Thesis Proposal Verb Semantics for Natural Language Understanding

A verb is the organizational core of a sentence. Understanding the meaning of the verb is therefore key to understanding the meaning of the sentence. Natural language understanding is the problem of mapping natural language text to its meaning representation: entities and relations anchored to the world. Since verbs express relations over their arguments in text, a lexical resource about verbs can facilitate natural language understanding by mapping verbs in text to relations over entities expressed by their arguments in the world. In this thesis, we propose an automatic construction of a verb resource that contains important semantics for natural language understanding. We propose to learn three important semantics about verbs that will complement existing resources on verbs such as WordNet and VerbNet. The three semantics are (1) the mapping of verbs to relations in knowledge bases, (2) the preand post-conditions of each verb on its arguments: the entry condition of entities that allow an event expressed by the verb to take place and the condition of entities that will be true after the event occurs, (3) the temporal sequences of verbs. The mapping of verbs to relations in knowledge bases such as NELL, YAGO, or Freebase can provide a direct link between the text and the background knowledge about the world in the knowledge bases; enabling inferences over the world knowledge to better understand the text. The preand post-conditions of verbs on their arguments and the temporal sequences of verbs will open a range of new inference options for natural language understanding; for example for predicting events, for temporal scoping of events, for inferring the cascading effect of events, for inferring the states of entities in text, or for understanding the meaning of a sentence in the context of other sentences in the same document.