Types and Levels of Knowledge

- Overview

What distinguishes humans from other animals or machines is our "conscience". While scientists often find it difficult to explain what conscience is, one can agree that it is the sum of our memory, that is, all the knowledge we have gathered so far. This knowledge creates different personalities that make humans behave differently and act differently. 

All human abilities, therefore, derive from this aggregated knowledge. Therefore, the prior knowledge that the teacup is hot prevents us from touching it. If we are going to make AI more complex, we need to provide them with more and more complex information about our world to perform complex tasks, which leads to the concept of knowledge representation (KR) in AI.

Knowledge representation in AI refers to how human information is represented in a way that computers can see and understand. This translation involves using symbols, diagrams, and other structures to describe concepts, relationships, and constraints. In this approach, we provide AI with the knowledge to perform logical reasoning and solve specific problems.

For example, suppose we have an AI program that can identify various diseases. Specifically, it must transmit information about symptoms, diseases, and treatments. The model allows the system to evaluate patient data, identify underlying disease and provide appropriate treatment options.

- Levels of Knowledge Needed To Be Represented in AI Systems

We have some concepts that are completely foreign to machines, such as intuition, intention, prejudice, belief, judgment, common sense, etc., while some knowledge is straightforward, such as knowing certain facts, general knowledge about objects, events, people, academic Subjects, languages, and other immediate things that machines have been able to understand with some success. 

With Knowledge Representation and Reasoning (KRR), we now have to represent this information in a machine-understandable format and make AI systems truly intelligent. Knowledge here would mean providing and storing information about the environment, reasoning would deduce this stored information, and intelligence would mean making decisions and actions based on knowledge and reasoning.

In order to solve the complex problems encountered in AI, one generally needs a large amount of knowledge, and suitable mechanisms for representing and manipulating all that knowledge. Let us first consider what kinds of knowledge might need to be represented in AI systems.

Following are the kinds of knowledge which needs to be represented in AI systems: 

• Objects: All the facts about objects in our world domain. E.g., Guitars contains strings, trumpets are brass instruments.
• Events: Events are the actions which occur in our world. Steve Vai played the guitar in Frank Zappa's band.
• Performance: It describe behavior (like playing the guitar) which involves knowledge about how to do things.
• Meta-knowledge: It is knowledge about what we know.
• Facts: Facts are the truths about the real world and what we represent. This can be regarded as the knowledge level.
• Knowledge Base: It is the main component of any human being to have a knowledge base. This refers to a set of information about any discipline, field, etc. For example, a knowledge base about building roads. 

We can structure these entities at two levels:

• The knowledge level - at which facts are described.
• The symbol level - at which representations of objects are defined in terms of symbols that can be manipulated in programs. 

- Types of knowledge in AI Systems

Knowledge is awareness or familiarity gained by experiences of facts, data, and situations. Given an understanding of the complexity of knowledge representation (KR) in AI, it is clear that to represent knowledge to machines, we must first identify and classify different types of knowledge. 

Knowledge representation can be described as the art of translating information into machine-readable form. These models assist AI systems in analyzing data, making sound judgments, and selecting options. Knowledge representation does more than just store data; it stores it in a way that the computer can use it intelligently.

Following are the types of knowledge in AI: 

• Declarative Knowledge: Facts and Concepts Explained
• Procedural Knowledge: Instructions and Task Execution
• Meta Knowledge: Understanding Knowledge About Knowledge
• Heuristic Knowledge: Expert Knowledge and Decision-Making
• Structural Knowledge: Relationships Between Concepts and Objects

- Knowledge Cycle in AI

In AI, the "Knowledge Cycle" refers to the iterative process where an AI system gathers information from its environment through perception, learns from that data, stores and organizes it as knowledge, uses reasoning to make inferences, plans actions based on that knowledge, and then executes those actions to interact with the world, essentially representing a continuous loop of acquiring, processing, and applying knowledge to make informed decisions.

The AI knowledge cycle comprises multiple elements or entities that are used to represent and utilize knowledge. These entities include perception, learning, knowledge, reasoning, planning, and execution. 

Key components of the Knowledge Cycle:

• Perception: Sensing and interpreting data from the environment, like seeing images or hearing sounds.
• Learning: Analyzing the perceived data to extract patterns and gain new knowledge.
• Knowledge Representation: Storing and organizing the learned information in a structured way that allows for reasoning.
• Reasoning: Using the stored knowledge to make logical inferences and draw conclusions.
• Planning: Creating a strategy or set of actions based on the knowledge and current situation.
• Execution: Carrying out the planned actions to interact with the environment.

It is necessary to explain the seven stages of the knowledge cycle in AI that may help in constructing intelligent systems. This cycle describes how AI systems capture, acquire, create, and apply knowledge to make decisions and solve problems. 

Each of the stages in this cycle enhances AI’s intelligence: 

• Data Collection: Gathering Relevant Data from Various Sources 
• Data Pre-processing: Cleaning and Transforming Data for Analysis 
• Knowledge Representation: Encoding Data for AI Systems 
• Knowledge Inference: Applying Algorithms to Make Predictions or Decisions 
• Knowledge Evaluation: Testing the Accuracy and Effectiveness of Inferred Knowledge 
• Knowledge Refinement: Updating and Improving Knowledge Representation 
• Knowledge Utilization: Applying Knowledge to Perform Tasks

[More to come ...]