|
Features
|
| Fuzzy Logic and the Measures of Certainty in eCommerce Expert Systems
Earl Cox examines the organization of business rule systems and how,
through the use of Fuzzy Logic, rules should be organized and written to
address the uncertainty and imprecision of business decisions. |
| AI@Work Prolog Knowledge Base Supplies Embedded and Web Remote
Control: Intelligence for Microscopes; Automated Knowledge Discovery with
Data-Driven Knowledge Engineering: Fuzzy Logic Based Approach - Simulates
Naval replenishment maneuvers at sea. |
| The Rise of Knowledge for Sale Dan Rasmus discusses how Internet
knowledge markets might facilitate the discovery of those that have knowledge
and match them with those that need knowledge. |
| Genetic Algorithms and Intelligent Agents Team Up: Techniques for Data
Assembly, Preprocessing, Modeling, and Decision Optimization Larry M.
Deschaine, Jennifer McCormack, Dorian Pyle, Frank Francone discuss a set
of techniques for optimal real-time decision making from distributed, heterogeneous
information found in financial, industrial, and scientific data. |
| Simulation of Petri Nets in Prolog: Modeling Dynamic System Behavior
Girish Keshav Palshikar illustrates how symbolic logic programming can
model dynamic, concurrent, asynchronous distributed or stochastic behavior
found in protocols, office automation, industrial control, and real-time
embedded systems. |
| "Ruling" the World of Web Services: Application Servers from
Java Beans Tom Ronk and Jeff Weyer describe how Enterprise Java Beans
and Component-Based application design is becoming the foundation of the
next generation B2B application integration. |
| Regulars |
|
| Editorial |
|
| AI and the Net - Research in Intelligent
Technology - Research projects that may lead to future intelligent applications |
|
| The Book Zone - SQL for Smarties
- Obtaining the most from a database |
|
| Product Updates -------------------------------> |
25 late breaking product announcements from
around the world in the fields of: |
| |
Business Forecasting |
Business Rules |
| |
Data Mining |
Decision Support |
| |
Expert System Development Tools |
Intelligent Tools |
| |
Knowledge Based Systems |
Modeling and Simulation |
| |
Natural Language Processing |
Voice and Speech Recognition |
| |
Announcements |
Conferences |
| |
Training |
|
| Product Service Guide - Provides access to information
on an entire category of products |
|
|
| PC AI Blackboard - AI advertisers bulletin board |
|
|
|
|
Artificial Intelligence, a sub-domain of computer science,
focues on the development and deployment of intelligent software and hardware
systems that emulate human reasoning techniques and capabilities. Knowledge
base systems, a sub-field of AI, emulates the human decision-making process
by employing different information representation and manipulation schemes: |
| |
Rules-Based systems, which capture knoweldge in if-tehn
statements or rules, are an intuitive and wide spread knowledge representation.
Knowledge related to conditions or limits encodes easily, as does knowlege
that must match patterns. (see Prolog Knowledge Base Supplies Embedded
and Web Remote Control: Intelligence for Microscopes) |
| |
Artificial neural networks, derived from computer-based
reserach on modeling human brain behavior, are excellent at associative
problems - the type of progblem where finding an existing input set that
is similar provides an acceptable solution. That is, relevant concepts,
represented by the neural network, are compared against the new problem
data. The key advantage here, is that example data sets train the neural
network versus developing a program comes at a price - this knowledge is
not in human readable form. |
| |
Fuzzy logic, with its roots in set theory, was developed
to handle situations where set membership is not clearly defined. An exapmle
would be the thermometer vlues for the terms "hot", "cold"
and "warm" - at what temperature does boiling water become warm
or cold, as it cools. This can also vary with the membership designer. (see
Fuzzy Logic and the Measures of Certainty in eCommerce Expert Systems)
|
| |
Decision trees, which capture the structure of the
decision-making processes simplify development of configuration and diagnostic
applications. When knowledge is structured into a series of steps and decision
points, this can be a very efiicient representation. Since decision trees
are developed before their use, they may lack the flexibility of other representation. |
| |
Model-Based Reasoning, which uses a mathematical model
to emulate real processes or catpure knowledge, was initially created to
support industrial processes (such as refinery or manufacturing). Since
changes are applied to the model and the results verified, this representation
can predict the effect of future activities - such as valve or environmental
changes. (see Automated Knowledge Discovery with Data-Driven Knowldge
Engineering: Fuzzy Logic Based Apporach as well as Simulation of
Petri Nets in Prolog: Modeling Dynamic System Behavior) |
| |
Each approach has its strengths and weaknesses but when combined
they improve their efficiency and overcome inherent weakness of the standalone
represenation. Many of the articles in this issue illustrate this hybrid
approach to using AI technolgy to solve real world problems (see Genetic
Algorithms and Intelligent Agents Team Up: Techniques for Data Assembly,
Preprocessing, Modeling, and Optimizing Decisions) |
| |
Models can be constructed from these, and a number of other
knowledge representations, capturing the crucial part of business processes.
By manipulating these models, it is possible to predict the effects of various
actions. Built as part of a knowlege base system, the models can predict
outcomes based on different business scenarios. This type of reasoning is
very useful as part of a sophisticated decision support system. |
| |
Once these systems are in place, it is possible to collect
valuable information and knowledge. It may now be possible to buy and sell
that knowledge on the Internet (seeThe Rise of Knowlege For Sale) |
| |
We find that the term business rules, referring to the knowlege
that is important to operating a business, is frequently confused with rule-based
systems (a type of knowlege representation). We hope that after reading
some of the articles in this issue, this distinction will be clear. (see
"Ruling" the World of Web Services: Application Servers from Java
Beans). |
| |
|
| |
Terry Hengl |
