Is FOPC the best way to represent English for all NLP operations?
According to Wikipedia, First Order Predicate Calculus is also known as First-order logicand Predicate logic. I consider semantic representation to be an unsolved problem. Is FOPC not the basis of Semantic Web triples or tuples? I don’t know of any publicly available dialog system based primarily on semantic web technology….
It seems to me that machine learning could be applied to corpora containing semantic (web) metadata; but, I am unaware of any specific project doing this in the context of dialog systems. Strictly speaking, IBM Watson is not a dialog system, but does combine semantic (web) metadata in a kitchen sink approach involving additional hypothesis generation and ranking technologies.
See also my quick and dirty webpage: