Artificial Intelligence Knowledge Representation and Reasoning 2 — Questions and Answers
Question 1: What distinguishes deductive reasoning from inductive reasoning in AI?
- Deductive reasoning draws specific conclusions that are guaranteed from general premises; inductive reasoning generalizes from specific observations (Correct answer)
- Inductive reasoning is always more accurate than deductive reasoning
- Deductive reasoning works only with probabilistic data
- Inductive reasoning requires complete knowledge of all facts
Correct answer: Deductive reasoning draws specific conclusions that are guaranteed from general premises; inductive reasoning generalizes from specific observations
Deductive reasoning moves from general rules to specific conclusions that are logically certain, while inductive reasoning infers general rules from specific examples, producing conclusions that are probable but not guaranteed.
Question 2: In forward chaining, an AI system:
- Starts from a goal and works backward to find supporting facts
- Starts from available facts and applies rules to derive new conclusions until the goal is reached (Correct answer)
- Randomly selects rules to apply until a solution is found
- Uses neural network layers to propagate activations forward
Correct answer: Starts from available facts and applies rules to derive new conclusions until the goal is reached
Forward chaining (data-driven reasoning) begins with known facts and fires applicable rules to generate new facts, continuing until the target goal is derived or no more rules can fire.
Question 3: Backward chaining is best described as:
- Processing data in reverse chronological order
- Starting from a goal hypothesis and working backward to find facts that support it (Correct answer)
- Training a neural network by reversing gradient flow
- Reversing the order of rules in a knowledge base
Correct answer: Starting from a goal hypothesis and working backward to find facts that support it
Backward chaining (goal-driven reasoning) begins with the goal and recursively identifies sub-goals or facts needed to prove it, making it efficient when the goal is known in advance.
Question 4: Fuzzy logic is used in AI primarily to:
- Handle binary true/false decisions more efficiently
- Represent and reason about degrees of truth in situations involving uncertainty or vagueness (Correct answer)
- Encrypt knowledge base entries for security
- Speed up gradient computation in neural networks
Correct answer: Represent and reason about degrees of truth in situations involving uncertainty or vagueness
Fuzzy logic extends classical binary logic by allowing truth values between 0 and 1, making it suitable for reasoning about imprecise or vague concepts like 'tall', 'warm', or 'fast'.
Question 5: Which of the following best describes a production rule system?
- A system that generates images using generative adversarial networks
- A rule-based system using IF-THEN rules to represent knowledge and trigger actions (Correct answer)
- A database system that handles CRUD operations for AI data
- A scheduling algorithm for parallel model training jobs
Correct answer: A rule-based system using IF-THEN rules to represent knowledge and trigger actions
A production rule system represents knowledge as a collection of IF (condition) THEN (action) rules; the inference engine matches conditions against working memory and fires matching rules.
Question 6: What is the Closed World Assumption (CWA) in knowledge representation?
- Any fact not known to be true is assumed to be false (Correct answer)
- All knowledge in the knowledge base is encrypted and inaccessible
- The system can only reason about a fixed, finite set of objects
- Unknown facts are assigned a probability of 0.5
Correct answer: Any fact not known to be true is assumed to be false
Under the CWA, if a fact cannot be proven true from the knowledge base, it is assumed to be false — this is typical in databases and logic programming (e.g., Prolog).
Question 7: The Open World Assumption (OWA), used in ontologies like OWL, means that:
- All facts in the world are already stored in the knowledge base
- A fact not currently known may still be true — lack of information does not imply falsehood (Correct answer)
- The knowledge base is publicly accessible on the internet
- The system treats all unknown facts as equally likely
Correct answer: A fact not currently known may still be true — lack of information does not imply falsehood
Under the OWA, the absence of a fact from the knowledge base does not mean the fact is false — it may simply be unknown, which is appropriate for open-ended, evolving domains like the Semantic Web.
Artificial Intelligence Knowledge Representation and Reasoning 2
What distinguishes deductive reasoning from inductive reasoning in AI?