The Relationship Between Artificial Intelligence and Data Mining: Application to Future Military Information Systems. Marion G. Ceruti, Senior Member, IEEE.
The Relationship Between Artificial Intelligence and Data Mining: Application to Future Military Information Systems Marion G. Ceruti,Senior Member, IEEE Space and Naval Warfare Systems Center, San Diego San Diego, CA 92152-5001 in AI and DM. For example, knowledge engineering is becoming more automated so that a subject-matter expert, such as one skilled in battle planning and execution, soon will be able to populate a KB with the help of user-friendly tools, and without the direct assistance of a knowledge engineer. (See, for example, [1]). Combinations of technology will have a greater impact on future capabilities than any one technology taken separately. Advances in networking, medical remote sensing, and DM can be combined to achieve a whole that is far greater than the sum ff its parts. For example, in the future, the preseme and origin of new chemical weapom can be detected by mining geographic patterns from networks of devices worn by troops in the field designed to record and transmit a soldier's or a marine's vital health data and environmental data. These geographic patterns will help to identify the origin of the attack. It also will affect the early response and treatment of wartime casualties with a result of more lives saved on the battlefield. This is one of many examples of future applications of DM in the military domain. The use of AI technology for DM will grow more rapidly in comparison to the use of non-intelligent techniques. AI will continue to have a greater and more significant role in DM. The field of DM is in its infancy and we can only begin to imagine future applications.
Abstract This position paper describes the relationship between data mining (DM) and artificial intelligence (AI) and how both can improve future military information systems. As these information systems take advantage of the technology trends, they will derive their upgrades from expert systems, knowledge-bases ryes), inference engines, robotics, natural language, remote sensing, knowledge discovery (KD), and data-mining over networks. This will result in an expansion to all echelons of present capabilities that now me available to warfighters only at the higher echelons. Moreover, it will result in an improved battlefield awareness for all warfighters beyond anything that is implemented today. Different researchers have defined DM in various ways. It is defined most comprehensively as the search for and extraction of hidden and useful patterns, suucaues and trends in large, multidimensional, and heterogeneous data sets that were collected originally for another purpose [2]. As technology evolves, our notion of DM will evolve with it. For example, the distinction between DM and some aspects of AI is becoming diffuse and vague. Historically, they originated from different sectors of the informationmanagement community. As both fmlds mature, they will have more technical areas of application in common. Progressively, more DM techniques are derived from AI technologies that hemofore have been associated with the AI community. An example is the use of Bayesian networks in DM classification, [2]. A trend in the KB community is to integrate more pmbabilistic and possibilistic technologies into knowledgebased systems. Just as AI enables DM, DM also enables AI. In facL DM has the potential to make substantial contributions to AI, an example being the population of KBs using KD and DM. What does this interaction imply for future military informarion systems that will rely increasingly on AI and DM technology? These systems include tacticalsystems, such as those that support and enable command and control, intelligence, surveillanee, and network-centric warfa~. As the technology of DM expands to include more techniques, the number of application domains to which DM and AI can contribute also increases. Both tactical and non-tactical systems can derive benefit from the technological advances
0-7803-6583-61001510.00 © 2000 IEEE
.
.
.
.
.
.
.
.
.
.
.
.
I
Acknowledgments The author thanks the Defeme Advanced Research Projects Agemy financial support. This work was produced by a U.S. government employee as part of official duties and is not subject to copyright. It is approved for public release with an unlimited distribution.
References [1] M. Burke, Rapid Knowledge Formation (RKF) Program Description, http:lldtan.darpa.milliso/programtemp.asp?mode=331 [2] M.G. Cemti and SJ. McCarthy, "Establishin 8 a DataMining Environment for Wmtime Event Prediction with an Object-Oriented Command and Control Database," Proceedings of the IEEE International Symposium on Object-oriented Real-time Distributed Computing, ISORC2K, pp. 174-179, March 2000.
1875
.
.
.
.
.
.
I
Ill
lull
I