3.12 Bibliography

3.12 Bibliography

AAAI (1998) Symposium on Artificial Intelligence and Link Analysis.

Chabrow, E. (January 14, 2002) "Tracking The Terrorists," Information Week.

Grady, N.,Tufano, D., and Flanery, R. (1998) "Immersive Visualization for Link Analysis," Oak Ridge National Laboratories Publication.

Picarelli, J. (1998) "Transnational Threat Indications and Warning: The Utility of Network Analysis," Pacific-Sierra Research Corporation Publication.

Westphal, C. and Blaxton, T. (1998) Data Mining Solutions, New York: Wiley Computer Publishing.

Wong, R. (1998) "Financial Crimes Enforcement Network," U.S. Department of the Treasury Publication.

Chapter 4: Intelligent Agents: Software Detectives

4.1 What Can Agents Do?

In a networked environment such as ours, a new entity has evolved: intelligent agent software. Over the past few years, agents have emerged as a new software paradigm; they are in part distributed systems, autonomous programs, and artificial life. The concept of agents is an outgrowth of years of research in the fields of AI and robotics. They represent concepts of reasoning, knowledge representation, and autonomous learning. Agents are automated programs and provide tools for integration across multiple applications and databases, running across open and closed networks. They are a means of retrieving, filtering, managing, monitoring, analyzing, and disseminating information over the Internet, intranets, and other proprietary networks.

Agents represent a new generation of computing systems and are one of the more recent developments in the field of AI. Agents are specific applications with predefined goals, which can run autonomously; for example, an Internet-based agent can retrieve documents based on user-defined criteria. They can also monitor an environment and issue alerts or go into action based on how they are programmed. In the course of investigative data mining projects, for example, agents can serve the function of software detectives, monitoring, shadowing, recognizing, and retrieving information for analysis and case development or real-time alerts.

Agents can be used by investigators and analysts to work on their behalf; for example FinCEN, the U.S. Treasury agency set up to detect money laundering, must review all cash transactions involving dollar amounts of above $10,000. This amounts to roughly 10 million transactions a year, which cannot be manually monitored. The FinCEN Artificial Intelligence System (FAIS) uses an agent to weed through this large data space and search for abnormalities and fraud through the use of neural network and link analysis. However, before continuing, a definition should be established regarding what comprises an intelligent software agent.

4.2 What Is an Agent?

An intelligent agent is software that assists users and acts on their behalf. Agents autonomously perform tasks delegated by their creators and users. Agents can automate repetitive tasks, remember events, summarize complex data, learn, and make recommendations. For example, an agent can be used to monitor and search for a suspect's name from multiple government and commercial databases, or it can be set up to assemble evidence for use in a prosecution case.

Intelligent agents continuously perform three main functions, which differentiates them from other software programs:

  1. They are capable of perceiving dynamic conditions in an environment.

  2. They can take action to affect conditions in an environment.

  3. They can reason to interpret findings, solve problems, draw inferences, and determine future actions.

For example, agent software can act on behalf of investigators and, thus, reduce their workload by sifting through large amounts of data for evidence gathering. Agents have the capability to interact with the external environment and perceive changes in it; hence, they can then either inform investigators of changes, such as that a suspect on the INS list has entered the country. Or, they can be set up to react dynamically to findings, issuing an alert at the point-of-entry station, once a match of a suspect on the INS list is found. While there are multiple definitions of intelligent agents, this is their essential characteristic: a software agent is a computing entity that performs user-delegated tasks autonomously. An agent can perform many tasks; however, for investigative data mining the most dominant ones are likely to be information monitoring, retrieval, organization, and reporting.

Agent technology is not a single, new technology, but rather the integrated application of a number of network, Internet, and AI technologies. As such, developers normally do not set out to construct an agent; more commonly they set out to add new functionality to a new or existing application that posses agent-like features. These agent programs possess various forms of learning, creating, and modifying rules of behavior and developing strategies for collaborating among other programs, databases, networks, and users and even other agents. As you will come to see in subsequent chapters where we discuss other data mining technologies, agents can be integrated with other applications, enabling investigators and analysts to automate many tasks. In Chapter 11 we will propose a system using agent technology for the integration of human investigators and machine-learning algorithms to create a new type of evolutionary investigative system resulting in a fusion of human and machine intelligence.