Issues in Knowledge Representation in AI with Example

Issues in Knowledge Representation

Having a great deal of knowledge and a way to manipulate it are essential for addressing problems. Using knowledge to support the derivation of conclusions is the fundamental purpose of knowledge representation. When numerous knowledge-based approaches are applied to a single topic, numerous problems could occur.


The following are a few problems that can occur while applying the knowledge representation technique:

  • Important attributes
  • Relationship between attributes
  • Choosing Granularity 
  • Set of objects
  • Finding the Right structures 

1. Important attributes

Is there a property of an object so fundamental that it appears in nearly every area of concern? Two general importance terms, "instance" and "is a," are credited. The fact that these qualities facilitate property inheritance makes them significant.

2. Relationship between attributes

The entities we represent are the attributes we utilize to characterize objects. Regardless of the particular knowledge they encode, an object's attributes may relate to one another to hold properties like:

  • Inverses: This refers to verifying consistency when one attribute receives an addition. There are numerous connections between the entities.
  • An "is a" hierarchy of attributes:
    • Being in an is a hierarchy involves generalization-specification; examples include classes of objects and specialized subsets within them, as well as attributes and attribute specialization. As an illustration, the attribute height is a subset of the generic attribute physical size, which is a subset of the physical attribute.
    • Because they facilitate inheritance, these generalization-specialization interactions are significant for characteristics.
  • Techniques for reasoning about values: This involves determining the values of characteristics that aren't mentioned directly. Reasoning uses various information types, such as height: which must be in a unit of length, and A person's age cannot be higher than their parents' age. When creating a knowledge base, the values are frequently supplied. 
  • Single value attributes: These are particular characteristics that are certain to have a single value. A baseball player, for instance, might only ever belong to one team and have one height. Different strategies are used by KR systems to support single-valued attributes.


3. Choosing Granularity

Granularity refers to the level of information that should be represented and the primitives. Do low-level primitive facts or high-level information need to be present in greater or lesser quantities? 

While low-level primitives might need a lot of storage, high-level truths might not need extra memory. One crucial component of knowledge representation is the level of granularity. It addresses the complexity of detail. The size of the knowledge base is determined by the granularity.

4. Set of objects

Certain characteristics of an object are true while it is part of a set but false when it is an individual;

The rationale behind representing sets of objects is that it is more efficient to associate a property once with the set rather than explicitly associate it with each and every element if it is true for all or most of the set's elements.


5. Finding the Right Structure

The appropriate structure is chosen to describe a particular situation. For this, a basic framework must be chosen, and then it must be revised or reviewed.

It's crucial to understand the following structure for this.

  • How to choose the best possible structure in the beginning.
  • How to add the necessary information based on the existing circumstances. 
  • How to choose a better structure in case the first one selected proves to be inappropriate. 
  • What should be done if none of the constructions are suitable?
  • When to design a new structure and when to forget it.

Post a Comment

0 Comments
* Please Don't Spam Here. All the Comments are Reviewed by Admin.

Top Post Ad

Below Post Ad