Computer Science 456: Artificial Intelligence and Expert Systems
Unit 6: Knowledge Representation
This unit discusses the question of how to best capture and represent knowledge for use in computer-based problem solving. In addition to a brief historical review of early research in representation, this unit discusses the classical topics in knowledge representation: semantic networks, scripts, frames, and conceptual graphs. Current topics in knowledge representation deal with the creation non-centralized and non-explicit representational schemes. These topics are discussed through the presentation of distributed and cooperative models and agent-based systems.
Activities
Assignment 3
Complete Assignment 3, and submit it to your tutor for evaluation and feedback.
Section 6.1: Knowledge Representation
This section discusses the issues in knowledge representation and presents different representation models for creating centralized and explicit representational schemes. It discusses the sematic network model, the natural language scripts model, the frames model, and the conceptual graphs model.
Learning Objectives
- Present a brief historical review of early research in knowledge representation
- Describe in detail centralized and explicit representational models based on sematic networks and conceptual graphs.
- Present the concepts related to frames and scripts as tools for representing knowledge.
- Demonstrate how conceptual graphs can be used to express modal logic.
Key Terms
semantic network, case frame, conceptual dependency, scripts, frames, conceptual graphs
Readings
Read the sections “Issues in Knowledge Representation,” “A Brief History of AI Representational Systems,” and “Conceptual Graphs: a Network Language” from Chapter 7, Knowledge Representation.
Tasks
Practice the following exercises from Chapter 7 of the textbook:
- Exercises 3, 9, 10, and 17. You may want to use the Personal Workspace wiki on the course home page and/or share your observations with classmates in the COMP 456 General Discussion forum.
Section 6.2: Alternative Knowledge Representations and Agent-based Systems
This section presents alternatives for knowledge representations in intelligent systems. The concepts related to Brooks’ subsumption architecture, Copycat, and ontologies are discussed. Models for creating distributed and active representational schemes for agent-based systems are also introduced.
Learning Objectives
- Present alternative knowledge representation models such as the Brooks’ subsumption architecture and the Copycat architecture.
- Discuss the use of multiple representations and ontologies in intelligent systems.
- Introduce the agent-based model for knowledge representation.
Key Terms
Brooks’ subsumption architecture, Copycat architecture, ontology, knowledge service, distributed AI, agent-based systems
Readings
Read the sections “Alternative Representations and Ontologies” and “Agent Based and Distributed Problem Solving” from Chapter 7, Knowledge Representation.
Tasks
Practice the following exercises from Chapter 7 of the textbook:
- Exercises 19, 20, and 22. You may want to use the Personal Workspace wiki on the course home page and/or share your observations with classmates in the COMP 456 General Discussion forum.