A knowledge base's external schema formalizes its practical language, its conceptual schema specifies concepts, and its internal schema describes the organization of the knowledge base.
The following are the widely known knowledge representation schemes:
1. Semantic nets
2. Frames
3. Conceptual dependency
4. Scripts
1. Semantic Nets Knowledge Representation scheme in AI
Semantic net is a Knowledge representation scheme that captures knowledge as a graph. In artificial intelligence, semantic networks are graphical frameworks created to describe and arrange knowledge so that computers can understand and interpret data in a way that is understandable to humans. These networks are made up of links that indicate the connections between the nodes, which stand in for ideas or objects. Reasoning and data retrieval are made efficient by this structured style.
Links, link labels, and nodes make up a semantic net. Nodes, which stand in for actual objects, ideas, or circumstances, are represented by circles, ellipses, or even rectangles in these network diagrams. Links represent the relationships between things as arrows, while link labels provide details about those relationships.
Because the nodes in semantic nets are connected to one another, they are also known as associative nets.
Semantic Network Types
Semantic networks come in six varieties, each with a distinct value in terms of visual aids for automated reasoning and information management:
Definitional Network
Definitional networks show the connection between ideas and the subcategories of those concepts. Definitional networks are essential for guaranteeing accuracy and clarity in knowledge representation since they make it possible to distinguish exactly the relationships between various ideas within a certain domain.
Assertional Networks
They are employed for factual information transmission and proposition statements. These networks are essential to knowledge-based systems and databases because they are especially helpful for storing and exchanging structured data.
Implicational Networks
These networks emphasize cause-and-effect linkages by using implications as the main links between nodes. These networks are essential for scenario analysis, risk assessment, and predictive modeling because they highlight causal interconnections, which allow possible outcomes to be inferred from known associations.
Executable Networks
Have internal mechanisms that can be used to modify the network itself and enable dynamic adjustments. Because these networks may autonomously change their behavior or structure to enhance efficiency or react to new information, they are crucial for systems that need to make decisions in real-time.
Learning Networks
By stressing insights from instances and emphasizing adaptive learning, these broaden knowledge representations. Because they allow computers to continuously develop and update their understanding of the world through access to new information, learning networks are fundamental to machine learning and artificial intelligence applications.
Hybrid Networks
To meet various knowledge representation needs, this kind of network combines two or more of the methods mentioned above, either inside a single network or between closely interacting networks.
Semantic Network Components in Knowledge Representation Schemes
Semantic networks are architectures that effectively represent and process knowledge across a wide range of applications. These components come together to define their design. Semantic networks can be further categorized:
Lexical Component
Nodes: Within the network, these stand in for ideas or objects.
Links: They show the connections between the nodes.
Labels: Labels that attach to nodes and links provide context by identifying certain objects and relations.
Structural Component
A directed graph is created in this component by the combination of nodes and links. Nodes and links have labels carefully put on them to indicate their responsibilities in the network hierarchy.
Semantic Component
This part provides information for the labels and relationships connected to nodes. Knowledge representation and inference are made possible by the network's operation, which is guided by the interpretations of these meanings.
Procedural Part
Constructors enable the network to grow by allowing the creation of additional nodes and linkages. In contrast, destroyers make it easier to remove nodes and links, keeping the network flexible and current.
Advantages of Semantic Network in Artificial Intelligence
- The semantic network makes it possible to use graphical algorithms for successful inference.
- They are straightforward, simple to use, and easy to comprehend.
- Semantic networks are commonly used to establish connections between different domains of knowledge, such as computer science and anthropology.
- A straightforward method for exploring the problem space is made possible by the semantic network.
- The semantic network provides a method for creating branches out of related parts. The way that individuals process data is also impacted by the semantic network.
Disadvantages of Semantic Network in Artificial Intelligence
- Link names don't have a standard meaning.
- Semantic Nets depend on their author and lack intelligence.
- There is variation in the shape and purpose of links, as well as in the connections that declare relationships and structural ties.
- Distinct nodes that stand in for particular objects and classes Links within an object only show binary relationships.
Example of Semantic Network in Knowledge Representation
