Brought to you by EarthWeb
IT Library Logo

Click Here!
Click Here!

Search the site:
 
EXPERT SEARCH -----
Programming Languages
Databases
Security
Web Services
Network Services
Middleware
Components
Operating Systems
User Interfaces
Groupware & Collaboration
Content Management
Productivity Applications
Hardware
Fun & Games

EarthWeb Direct EarthWeb Direct Fatbrain Auctions Support Source Answers

EarthWeb sites
Crossnodes
Datamation
Developer.com
DICE
EarthWeb.com
EarthWeb Direct
ERP Hub
Gamelan
GoCertify.com
HTMLGoodies
Intranet Journal
IT Knowledge
IT Library
JavaGoodies
JARS
JavaScripts.com
open source IT
RoadCoders
Y2K Info

Previous Table of Contents Next


3.2. RELATIONAL REPRESENTATION

Conceptual representation is the main way for putting the patterns of the world into a computer program. The task of concatenation of pattern situations, preparation of the ways of reasoning, inference is the procedural representation of the metapatterns. The essential tool is the rule; that is the reason why another name for expert systems is rule-based system. The rule is the first Aristotelian syllogism in the form: if...then.... The variables of the syllogism are joined by the and/or logical connections. The selector else points to other cases. This formula is a rule; the rules are sequenced in the succession of logical thinking or pointed at a jump in the sequence (else-->go to).

The sequencing of the rules, internal and external control of reasoning, offers visual representation by graphs. Object representation with its messages, selectors, is a good basis for graph representation; similarly, the frame has a graph structure.

More complex relations are organized in nets. These contain not only the straightforward sequences of logical inference, and some permitted loops for return to check or to try another way if the previously chosen one is not the right solution. Nets are organizations of many returning alternatives, offering ways of multiple interactions, controls, intermediate choices. The complexity of real-life events can be represented in a more flexible way though the net representation and the frame one are not contradictory, they can be connected as this is shown on Figure 2. These nets express the semantic content of the event, situation representing a more or less complete pattern of conceptual generalization. Connecting single objects into a relational entity, these nets are named semantic or conceptual nets. They usually describe a story or a part of that and, therefore, the verbal correspondent is a script. Nodes of a graph are usually objects, the branches represent relations (is a, part of), actions. In this latter, important action representational scheme which lends the meaning of the script name, the node-objects are actors. The actor-type representation is a relevant emerging concept. We shall come back to that later, discussing different possible worlds of different actors.

The representation scheme is also a guidance for knowledge acquisition, as was mentioned before. Knowledge engineers first clarify the concepts, putting them into an object frame; then the environment of the object in a more elaborated and general frame scheme, and continue the inquiry by representing typical situations, actions, events, drafting synopses, scripts, and putting them into net-like schemes.

Although the basic ideas are very simple, the software quality requirements of the schemes are even higher than in binary graph-organized databases, controlled externally. The multiple loops of the nets, the consequent, unambiguous process of data exchanges, transformations of values, dimensions of those, especially due to the actions on the branches are very sensitive. Because one scheme should cover several individual scripts, this sensitivity is even higher, the automatic operation of concatenations, backtracks, trials on different routes, embedded loops on the net should be unambiguous, the result must be somehow verifiable if the algorithm of the operation delivered the required, feasible solution. The idea of internal control was one of the decisive ones to make the machine procedure more intelligent than the usual, externally controlled databases. Decisions on branching, routings on the graphs, i.e., directions of inference, are made by the algorithm itself in the process of machine solution, instead of the human user (Figure 3). The control can work by decisions on intermediate data and estimation of a best or suitable way to a certain goal. The algorithm should be efficient, applicable in real time interaction with the user on the dedicated computer platform.

All these circumstances advocate for the application of professional tools, shells, which should prove the performance of all these conditions, and offer a wide variety of representations, objects, frames, nets, all programmable and graphically representable, providing guidance and flexibility, automatism and control features simultaneously.


Previous Table of Contents Next

footer nav
Use of this site is subject certain Terms & Conditions.
Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details.