[工程科技]Representing Time Situation Calculus- branching_ constructed from hypothetical actions.pdfVIP

Modular-E : A Logic for Assimilating Unexpected Observations Along a time Line Antonis Kakas Loizos Michael Rob Miller (Cyprus) (Harvard) (UCL) part 1: Background and aims part 2: Problems with existing approaches part 3: Modular-E syntax and semantics part 4: Properties and extensions of Modular-E part 5: Ideas for future work Read more in LPNMR’05 and forthcomming AIJ papers 1 Historical Context The idea of using symbolic logic to model human reasoning goes back a long way ... • Before A.I. - Aristotle, Chrysippus, Descartes, Leibniz (a librarian!), de ¨ Morgan, Frege, Boole, Venn, Russell, Tarski, Godel, ... • A.I. - John McCarthy – “Programs with Common Sense” (1959). – Concepts such as “cause”, “effect” and “action” are essential for key human reasoning processes such as planning . – In 1969 McCarthy and Hayes invent the Situation Calculus. – Broadens out into the A.I. sub-field of Reasoning about Action. 2 Modeling Cause and Effect in Logic • Example Applications: – Automated planning - underpins much contemporary research. – Causal Modeling - e.g. simulation of complex biological systems. – Software Specification - e.g. of “event-driven” systems. • Issues and Problems: – The Frame Problem - reasoning about non-effects of actions. – The Ramification Problem - reasoning about indirect effects. – The Qualification Problem - reasoning about unexpected effects. 3 Event Calc