The Generator of the Event Structure Lexicon (GESL): Automatic Annotation of Event Structure for Textual Inference Tasks.

摘要

This dissertation aims to develop the Generator of the Event Structure Lexicon (GESL) which is a tool to automate annotating the event structure of verbs in text to support textual inference tasks related to lexically entailed subevents. The output of the GESL is the Event Structure Lexicon (ESL), which is a lexicon of verbs in text which includes verb class, event structure, tense, aspect, and dependencies. The GESL consists of two parts. First, the generator of the basic ESL classifies verbs contextually to one of the pre-defined 23 verb classes and assigns their appropriate event structures. Second, the enricher of the basic ESL increases extractable entailments by lexical substitution of predicates in event structure. This thesis is unique and important in that it provides event structure-related entailments whose necessity has been recognized by very recent research.