05 Knowledge Engineering in FOL

Question 1

Knowledge engineering

Answer 1

Knowledge engineering is the process of building a knowledge base (KB) for a domain. Carried out by knowledge engineers (KE) doing knowledge acquisition, often by interviewing domain experts. The KE investigates the domain, learns important concepts, and creates a formal representation of domain objects and relations.

Most KBs are special purpose, covering a specific domain in detail. Other KBs are general purpose, valid across many domains.

Question 2

Knowledge engineering vs. programming

Answer 2

Programming
• Choosing a programming language
• Writing a program of data and control structures
• Choosing a compiler for the language
• Running the program

Knowledge engineering
• Choosing a logic for knowledge representation
• Building a knowledge base of facts and rules
• Implementing the inference procedure
• Inferring new facts

Question 3

Knowledge engineering: Approach

Answer 3

Declarative approach
Main point is that knowledge engineering is declarative
• The knowledge engineer tells the system what is true
• The system knows how to infer new facts and solutions

Some advantages:

• More efficient/high-level development, less debugging
• The knowledge base can be (re-)used for other tasks
• The inference engine can be (re-)used for other knowledge bases

Question 4

Knowledge engineering process (FOL)

Answer 4

• *Knowledge engineering process (FOL)**
  1. Identify the task
  2. Assemble relevant knowledge
  3. Decide on a vocabulary of predicates, functions and concepts (ontology)
  4. Encode general domain knowledge (axioms)
  5. Encode specific problem instance
  6. Pose queries to inference engine - get answers
  7. Debug knowledge base

See Electronics circuits domain for examples