TOPICS AND FAQS

A knowledge representation is a a scheme for collecting and preserving knowledge in an organized fashion. A table with rows and columns is one simple form of knowledge, but most often when one refers to a knowledge representation they mean a more flexible method which is able to capture richer relationships among elements. Examples include logic structures such as a rule base, semantic network, and logic languages like horn clause structure. The collection of knowledge in a system of knowledge representation forms a knowledge base.

Are there different kinds of knowledge?

Knowledge can be classified across several dimensions. The following are a few types:

Declarative Knowledge - factual knowledge about items
Procedural Knowledge - knowledge about how to perform a task
Static knowledge - knowledge which does not change
Dynamic knowledge - knowledge which changes or which incrementally grows
Shallow knowledge - abstracted knowledge as one would typically "take for granted"
Deep knowledge - knowledge built vertically upon basic first principles understandings

What is a Knowledge Based System?

Knowledge-Based Systems (KBS) are software programs which have as a central component an abstracted collection of knowledge (knowledge base) in some knowledge representation. The program uses this knowledge along with logical control to make decisions in the process of executing. These systems have advantages, for example, they are easily modified as new knowledge is learned and the knowledge can be easily reviewed by human inspection. Typical knowledge representations used in the system might include a rule base, semantic network, or criteria table representation. An rule-based expert system is an example of a Knowledge-Based Systems.

What is a semantic network?

A semantic network in a graphical knowledge representation. In the representation, individual items are related to one another by lines connecting them. For example, two individual's names might be connected by a named line (arc) called siblings that indicates they are brother and sister. This form of knowledge has an advantage because it is visually meaningful to humans since it parallels a similar kind of diagrammatic representation in educational psychology called a concept map. Semantic networks can be translated into other forms of representation, such as a rule base, and can be used by knowledge-based computer systems. A visual example of a semantic network can be found in the IPLI site under fun stuff.

What do we mean by "Intelligent" software?

The term "intelligent," when applied to software, is generally interpreted to refer to software which behaves in a way which seems like human intelligence. The status has been somewhat elusive over time. Many years ago, machine language programmers dreamed of an intelligent program that would allow them to specify their programs at a high level. This need was answered with the development of the software compiler, yet today few people would consider a compiler program intelligent. The bar has continued to raise as systems have become more sophisticated. This term can be contrasted with Artificial Intelligence and Artificially Intelligent systems.

What is Artificial Intelligence and when is a program considered Artificially Intelligent?

In a general sense, Artificial Intelligence in a branch of study considered to be the pursuit of development intelligent software. However, more strict, useful, definitions are sometimes applied. Clearly, it is a discipline of disciplines combining areas of Cognitive Science, Computer Science, and Theoretical Mathematics. More strict definitions stipulate that it is study of intelligent methods using the tools which have been developed in the AI field, such as knowledge representations and forms of algebraic manipulation of logic (e.g. First Order Predicate Calculus and Resolution Theorem Proving). By this definition, AI might be considered a discipline of advanced programming techniques, and a system which uses those techniques could be considered Artificially Intelligent.

What is an Intelligent Tutoring System?

Intelligent Tutoring Systems (ITS) are software programs which provide instruction for a learner with guidance and insight in the way a teacher would guide a student. In an ITS program the knowledge of "how to teach" is distinct from that which is to be taught and from that which the student knows. Each of these areas of knowledge may be captured in a knowledge base or at least some form of an abstraction which the program operates upon to control its execution.

What is a WebSim?

WebSim is a trademark of Powersim, Inc (Herndon, VA) and refers to a web-based simulation which individuals can play across the internet from around the world. These simulations are often used for business or education (though other applications can be created) and allow employees to play roles in their organization and explore various scenarios within the safety of a computer simulation. The model of a "dynamic system" is developed from a Systems Thinking approach, "wrapped" in an implementation using Java technology, and deployed on the internet or intranets. In fall 1997, IPLI acted as beta tester for Powersim's Metro Server product which functions as a server for WebSims.

What is a cognitive strategy?

In education, influencing a learner and assisting them in retention of information can be central and important tasks. To achieve these, the use of various cognitive strategies can be very effective, when carefully chosen to be appropriate for each learning situation. To influence a learner about the importance adequate pet care, the developer may appeal to affect (individual feelings), or to reinforce the importance of diet in nutrition, an authority figure such as a doctor might be enlisted to provide the advise. To assist in retention of information, cognitive techniques might be incorporated into the design and presentation of content which mimic mental processing of material. In data interpretation, for example, pictorials grouping related information together and emphasizing differences can be very useful. A conscious awareness and attention to these strategies and techniques results in systems which are better designed for human factors, interface, and learning.

How can we define "distance learning"?

On a basic level, distance learning is what the words say, "leaning which takes place at a distance". There are some basic features identified by individuals in the field, however, which are useful for characterizing this kind of educational effort. These include the following:

A defined instructional curriculum or program
A physical separation of the student and the instructional source
Some form of media supported by technology
Interaction between the student and the source of instruction

How does IPLI use a rational design for instruction?

IPLI uses a "rational" methodology to design instructional systems. This method can be broken into a four step process as follows:

Goal identification - The goal of the educational effort, potentially unattainable, is explicitly identified as an initial step (e.g., improvement of an employee's work performance).
Objectives definition - Real-world objectives are defined. These are attainable aims which will address achievement of the goal identified in step one (e.g., making an employee aware of factors which may improve is performance).
Learning paradigm selection - A learning paradigm is selected in conjunction with appropriate cognitive strategies and techniques in order to achieve the objectives set forth in step two (e.g., a drill and practice scenario).
Cognitive strategies selection - Cognitive strategies and techniques are selected in conjunction with, and to complement, learning paradigms to result in an overall system which achieves the defined objectives and addresses the identified goal.

What is Client / Server architecture?

A client / server (C/S) system is a software architecture which distributes processing between a client (user) computer and a central server. Part of the computer program is actually executed on each machine and the two components communicate over a network. The client portion of a C/S database may form queries on the user's machine while those queries are executed on the server returning only the results. The internet can also be considered a C/S system where the user's client (browser software) interacts with a server which returns HTML pages. In this kind of software, the software can be developed to create either thin or thick client systems.

How is a thin client different from a thick client?

In a client / server (C/S) system, the amount of software and processing carried out on the client machine versus the server machine may be shifted to place more of the software on the client or more on the server. When more software is shifted to the client, we say the client is thicker. Systems with large, task specific client software (e.g., traditional two-tier C/S database systems), are typically referred to as thick clients while small programs which are more generic such as a browser are typically referred to as thin client systems. There is a spectrum in-between which can be exploited to meet special constraints of specific software behavior, responsiveness, and bandwidth.

Are there different kinds of intranets?

There is some variation in what individuals identify as an intranet. Some individuals characterize any system which uses TCP/IP as an intranet regardless of how the software is distributed in a client/server architecture. At another extreme, some implicitly assume that the client software will be limited to a browser only and all other software will be located on a server machine or machines. The latter kind of intranet is often sighted as a low maintenance system which can even be implemented on a network computer do to its small software profile. Other installations may be somewhere in-between, having a browser - HTML server front end and other software both on the network servers and the client machines. To some extent, these factors may be chosen depending upon the needs of the users and constraints of the application and network connections.

What do we mean by CD-ROM development?

The desire to "develop a CD-ROM" is commonly stated as a way of saying there is an interest in developing a multimedia program, however the latter is perhaps a better way to approach it. It is really no more accurate to say, "we want to make a floppy disk." A CD-ROM, like a floppy disk, digital video disk, or magnetic tape, is really a medium for distribution of a software program or data. What is put on a CD-ROM may be a multimedia program, or simply all of the collected MS word documents you wrote last year. Also, a multimedia program can be distributed on any media, run from a hard drive, or even run over an intranet, if properly designed. At risk in this thinking is automatic exclusion of these potential alternatives (e.g., network delivery). It is best to think in terms of developing software and a solution. The selection of a platform, both for running any program and distributing it (multimedia or otherwise), can be considered in terms of the solution required to fulfill needs and meet development constraints.

How are learning paradigms used for education?

A learning paradigm is a systematic methodology and situation in which the user is placed for the purpose of promoting learning. Examples of learning paradigms include but may not be limited to:

Drill and practice - The learner is lead through information and then forced to reiterate points or behaviors learned
Simulation - The user is placed in a situation which approximates the reality in which they must perform. Experience gained in the simulation constitutes learning which is transferable to the real-world situation.
Didactic session - Information is presented by an instructor to the learner. The learner may interact to direct the presentation to areas of particular interest.
Discovery (Exploration) - The learner chooses a path through material without direction from an instructor. In this process the learners motivation and interest drives the pathway and learning.

What is Medical Informatics?

Medical Informatics is an applied discipline which lies in the intersection of computer science and medicine. It also draws from areas of library and information science, statistics, health policy making, and others. It is the science of how to collect and manage medical and health information using computers.

Design of computing solutions for medical informatics benefits best from expertise in computer science, medical and health sciences, and cognitive sciences. The former two due to the need for deep insight in both computing and health disciplines, and the latter due to the need for careful attention to human factors and interfaces which may be of particular importance to success in healthcare.

Among the major organizations assuming leadership in this area the American Medical Informatics Association (AMIA, sponsors of the annual Symposium for Computer Applications in Medical Care) and the Health Information Management Systems Society (HIMSS).

Why are clinical outcomes important in Medical Informatics today?

According to experts in the field, there are three general classes of clinical outcomes of importance to today's practice of medicine:

Costs - The actual cost for a patient treatment or procedure
Service outcomes - The satisfaction of the patient with care
Clinical outcomes - The improvement in medical condition, including the patient's perception of their health status.

Traditionally, it has been relatively speaking, easy to collect cost information and difficult to measure other areas. In this time of managed care, it is increasingly important to monitor all areas to assure reductions in costs to not adversely affect patient care. This is an important challenge which can in part be addressed through efforts in medical informatics development.

What is Systems Thinking?

Systems Thinking is an approach to gaining an understanding of a dynamic system through an understanding of the interaction of its parts. Studying a system from this perspective allows discovery of causal relationships, complexities, and predictive factors related to both desirable and undesirable outcomes. It is a useful tool for understanding any system ranging from scientific arenas to business management.

Powersim, Inc.(Herndon, VA) is one company manufacturing tools for web-based (WebSims) and stand-alone simulations developed from a systems thinking approach.

What is a dynamic system and the field of System Dynamics?

A system may be static (unchanging) or dynamic (changing). The study of dynamic systems and how they change over time is called System Dynamics. Understanding the behavior of dynamic systems is an important task in being able to control those systems and predict outcomes of decisions. Systems Thinking is an approach offering tools and methods to address System Dynamics.

Powersim, Inc.(Herndon, VA) is one company manufacturing tools for web-based (WebSims) and stand-alone simulations developed from a systems thinking approach.

How are developments for education, web applications, and assembly programs alike?

In many ways, "quality software development" is "quality software development". The factors and techniques used in one area are applicable to another and should begin independent of the underlying hardware, software, and language choice. Comparisons can be made in areas as follows:

Human factors - Careful attention to, and engineering of, human factors is important to all software interfaces and in particular to media oriented needs. These same design considerations should be applied based upon factors of the requirements and not the implementation or language chosen. In actuality, assembly language could be used to create the same exact presentations as a high-level authoring language, however the amount of code required would be unbearable by comparison. As such, we see that requirements may drive choice of an appropriate implementation language but the converse should be avoided except for some special situations.

Software engineering - The steps in design of software are similar across all developments. Typical traditional software engineering steps are as follows:

Needs and problem analysis
Requirements definition
High level design
Detailed design
Coding
Implementation
Evaluation and refinement

These steps are common across all areas of software development though in slightly different form. For example, in multimedia circles, the term high level design is replaced by "story boarding" and coding with "authoring". In many ways, these basic steps are universal to assessment of any problem and creation of a solution.

Sound software practices - Software development should always subscribe to sound design and coding principles. These typically include, but are not limited to principles such as:

Modular design - Software designed in a modular way is easier to maintain, understand, and modify.
Understandability - Given tradeoffs of efficiency versus understandable code, choices should be made in favor of more understandable code unless efficiency is an issue
Vision - Software should be built with an eye for future needs and potential extension

Similar to terms used for software engineering, depending upon the kind of application, different kinds of terminology may be applied. For example, subroutines in a traditional program may be compared in analogy to icon groups in an icon based authoring system.