values from those in Sadie's generalization hierarchy, and inferences based on rules mined from the information.

This system includes the following modules:

• The Sadie knowledge base engine, which integrates several technologies into an intelligent software agent development shell, including traditional and fuzzy logics, expert systems, learning systems, inter-agent communication, and problem-solving systems.
• Record Manager, which permits any user to select whatever fields best represent the information for a given case and to easily create new fields as needed.
• Query Manager, which lets a user quickly and easily create queries to promptly retrieve needed case information.
• System Administrator, which takes care of system security and permissions for modifying records and assigning passwords.

Other Law Enforcement Applications
This system could be used to monitor transportation portals (such as airports) for known terrorists, drug dealers, or other known criminals on a "watch list." It can also be extended to provide access authentication, monitoring, and control functions for physical, computer, and network security systems.