Page 315 -

298 Chapter 8

troubleshooting, and other forms of support to knowledge workers. Expert systems
are decision support systems that do not execute an a priori program but instead
deduce or infer a conclusion based on the inputs provided. Natural language process-
ing also grew out of AI research. Linguistic technologies resulted in automating the
parsing (breaking into subsections) and analysis of text. Common applications today
are voice interfaces or natural language queries that can be typed in to search data-
bases. Similar AI technologies can also be applied to analyze and summarize text or
to automatically classify content (e.g., automated taxonomy tools). Many of the auto-
mated reasoning capabilities studied in AI research were encapsulated in autonomous
pieces of software code, called intelligent agents or software robots (softbots). These
agents act as proxies for knowledge workers and they can be tasked with information
searching, retrieving, and fi ltering tasks.

Intelligent Filtering Tools
Intelligent agents can generally be defi ned as software programs, which assist their
user and act on his or her behalf, such as a computer program that helps you in
newsgathering, acts autonomously and on its own initiative, has intelligence and can
learn, and improves its performance in executing its tasks ( Woolridge and Jennings
1995 ). They are autonomous computer programs, where their environment dynami-
cally affects their behavior and strategy for problem solving. They help users deal with
information. Most agents are Internet based, that is, software programs inhabiting the
Net and performing their functions there.
The following features are necessary to defi ne a true intelligent agent ( Khoo, Tor,
and Lee 1998 ):

Autonomy The ability to do most of their tasks without any direct assistance from an
outside source, which includes human and other agents, while controlling their own
actions and states.
Social ability The ability to interact with, when they deem appropriate, other software
agents and humans.
Responsiveness The ability to respond in a timely fashion to perceived changes in the
environment, including changes in the physical world, other agents, or the Internet.
Personalization The ability to adapt to its users needs by learning from how the user
reacts to the agent ’ s performance.
Initiative The ability of an agent to take initiatives by itself, autonomously (out of a
specifi c instruction by its user) and spontaneously, often on a periodical basis, which
makes the Agents a very helpful and time saving tool.