In this post are gathered the tutorials and exercises, called “mega-recitations”, to apply the concepts presented in the Artificial Intelligence course of Prof. Patrick Winston. Continue reading “24. Reviews and Exercises”
3. Reasoning: Goal Trees and Rule-Based Expert Systems
Reporting on a Goal Tree program
A Goal Tree program can answer questions about its own behavior by reporting steps up (why questions) or down (how questions) in the actions it takes.
Simon’s Ant
The complexity of a program’s behavior is a consequence of the environment, not the complexity of the program.
Complexity (behavior) = Max ( Complexity (Program) , Complexity (Environment) )
An ant’s apparently complex behavior stems from the obstacles in its environment to avoid on its way home.
Rule-Based Expert Systems
Rule-based “expert” systems are deductions systems, they can answer questions about their behaviors as they are a form of goal trees.
Forward-chaining rule-based expert systems work from characteristics to deduce a result by a set of rules. Backward-chaining rule-based expert systems verify a proposition is true by going back to the check its characteristics through the rules.
Heuristics of knowledge engineering
How to set the program’s rules:
- Deal with specific cases: know all details of each specific case, not vague, general ideas from people (potato chips, tomatoes vs “squishy”)
- Understand the vocabulary items that make two cases different (frozen vs canned)
- Analyze when the program breaks down to understand the missing rule.