Artificial Intelligence System

All human beings have intelligence, which they use for problem solving. Intelligence when supported by knowledge and reasoning abilities becomes an artificial intelligence. When such an artificial intelligence is packed into a database as a system, then what we have is an AI system.

AI Systems Contd.
AI systems fall into three basic categories:
The Expert Systems (Knowledge Based) The Natural Language (Native Languages) Systems and The Perception System (vision, speech, touch)

Structure of AI system
Artificial Intelligence System

Natural Language

Expert uses

Perception
uses
Size, Shape Image, Voice

uses
Native Language Knowledge

Knowledge
Applies
Human Like Reasoning

Applies
Language Reasoning

Applies
Sensing Abilities for Reasoning

Some more facts about AI

Artificial intelligence is a software technique applied to the non-numeric data expressed in terms of symbols, statements and patterns. It uses the methods of symbolic processing, social and scientific reasoning and conceptual modeling for solving the problems. The AI systems are finding applications in configuration, design, diagnosis, interpretation, analysis, planning, scheduling, training, testing and forecasting.

Facts Contd.
The AI systems do not replace people. They liberate experts from solving common/ simple problems, leaving the experts to solve complex problems. Artificial intelligence systems help to avoid making same mistakes, and to respond quickly and effectively to a new problem situation.

Knowledge Based Expert System (KBES)

Decision making or problem solving is a unique situation riddled with uncertainty and complexity, dominated by the resource constraints and a possibility of several goals. In such cases, flexible systems (open system) are required to solve the problem. Most of such situations, termed as the unstructured situations, adopt two methods of problem solving, generalized or the KBES.

Generalized Vs KBES
Generalized Approach 1. It considers the generally applicable constraints, examines all possible alternatives and selects one by trial and error method with reference to a goal. Knowledge Based 1. It considers the specific constraints within a domain, examines the limited problem alternatives within a knowledge domain and selects the one with knowledge based reasoning with reference to goal.

Comparison Contd.
2. All alternatives are 2. Only limited considered and the alternatives are resolution of the considered and problem is by trial resolution is made and error, with no by a logical assurance, whether reasoning with the it is the best or the assurance of the optimum. local optimum. 3. This approach is 3. This approach is dominated by dominated by the procedure or method reasoning process based on the knowledge.

KBSE Model
To build a knowledge based system certain prerequisites are required.
1. That a person with the ability to solve the problem with knowledge-based reasoning should be available. 2. That such an expert should be able to articulate the knowledge to the specific problem characteristic.

Components of KBES
It has three basic components which are necessary to build the system
Knowledge Base
It is a database of knowledge consisting of the theoretical foundations, facts, judgments, rules, formulae, intuition and experience. It is a structural storage with facilities of easy access.

Interface Mechanism
It is a tool to interpret the knowledge available and to perform logical deductions in a given situation.

User Control Mechanism

It is a tool applied to the inference mechanism to select, interpret and deduct or infer, the user control mechanism uses the knowledge base in guiding the inference process.

Contd.
In KBES three components are independent of each other. This helps in modifying the system without affecting all the components. The KBES database, stores the data, the cause-and-effect relation rules, and the probability information on event occurances.

Methods of Knowledge Representation

Semantic Networks Frames Rules

Semantic Networks
Knowledge is represented on the principles of predicate function.