Terrance Swift, SUNY at Stony Brook; Calvin C. Henderson, Richard Holberger and Edward Neham, Systems Development and Analysis; John Murphy, DHD Systems, Inc.
CCTIS, the Cargo Container Targeting Information System, was developed for the U.S. Customs Service to help monitor and control goods imported by ship. As an expert system, CCTIS has a combination of features which make it of interest to the applied A.I. community. First of all, CCTIS interacts with a large database -- but unlike most data-oriented expert systems CCTIS is used in a transactions-oriented environment and needs the speed of such a system. Secondly, there exists no single cognitive model for the domain of import control, and it is unlikely that such a model can be developed in the near future. To address this problem CCTIS includes the ability for users to weigh and parameterize rules. And thirdly, the information CCTIS uses is often derived from free text of low quality that must be corrected and analyzed through natural language analysis techniques.
The system uses a logic-based approach to solving these problems, defining explicit algorithms to extract data from text, and logical rules to analyze transactions. Our experience shows that this approach can produce a robust system. CCTIS has been used every business day for over a year in the two largest ports in the country, and has aided in the seizure of a number of illicit goods. Design is underway to merge CCTIS with another Customs A.I. system and to deploy the resulting system nationally, processing every sea-based import into the U.S.
(This paper reflects the opinions of the authors only, and does not represent policy of the U.S. Customs Service.)