Discourse Processing here is framed as marking up a text with structural descriptions on several levels, which can serve to support many language-processing or text-mining tasks. We first explore some ways of assigning structure on the document level: the logical document structure as determined by the layout of the text, its genre-specific content structure, and its breakdown into topical segments. Then the focus moves to phenomena of local coherence. We introduce the problem of coreference and look at methods for building chains of coreferring entities in the text. Next, the notion of coherence relation is introduced as the second important factor of local coherence. We study the role of connectives and other means of signaling such relations in text, and then return to the level of larger textual units, where tree or graph structures can be ascribed by recursively assigning coherence relations. Taken together, these descriptions can inform text summarization, information extraction, discourse-aware sentiment analysis, question answering, and the like. Table of Contents: Introduction / Large Discourse Units and Topics / Coreference Resolution / Small Discourse Units and Coherence Relations / Summary: Text Structure on Multiple Interacting Levels Acknowledgments......Page 11 Introduction......Page 13 Genre-induced text structure......Page 19 Logical document structure......Page 21 Example: Scientific papers......Page 25 Example: Film reviews......Page 27 Introduction: ``What is this all about?''......Page 28 Exploiting surface cues......Page 31 Lexical chains......Page 34 Word distributions......Page 40 Probabilistic models of segmentation and topics......Page 46 Combining evidence......Page 48 Summary......Page 49 Coreference Resolution......Page 51 Reference and coreference: An overview......Page 52 Corpus annotation......Page 58 Entity-based local coherence......Page 63 Determining anaphoricity and familiarity status......Page 66 Matching proper names......Page 69 Pronoun resolution......Page 70 Resolving definite noun phrases......Page 73 Web-assisted resolution of `other'-anaphora......Page 75 Statistical approaches to coreference resolution......Page 77 Features......Page 78 Mention-pair models......Page 80 Alternative models......Page 83 Evaluation......Page 86 Summary......Page 87 Coherence relations......Page 91 Defining EDUs......Page 99 A subproblem: Attribution......Page 102 Automatic EDU segmentation......Page 104 Connectives: An introduction......Page 109 Identifying connectives......Page 113 Interpreting connectives......Page 115 Detecting implicit coherence relations......Page 122 Finding relations: The problem at large......Page 124 Coherence-relational text structure......Page 125 Trees......Page 126 Parsing coherence-relational trees......Page 129 Graphs......Page 135 Summary: Guessing or underspecifying?......Page 138 Summary: Text Structure on Multiple Interacting Levels......Page 141 Sample text......Page 145 Bibliography......Page 149 Author's Biography......Page 167 1. Introduction 2. Large discourse units and topics 2.1 Genre-induced text structure 2.1.1 Logical document structure 2.1.2 Content zones 2.1.3 Example: scientific papers 2.1.4 Example: film reviews 2.2 Topic-based segmentation 2.2.1 Introduction: "What is this all about?" 2.2.2 Exploiting surface cues 2.2.3 Lexical chains 2.2.4 Word distributions 2.2.5 Probabilistic models of segmentation and topics 2.2.6 Combining evidence 2.3 Summary 3. Coreference resolution 3.1 Reference and coreference: an overview 3.2 Corpus annotation 3.3 Entity-based local coherence 3.4 Determining anaphoricity and familiarity status 3.5 Rule-based methods for resolving nominal anaphora 3.5.1 Matching proper names 3.5.2 Pronoun resolution 3.5.3 Resolving definite noun phrases 3.5.4 Web-assisted resolution of 'Other'-anaphora 3.6 Statistical approaches to coreference resolution 3.6.1 Features 3.6.2 Mention-pair models 3.6.3 Alternative models 3.7 Evaluation 3.8 Summary 4. Small discourse units and coherence relations 4.1 Coherence relations 4.2 Segmentation: finding elementary discourse units 4.2.1 Defining EDUs 4.2.2 A subproblem: attribution 4.2.3 Automatic EDU segmentation 4.3 Recognizing coherence relations 4.3.1 Connectives: an introduction 4.3.2 Identifying connectives 4.3.3 Interpreting connectives 4.3.4 Detecting implicit coherence relations 4.3.5 Finding relations: the problem at large 4.4 Coherence-relational text structure 4.4.1 Trees 4.4.2 Parsing coherence-relational trees 4.4.3 Graphs 4.5 Summary: guessing or underspecifying? 5. Summary: text structure on multiple interacting levels A. Sample text Bibliography Author's biography. Discourse Processing here is framed as marking up a text with structural descriptions on several levels, which can serve to support many language-processing or text-mining tasks. This title explores some ways of assigning structure on the document level. The focus then moves to phenomena of local coherence, and the role of connectives and other means of signalling such relations in text.
