Basic dependency parsing in natural language inference

摘要

Parsing is the process of analyzing a sentence for it structure, content and meaning, this process uncover the structure, articulate the constituents and the relation between the constituents of the input sentence. This paper described the importance of parsing strategy in achieving entailment in natural language inference. Parsing is the basic task in processing natural language and it is also the basis for all natural language applications such as machine learning, question answering and information retrieval. We have used the parsing strategy in natural language inference to achieve entailment through an approach called normalization approach where entailment is achieved by removing or replacing some nodes as well as relations in a tree. This process requires a detailed understanding of the dependency structure, in order to generate a tree that does not contain nodes and relations that are irrelevant to the inference procedure. In order to achieve this, the dependency trees are transformed by applying some rewrite rules to the dependency tree.