This thesis investigates automatic geolocation (i.e. identification of the location, expressed as latitude/longitude coordinates) of documents. Geolocation can be an effective means of summarizing large document collections and is an important component of geographic information retrieval. We describe several simple supervised methods for document geolocation using only the document’s raw text as evidence. All of our methods predict locations in the context of geodesic grids of varying degrees of resolution. We evaluate the methods on geotagged Wikipedia articles and Twitter feeds. For Wikipedia, our best method obtains a median prediction error of just 11.8 kilometers. Twitter geolocation is more challenging: we obtain a median error of 479 km, an improvement on previous results for the dataset. / text

In this dissertation, we introduce a graph-based model of instance-based, usage meaning that is cast as a problem of probabilistic inference. The main aim of this model is to provide a flexible platform that can be used to explore multiple hypotheses about usage meaning computation. Our model takes up and extends the proposals of Erk and Pado [2007] and McCarthy and Navigli [2009] by representing usage meaning as a probability distribution over potential paraphrases. We use undirected graphical models to infer this probability distribution for every content word in a given sentence. Graphical models represent complex probability distributions through a graph. In the graph, nodes stand for random variables, and edges stand for direct probabilistic interactions between them. The lack of edges between any two variables reflect independence assumptions. In our model, we represent each content word of the sentence through two adjacent nodes: the observed node represents the surface form of the word itself, and the hidden node represents its usage meaning. The distribution over values that we infer for the hidden node is a paraphrase distribution for the observed word. To encode the fact that lexical semantic information is exchanged between syntactic neighbors, the graph contains edges that mirror the dependency graph for the sentence. Further knowledge sources that influence the hidden nodes are represented through additional edges that, for example, connect to document topic. The integration of adjacent knowledge sources is accomplished in a standard way by multiplying factors and marginalizing over variables. Evaluating on a paraphrasing task, we find that our model outperforms the current state-of-the-art usage vector model [Thater et al., 2010] on all parts of speech except verbs, where the previous model wins by a small margin. But our main focus is not on the numbers but on the fact that our model is flexible enough to encode different hypotheses about usage meaning computation. In particular, we concentrate on five questions (with minor variants): - Nonlocal syntactic context: Existing usage vector models only use a word's direct syntactic neighbors for disambiguation or inferring some other meaning representation. Would it help to have contextual information instead "flow" along the entire dependency graph, each word's inferred meaning relying on the paraphrase distribution of its neighbors? - Influence of collocational information: In some cases, it is intuitively plausible to use the selectional preference of a neighboring word towards the target to determine its meaning in context. How does modeling selectional preferences into the model affect performance? - Non-syntactic bag-of-words context: To what extent can non-syntactic information in the form of bag-of-words context help in inferring meaning? - Effects of parametrization: We experiment with two transformations of MLE. One interpolates various MLEs and another transforms it by exponentiating pointwise mutual information. Which performs better? - Type of hidden nodes: Our model posits a tier of hidden nodes immediately adjacent the surface tier of observed words to capture dynamic usage meaning. We examine the model based on by varying the hidden nodes such that in one the nodes have actual words as values and in the other the nodes have nameless indexes as values. The former has the benefit of interpretability while the latter allows more standard parameter estimation. Portions of this dissertation are derived from joint work between the author and Katrin Erk [submitted]. / text

To gain greater insight into the operation of online social networks, we applied Natural Language Processing (NLP) techniques to text-based communication to identify and describe underlying social structures in online communities. This paper presents our approach and preliminary evaluation for content-based, automated discovery of social networks. Our research question is: What syntactic and semantic features of postings in a threaded discussions help uncover explicit and implicit ties between network members, and which provide a reliable estimate of the strengths of interpersonal ties among the network members? To evaluate our automated procedures, we compare the results from the NLP processes with social networks built from basic who-to-whom data, and a sample of hand-coded data derived from a close reading of the text. For our test case, and as part of ongoing research on networked learning, we used the archive of threaded discussions collected over eight iterations of an online graduate class. We first associate personal names and nicknames mentioned in the postings with class participants. Next we analyze the context in which each name occurs in the postings to determine whether or not there is an interpersonal tie between a sender of the posting and a person mentioned in it. Because information exchange is a key factor in the operation and success of a learning community, we estimate and assign weights to the ties by measuring the amount of information exchanged between each pair of the nodes; information in this case is operationalized as counts of important concept terms in the postings as derived through the NLP analyses. Finally, we compare the resulting network(s) against those derived from other means, including basic who-to-whom data derived from posting sequences (e.g., whose postings follow whose). In this comparison we evaluate what is gained in understanding network processes by our more elaborate analyses.

The role of the lexicon has been ignored or minimized in most work on computational stylistics. This research is an effort to fill that gap, demonstrating the key role that the lexicon plays in stylistic variation. In doing so, I bring together a number of diverse perspectives, including aesthetic, functional, and sociological aspects of style. The first major contribution of the thesis is the creation of aesthetic stylistic lexical resources from large mixed-register corpora, adapting statistical techniques from approaches to topic and sentiment analysis. A key novelty of the work is that I consider multiple correlated styles in a single model. Next, I consider a variety of tasks that are relevant to style, in particular tasks relevant to genre and demographic variables, showing that the use of lexical resources compares well to more traditional approaches, in some cases offering information that is simply not available to a system based on surface features. Finally, I focus in on a single stylistic task, Native Language Identification (NLI), offering a novel method for deriving lexical information from native language texts, and using a cross-corpus supervised approach to show definitively that lexical features are key to high performance on this task.

The role of the lexicon has been ignored or minimized in most work on computational stylistics. This research is an effort to fill that gap, demonstrating the key role that the lexicon plays in stylistic variation. In doing so, I bring together a number of diverse perspectives, including aesthetic, functional, and sociological aspects of style. The first major contribution of the thesis is the creation of aesthetic stylistic lexical resources from large mixed-register corpora, adapting statistical techniques from approaches to topic and sentiment analysis. A key novelty of the work is that I consider multiple correlated styles in a single model. Next, I consider a variety of tasks that are relevant to style, in particular tasks relevant to genre and demographic variables, showing that the use of lexical resources compares well to more traditional approaches, in some cases offering information that is simply not available to a system based on surface features. Finally, I focus in on a single stylistic task, Native Language Identification (NLI), offering a novel method for deriving lexical information from native language texts, and using a cross-corpus supervised approach to show definitively that lexical features are key to high performance on this task.

This work addresses issues in the automatic preprocessing of historical German input text for use by conventional natural language processing techniques. Conventional techniques cannot adequately account for historical input text due to conventional tools' reliance on a fixed application-specific lexicon keyed by contemporary orthographic surface form on the one hand, and the lack of consistent orthographic conventions in historical input text on the other. Historical spelling variation is treated here as an error-correction problem or "canonicalization" task: an attempt to automatically assign each (historical) input word a unique extant canonical cognate, thus allowing direct application-specific processing (tagging, parsing, etc.) of the returned canonical forms without need for any additional application-specific modifications. In the course of the work, various methods for automatic canonicalization are investigated and empirically evaluated, including conflation by phonetic identity, conflation by lemma instantiation heuristics, canonicalization by weighted finite-state rewrite cascade, and token-wise disambiguation by a dynamic Hidden Markov Model. / Diese Arbeit behandelt Themen der automatischen Vorverarbeitung historischen deutschen Textes für die Weiterverarbeitung durch konventionelle computerlinguistische Techniken. Konventionelle Techniken können historischen Text wegen des hohen Grads an graphematischer Variation in solchem Text ohne eine solche Vorverarbeitung nicht zufriedenstellend behandeln. Variation in der historischen Rechtschreibung wird hier als Fehlerkorrekturproblem oder "Kanonikalisierungsaufgabe" behandelt: ein Versuch, jedem (historischen) Eingabewort eine eindeutige extante Äquivalente zuzuordnen; so können konventionelle Techniken ohne weitere Modifikation direkt auf den gelieferten kanonischen Formen arbeiten. Verschiedene Methoden zur automatischen Kanonikalisierung werden im Rahmen dieser Arbeit untersucht, unter anderem Konflation durch phonetische Identität, Konflation durch Lemma-Instanziierungsheuristiken, Kanonikalisierung durch eine Kaskade gewichteter endlicher Transduktoren, und Disambiguiierung von Konflationskandidaten durch ein dynamisches Hidden Markov Modell.

This thesis studies disfluency in spontaneous Swedish speech, i.e., the occurrence of hesitation phenomena like eh, öh, truncated words, repetitions and repairs, mispronunciations, truncated words and so on. The thesis is divided into three parts: PART I provides the background, both concerning scientific, personal and industrial–academic aspects in the Tuning in quotes, and the Preamble and Introduction (chapter 1). PART II consists of one chapter only, chapter 2, which dives into the etiology of disfluency. Consequently it describes previous research on disfluencies, also including areas that are not the main focus of the present tome, like stuttering, psychotherapy, philosophy, neurology, discourse perspectives, speech production, application-driven perspectives, cognitive aspects, and so on. A discussion on terminology and definitions is also provided. The goal of this chapter is to provide as broad a picture as possible of the phenomenon of disfluency, and how all those different and varying perspectives are related to each other. PART III describes the linguistic data studied and analyzed in this thesis, with the following structure: Chapter 3 describes how the speech data were collected, and for what reason. Sum totals of the data and the post-processing method are also described. Chapter 4 describes how the data were transcribed, annotated and analyzed. The labeling method is described in detail, as is the method employed to do frequency counts. Chapter 5 presents the analysis and results for all different categories of disfluencies. Besides general frequency and distribution of the different types of disfluencies, both inter- and intra-corpus results are presented, as are co-occurrences of different types of disfluencies. Also, inter- and intra-speaker differences are discussed. Chapter 6 discusses the results, mainly in light of previous research. Reasons for the observed frequencies and distribution are proposed, as are their relation to language typology, as well as syntactic, morphological and phonetic reasons for the observed phenomena. Future work is also envisaged, both work that is possible on the present data set, work that is possible on the present data set given extended labeling and work that I think should be carried out, but where the present data set fails, in one way or another, to meet the requirements of such studies. Appendices 1–4 list the sum total of all data analyzed in this thesis (apart from Tok Pisin data). Appendix 5 provides an example of a full human–computer dialogue. / The electronic version of the printed dissertation is a corrected version where typos as well as phrases have been corrected. A list with the corrections is presented in the errata list above.

A digital library is a collection of services and information objects for storing, accessing, and retrieving digital objects. Automatic text summarization presents salient information in a condensed form suitable for user needs. This thesis amalgamates digital libraries and automatic text summarization by extending the Greenstone Digital Library software suite to include the University of Lethbridge Summarizer. The tool generates summaries, nouns, and non phrases for use as metadata for searching and browsing digital collections. Digital collections of newspapers, PDFs, and eBooks were created with summary metadata. PDF documents were processed the fastest at 1.8 MB/hr, followed by the newspapers at 1.3 MB/hr, with eBooks being the slowest at 0.9 MV/hr. Qualitative analysis on four genres: newspaper, M.Sc. thesis, novel, and poetry, revealed narrative newspapers were most suitable for automatically generated summarization. The other genres suffered from incoherence and information loss. Overall, summaries for digital collections are suitable when used with newspaper documents and unsuitable for other genres. / xiii, 142 leaves ; 28 cm.

Sentiment analysis is a challenging task today for computational linguistics. Because of the rise of the social Web, both the research and the industry are interested in automatic processing of opinions in text. In this work, we assume a multilingual and multidomain environment and aim at automatic and adaptive polarity classification.We propose a method for automatic construction of multilingual affective lexicons from microblogging to cover the lack of lexical resources. To test our method, we have collected over 2 million messages from Twitter, the largest microblogging platform, and have constructed affective resources in English, French, Spanish, and Chinese.We propose a text representation model based on dependency parse trees to replace a traditional n-grams model. In our model, we use dependency triples to form n-gram like features. We believe this representation covers the loss of information when assuming independence of words in the bag-of-words approach.Finally, we investigate the impact of entity-specific features on classification of minor opinions and propose normalization schemes for improving polarity classification. The proposed normalization schemes gives more weight to terms expressing sentiments and lower the importance of noisy features.The effectiveness of our approach has been proved in experimental evaluations that we have performed across multiple domains (movies, product reviews, news, blog posts) and multiple languages (English, French, Russian, Spanish, Chinese) including official participation in several international evaluation campaigns (SemEval'10, ROMIP'11, I2B2'11).

This work investigates the use of formal logic as a practical tool for describing the syntax and semantics of a subset of English, and building a computer program to answer data base queries expressed in that subset. To achieve an intimate connection between logical descriptions and computer programs, all the descriptions given are in the definite clause subset of the predicate calculus, which is the basis of the programming language Prolog. The logical descriptions run directly as efficient Prolog programs. Three aspects of the use of logic in natural language analysis are covered: formal representation of syntactic rules by means of a grammar formalism based on logic, extraposition grammars;. formal semantics for the chosen English subset, appropriate for data base queries; informal semantic and pragmatic rules to translate analysed sentences into their formal semantics. On these three aspects, the work improves and extends earlier work by Colmerauer and others, where the use of computational logic in language analysis was first introduced.