Artificial Intelligence Logic Lecture Notes

Author: Li Xiaowu
Publisher:
Publish Date: 2005-09-01
Features: Artificial intelligence logic is the study of how intelligent agents process knowledge using logical methods and achievements, primarily focusing on the subject's common-sense reasoning. This type of reasoning is characterized by non-monotonicity and incomplete information. This lecture note introduces three main types of artificial intelligence logic: default logic, non-monotonic modal logic, and circumscription logic. The main idea of default logic is to add default reasoning rules to classical logic to characterize common-sense reasoning, and the resulting extension concept describes the subject's belief set and its changes. Non-monotonic modal logic studies the subject's cognitive state using epistemic operators such as "knowing" and "believing," and characterizes the subject's cognitive state using a stable set concept with reflective properties. The main idea of circumscription logic is to reasonably limit the predicates described by ordinary predicate logic, thereby reasonably limiting the extension of the class of references of such predicates to exclude anomalous objects. Therefore, circumscription logic provides a general formal method to minimize any selected predicate or formula composed of predicates, thereby intuitively characterizing the subject's common-sense reasoning ability. This lecture note is rich in content, broad in scope, rigorous in expression, and rich in examples and exercises, making it suitable as a textbook for undergraduate and graduate students in the field of logic at higher education institutions.