Module 2. Information systems

Lesson 4

KNOWLEDGE MANAGEMENT SYSTEMS

4.1  Introduction

This lesson discusses the concept of knowledge management and various types of knowledge based systems. These topics will be useful for students to have idea of current trend in the information systems for improving decision making ability of managers.

Knowledge based information systems are referred as Knowledge Management Systems (KMS). These systems manage knowledge in the organization. They support creation, organization and dissemination of knowledge to managers and other workers of the organization. Many organizations are building knowledge management systems to manage organizational learning, solution to complex problems and decisions, policy issues, and business know-how. Such systems provide quick feedback to knowledge workers and encourage behavioral changes in employees to improve business performance. For example, taking a piece of advice on complex issue such as best practices on animal health care, ethics of production management, clean milk production, transfer of technology to farmers, etc. Main objective for creation of KMS is sharing of ideas.  For example, sharing the findings of “composition of cheese to increase its shelf life” by a researcher can lead to improved products that are using cheese as ingredient or it may also lead to some other innovative ideas. KMS may be stand alone or Web based systems but web based systems are more effective to disseminate knowledge to scattered and large number of stakeholders. Some of the advantages of KMS systems are:

· Sharing of valuable organizational information throughout organizational hierarchy.

· Can avoid re-inventing of wheel, reducing redundant work.

· May reduce training time for new employees

· Retention of intellectual property after an employee leaves organization

4.2  Artificial Intelligent Systems

Artificial intelligence is an outcome of synergy of many fields like computer science, mathematics, biology, philosophy, linguistics, neural science, engineering, etc. The primary goal of AI is to develop computers that can simulate the ability to think as well as see, hear, walk, talk, and feel like human beings. Computer functions are associated with human intelligence such as reasoning, learning, acquire and apply knowledge, creativity and imagination, and problem solving. AI systems learn from a given set of inputs and outputs. These store experiences in their memory, generalize them and become ready to deal with new circumstances (i.e., new inputs). AI applications may be classified under three major areas such as cognitive science, robotics, and natural interface. However, these classifications overlap each other sometimes. Figure 6.1 given below shows major application areas of AI.

Fig. 4.1 Important application areas of artificial intelligence

Information systems that make use of artificial intelligent techniques in decision making process are known as artificial intelligent systems. In such systems, knowledge is major point of focus rather than data or information. AI technologies have been successfully explored to its full potential in the areas of problem solving i.e. concepts and methods for building programs that reason about problems rather than calculate a solution. Development of expert systems for decision making is the most widely used AI application. However, significant results have also been achieved in other areas of AI like intelligent robotics, voice recognition, computer vision, natural language processing etc. AI systems have several advantages over human intelligence as given below:

· Knowledge in AI systems is permanent

· Can be easily duplicated and disseminated

· Consistent and thorough

· Knowledge can be documented

4.2.1  Characteristics of AI system

· Symbolic processing: AI is a branch of Computer Science that deals with symbolic, non algorithmic method of problem solving.

· Heuristics: It consists of intuitive knowledge or rules of thumb, learned from experiences.

· Drawing inferences: AI systems have the ability of reasoning or drawing inferences from facts and rules using heuristics or other search approaches.

· Machine learning: AI systems have mechanical learning capabilities called machine learning in parallel with human beings.

4.3  Expert Systems

Knowledge based expert systems or simply Expert Systems (ES) are computer programs that use human knowledge to solve problems which normally would require human intelligence. Expert systems represent expertise knowledge in form of data or rules within the computer program. These rules and data can be called upon when needed to solve problems. Expert system emulates the interaction with user just like a human expert to solve a problem. Program asks a series of questions and end user provides input by selecting one or more answers from list of answers or by entering data. Program will keep on asking questions until it reached to a conclusion. The conclusion drawn by system may be a selection of single solution or a list of possible solutions arranged in order of likelihood. Expert system can explain why data is needed and how conclusions were reached. Some expert systems are complex to build while some are easy. In a very simple case, consider a tree diagram on paper describing how to solve a problem. By making a selection at each branch point, tree diagram can help to make final decision. This type of tree structured logic can easily be converted into computerized program that is easier to use.

More elaborate systems may include confidence factors allowing several possible solutions to be selected with different degree of confidence. Output of an expert system may be information, an instruction or a risk judgment. A different problem within the domain can be solved using the same program without reprogramming. Ability of these systems to explain reasoning process through back traces and to handle levels of confidence and uncertainty provides an additional feature that simple expert systems do not have.

Most of the expert systems are developed using a specialized software tool called shells. A shell contains the user Interface, script language for developing expert system, a format for declarative knowledge base, and an inference engine, different data structures, connectivity with other programs and databases etc. Some of the commonly used shells for developing expert systems are CLIPS and JESS.

· CLIPS: CLIPS is a public domain software tool for building expert systems. The name is an acronym for "C Language Integrated Production System." It was developed by the Software Technology Branch (STB) at the NASA/ Lyndon B. Johnson Space Center. It was released in 1986 for the first time and has undergone continual refinement and improvement ever since. CLIPS is probably the most widely used expert system tools because it is fast, efficient and free. CLIPS is written in C, extensions can be written in C, and CLIPS can be called from C. Its user interface more closely resembles that of the programming language LISP. It supports rule-based, object-oriented and procedural programming. CLIPS inference engine provides only forward chaining.

· JESS: JESS is an expert system shell and it is even based on CLIPS. The name is acronym for Java Expert System Shell. JESS was originally a clone of the essential core of CLIPS, but has begun to “acquire a Java-influenced flavor of its own”. JESS is completely programmed in Java by Ernest J. Friedman-Hill at the Sandia National Laboratories for the U.S. Department of Energy. It was first released in 1995. It provides both forward and backward chaining.

These shells come equipped with an inference mechanism with several levels of sophistication "forward chaining", "backward chaining" and "mixed chaining" for drawing conclusions and require knowledge to be entered according to a specified format. Forward chaining (also known as data driven reasoning) is the questioning of an expert who has no idea of the solution and investigates progressively (e.g. fault diagnosis). Inference technique uses IF - THEN rules to deduce a problem solution from initial data. In backward chaining (also known as goal driven reasoning), the engine has an idea of the target (e.g. is it okay or not?). It starts from the goal in hopes of finding the solution as soon as possible. Inference technique uses IF - THEN rules to repetitively break a goal into smaller sub-goals which are easier to prove. In mixed chaining the engine has an idea of the goal but it is not enough: it deduces in forward chaining from previous user responses all that is possible before asking the next question. So quite often he deduces the answer to the next question before asking it.

4.4  Components of Expert System

Components of an expert system include knowledge base and software modules that perform inferences (or reasoning) on knowledge in the knowledge base and communicate with users for question and answer. Figure 6.2 illustrates interrelated components of an expert system.

Fig. 4.2 Components of expert system

The role of individual components is described as follows:

4.4.1  Domain expert

The individual person who is expert in solving the given problem? Domain knowledge is with him (expert). Here the system is intended to solve the problem as expert.

4.4.2  Knowledge engineer

The individual who encodes the expert knowledge in a declarative form that can be used by expert system.  Knowledge engineer acquires domain knowledge from expert and builds a knowledge base for particular problem using expert shell.

4.4.3  Knowledge base

Knowledge base of expert systems contains both factual and heuristic knowledge. Knowledge available in text book, journals and standard practices about a specific domain is known as factual knowledge. It is declarative knowledge either in true or false statements. Heuristics knowledge is informal, experiential, and judgmental knowledge of an application area that constitutes rules of good judgment in the field.  Heuristics cover the knowledge of how to solve problems efficiently and effectively, how to plan steps in solving a complex problem, how to improve performance etc. In comparison to factual knowledge, heuristic knowledge is less rigorous, rarely discussed, and is largely individualistic. It is the knowledge of good practice, good judgment, and plausible reasoning in the field. It is the knowledge that underlies the "art of good guessing".

Knowledge in expert systems may be represented in many ways. Widely used method is form of production rules or simply rules. A rule consists of IF and THEN part (also called a condition and an action). Expert systems whose knowledge is represented in rule form are called rule-based systems. Another commonly used method of knowledge representation is case-based. In this method knowledge base is created in the form of cases, that is, past performance, occurrences, and experiences. Other methods of representation are frame-based and object-based.

4.4.4 The inference engine

Inference engine processes the knowledge such as rules and facts related to a specific problem. It recommends the course of action for a user. It is a computer program designed to produce reasoning based on logic. Several kinds of logics are used in expert systems such as propositional logic, predicates of order 1 or more, epistemic logic, modal logic, temporal logic, fuzzy logic, etc. With logical reasoning, the engine is able to draw conclusions from the knowledge contained in the rule base. Knowledge is almost always incomplete and uncertain. To deal with uncertain knowledge, a rule may have associated with it a confidence factor or a weight. Set of methods for using uncertain knowledge in combination with uncertain data in reasoning process is called reasoning with uncertainty. For instance, an important subclass of methods for reasoning with uncertainty is called "fuzzy logic," and the systems that use them are known as "fuzzy systems."

The expert system process data in two ways i.e. batch or conversational.  In batch mode, the expert system has all the necessary data to process from the beginning. For the user, the program works as a classical program. User provides data and receives results immediately. Reasoning and processing of data is invisible to the user. In conversational method, user interactively answers the questions during execution of the system. This method is useful when the problem is big, complex and all necessary data is not available at the starting. The software tries to solve the problem by asking the missing data from the user, gradually approaching the goal as quickly as possible.

4.4.5 User interface

User interface is a program that controls the dialog between the user and the system. Users interact with system by putting their questions and answers and get advice from the expert system. User interface provides an easy and user friendly environment to operate the system.

4.4.6 Users

Users are the individual person who is consulting with the system to get advice on the particular problem.

4.5  Applications of Expert Systems

· Decision Management: Worker performance evaluation, Demographic forecast

· Diagnostic / problem shooting: Fault diagnosis, Help desk operations, debugging

· Design/ configuration: Manufacturability studies, optimum assembly plan

· Selection/ Classification: Material selection, Information classification

· Process Monitoring/ Control: Machine control, production monitoring

4.6  Benefits of Expert Systems

· Expert systems are faster and more consistent.

· The knowledge base may have knowledge of several experts.

· Does not tire and distract from overwork or stress.

· Preserve knowledge of experts.

4.7  Limitations of Expert Systems

· Can solve one type of problem in a limited domain of knowledge

· Inability to learn from experiences

· Maintenance problem

· Development cost may be too high. It includes cost of hardware, software, knowledge engineer, expert time etc.

4.8  Current Trends – Knowledge based DSS

Knowledge based DSS or intelligent DSS is the current buzz word in DSS research field. It is an extended version of DSS which derives knowledge through artificial intelligent tools like artificial neural networks, fuzzy logic, genetic algorithms, experts systems, knowledge representation methods etc. Such systems assist in decision making process where information is incomplete, imprecise, uncertain, non linear, unstructured and decisions are to be made using human judgment. Recent developments in field of information technology have further enhanced the strength of intelligent DSS for example WWW, Internet, intranet, multimedia, virtual reality, static/ mobile ad hoc networks, wired/ wireless sensor networks, RFID, image analysis, etc. have given a real momentum in quality data collection and processing.

With advances in micro-electronic equipment for example sensors, potential parameters about animal can be recorded continuously online without human intervention. The possibilities comprise electronic identification, automatic weighing, temperature and activity measurements, automatic regulation of feed and water intake, geographic positioning of individual animals and monitoring animals through video recordings using image analysis. Recent examples have shown the usage of such data for developing knowledge based DSS with the help of artificial intelligence techniques for heat detection, daily gain in body weight, water consumption, and monitoring of pregnancy rates, early diagnosis of mastitis and lameness. Purpose of these systems/ studies is to allow managers/ users to detect changes in the monitored variable for welfare of animals and increasing productivity of dairy farms.