自然言語解釈

現在スタンフォード大学のMacCartneyとManningが書いた論文を読書しています。
Natural Language Processing(NLP) is a discipline which designs to enable computers to generate and understand expressions in human languages. Natural Language Inference (NLI) attempts to derive a hypothesis from a premise, where the premise is expressed in actual human language(as opposed to formal logic). An example might be, given "No black cats visited our abode", conclude that the hypothesis "No cats were at our house" is not necessarily true.
The title of the paper is "Modeling Semantic Containment and Exclusion in Natural Language ". Semantic containment seems to allude to words which contain others. One example would be "clothing", which contains "trousers". Another example, using a quantifier, would be "every philosopher", which contains "a philosopher". Exclusion indicates that elements are members of a certain set. An example from the paper is "no delegate finished the report". This precludes the existence of "a delegate who completed the report".
It seems that the hypothesis is a contrived test case, as is the premise, and that the system under discussion is basically a decision engine which determines the validity of the hypothesis from the premise.