AI · Theory

Inference in First-Order Logic

Theory

37/215

Inference in First-Order Logic

FOL inference: unification + lifting of propositional methods, resolution, forward/backward chaining on Horn clauses. Undecidable in the general case — say that. For many expert-system rules, chaining is enough.

Don’t claim a magic complete fast algorithm for all FOL. Honesty scores.

Output — idea / do / check for Inference in First-Order Logic. Fill those three in the viva.

Exam tip

Name two inference methods + one limit.

Example

# Inference in First-Order Logic
print("idea :", "Inference in First-Order Logic")
print("do   : one tiny example on paper")
print("check: one failure case")

Output: idea / do / check for Inference in First-Order Logic. Fill those three in the viva.

Short notes

DefUnification, resolution, chaining.
RuleGeneral FOL can be undecidable.

AI · Theory

Inference in First-Order Logic

← All stacks

Theory

37/215

Inference in First-Order Logic

Don’t claim a magic complete fast algorithm for all FOL. Honesty scores.

Output — idea / do / check for Inference in First-Order Logic. Fill those three in the viva.

Exam tip

Name two inference methods + one limit.

Example

# Inference in First-Order Logic
print("idea :", "Inference in First-Order Logic")
print("do   : one tiny example on paper")
print("check: one failure case")

Output: idea / do / check for Inference in First-Order Logic. Fill those three in the viva.

Short notes

DefUnification, resolution, chaining.
RuleGeneral FOL can be undecidable.

Inference in First-Order Logic

Example

Short notes

Questions

Inference in First-Order Logic

Example

Short notes

Questions