Skip to main content
main-content

Über dieses Buch

The book offers a detailed guide to temporal ordering, exploring open problems in the field and providing solutions and extensive analysis. It addresses the challenge of automatically ordering events and times in text. Aided by TimeML, it also describes and presents concepts relating to time in easy-to-compute terms. Working out the order that events and times happen has proven difficult for computers, since the language used to discuss time can be vague and complex. Mapping out these concepts for a computational system, which does not have its own inherent idea of time, is, unsurprisingly, tough. Solving this problem enables powerful systems that can plan, reason about events, and construct stories of their own accord, as well as understand the complex narratives that humans express and comprehend so naturally.

This book presents a theory and data-driven analysis of temporal ordering, leading to the identification of exactly what is difficult about the task. It then proposes and evaluates machine-learning solutions for the major difficulties.

It is a valuable resource for those working in machine learning for natural language processing as well as anyone studying time in language, or involved in annotating the structure of time in documents.

Inhaltsverzeichnis

Frontmatter

Chapter 1. Introduction

Abstract
Humans developed natural language to communicate; over past millennia, it has been the most efficient form of transferring the majority of information between individuals. With the advent of computing, large amounts of natural language text are stored in digital format. The study of computational linguistics helps link the significant power of the computer with the efficiency of communicating in natural language.
Leon R. A. Derczynski

Chapter 2. Events and Times

Abstract
Time is a critical part of language. Without the ability to express it, we cannot plan, tell stories or discuss change.
Leon R. A. Derczynski

Chapter 3. Temporal Relations

Abstract
Having discussed timex and events in the previous chapter, we move on to discuss the temporal relations that exist between them.
Leon R. A. Derczynski

Chapter 4. Relation Labelling Analysis

Abstract
In Chap. 3, we discovered that automatic temporal relation typing is a difficult problem. This motivates an investigation into potential ways of improving performance in relation typing. This chapter details an attempt to discover potential ways of improving performance at the task. As humans are readily able to identify the nature of temporal links, one may a priori draw the conclusion that the information required to do so must be available somewhere. This knowledge is in a given document or in information known by the reader before encountering that document (referred to as world knowledge).
Leon R. A. Derczynski

Chapter 5. Using Temporal Signals

Abstract
In Chap. 4, we saw that a proportion of difficult temporal relations were associated with a particular separate word or phrase that described the temporal relation type – a temporal signal.
Leon R. A. Derczynski

Chapter 6. Using a Framework of Tense and Aspect

Abstract
This chapter investigates a linguistic framework for tense and aspect.
Leon R. A. Derczynski

Chapter 7. Conclusion

Abstract
Temporal annotation is difficult for both humans and machines. The task of determining how particular events are ordered or nested is part of this temporal annotation problem and has been the goal of this book. This is known as the temporal link labelling problem. The state of the art in this problem has advanced slowly in recent years, without reaching high enough performance levels to consider it solved. This book has investigated the problem of temporal link labelling.
Leon R. A. Derczynski

Backmatter

Weitere Informationen

Premium Partner

    Bildnachweise