Hiroyuki (Hiro) Uchida, Graduate Student


Hiroyuki (Hiro) Uchida, Graduate Student

Sage 4101

October 27, 2010 12:00 PM - 1:30 PM



 The theory that human cognition proceeds through mental simulations, if true, would provide an elegant explanation of how the mechanisms of reasoning and problem solving integrate with and develop from mechanisms underlying forms of cognition that occur earlier in evolution and development. However, common simulation mechanisms do not seem to be strong enough to model intelligent or general reasoning that humans can run. More specifically, it is not apparent how we can deal with the semantics of universally or existentially quantified natural language sentences by way of such a mental simulation schema. Contrary to such a general view, in this paper, we show that the universal force and the existential force as observed in natural language semantics and spontaneous inferences can be analyzed as emergent properties of a particular simulation schema using arbitrary objects. We then introduce a novel inference language that directly represents the internal structure of such a mental simulation. This inference language has several independent merits. Most importantly, this language is made out of atomic formulas only, configured into particular structured sets of formulas. This allows us to maintain the direct translatability between the logical language expressions and the other perceptual representations that we manipulate in our AI architecture. Also, quantifier-free language allows us to provide a uniform semantic analysis of natural language NP arguments. We show that crucial natural language data can be adequately analyzed in this inference language. Finally, we compare the expressive power of this novel language with those of Classical Propositional Logic (at the propositional level) and Predicate Calculus, two languages that are commonly used in the semantic analysis of natural language. We show that translation from our language into CPL and from CPL to our language is complete at the propositional level. We indicate that every PC formula can be translated into our language but some of the expressions in our language are not expressible in a common PC language.

AI Language without explicit quantifiers (updated)

Due to technical problems there is no video of this event
Add to calendar