Overview:

The purpose of PhraseNet is to build a context-sensitive lexical semantic knowledge system, which can help various Natural Language Processing tasks such as question answering and prepositional phrase attachment.

Research in natural language understanding necessitates significant progress in lexical semantics and the development of lexical semantics resources. In a broad range of natural language applications, from prepositional phrase attachment, co-reference resolution to question answering and summarization, semantic information is a necessary component in the inference, which provides a level of abstraction that is indispensable for robust decisions. Inducing that the prepositional phrase in They ate a cake with a fork is different from that in They ate a cake with strawberry, for example, depends on the knowledge that fork and strawberry have different hypernyms. However, fork as a noun has five senses listed in WordNet and each of them has a different hypernym. Choosing the correct one in each situation is a context sensitive decision.

Details:

PhraseNet is designed based on the assumption that, by and large, semantic ambiguity disappears when the local context of a word is taken into account. PhraseNet is generated automatically using WordNet and machine-learning-based processing of large English corpora. It enhances a WordNet synset with its contextual information and refines WordNet relational structure by maintaining only those links that respect contextual constraints. Specifically, contextual information is built hierarchically in PhraseNet. Each word list in the system has pointers referring to its corresponding contextual information. PhraseNet is an independent lexical semantic system allied with proper user interfaces and access functions that will allow researchers to access it and practitioners to use it in applications.