This dissertation proposes an experimental study of focus in Spanish, investigating, in particular, if two types of focus – Contrastive focus and Non-contrastive focus – are syntactically and prosodically distinguished. The evidence that the conceptual distinction between the focus subtypes can be represented linguistically has been found in languages (Drubig 2003, É. Kiss 1998, Gundel & Fretheim 2001, Zubizarreta 1998 to name a few). As for Spanish, Zubizarreta (1998) argued that the two types of focus most noticeably differ syntactically. While Non-contrastive Focus should appear at utterance-final position, Contrastive Focus may appear in-situ. Nevertheless, not all the studies seem to accept Zubizarreta’s (1998) syntax-oriented distinction between the two focus types. A few studies suggest that not only Contrastive Focus but also Non-Contrastive Focus can indeed occur sentence-internally (Cabrera Abreu & García Lecumberri 2003, Kim & Avelino 2003, Toledo 1989). Inspired by a handful of studies and motivated by empirical data gathered for the pilot study, the current study sets out to investigate Zubizarreta’s (1998) syntax-oriented claim on the distinction between the focus subtypes. Focus in Spanish is known to be prosodically marked by its particular intonational contour- higher pitch and the early peak, and secondarily longer duration and/or higher intensity, compared to unfocused elements in a given utterance (Cabrera Abreu & García Lecumberri 2003, Domínguez 2004a & b, Face 2000, 2001, 2002b, Hualde 2003, 2005, Kim & Avelino 2003, de la Mota 1995, 1997, Navarro Tomás 1918, Nibert 2000, Quilis 1971, Sosa 1998, Toledo 1989, Zubizarreta 1998). We assume that the distinction between the two types of focus would also be made using the existing cues, as suggested by a handful of studies on focus types (Cabrera Abreu & García Lecumberri 2003, Kim & Avelino 2003, Zubizarreta, 1998). The findings of our experiments clearly indicate that Spanish speakers consistently use different phonetic and phonological cues such as duration and pitch in order to make a distinction between the two types of focus. These findings give clear evidence that the pragmatically defined notion of focus (Lambrecht 1994) is indeed further divided into two types in Castilian Spanish, somewhat similar to the distinction made in English (Selkirk 1984, 1995). / text

This dissertation addresses the task of learning to rank, both in the supervised and unsupervised settings, by exploiting the interplay of convex functions, monotonic mappings and their fixed points. In the supervised setting of learning to rank, one wishes to learn from examples of correctly ordered items whereas in the unsupervised setting, one tries to maximize some quantitatively defined characteristic of a "good" ranking. A ranking method selects one permutation from among the combinatorially many permutations defined on the items to rank. Accomplishing this optimally in the supervised setting, with minimal loss in generality, if any, is challenging. In this dissertation this problem is addressed by optimizing, globally and efficiently, a statistically consistent loss functional over the class of compositions of a linear function by an arbitrary, strictly monotonic, separable mapping with large margins. This capability also enables learning the parameters of a generalized linear model with an unknown link function. The method can handle infinite dimensional feature spaces if the corresponding kernel function is known. In the unsupervised setting, a popular ranking approach is is link analysis over a graph of recommendations, as exemplified by pagerank. This dissertation shows that pagerank may be viewed as an instance of an unsupervised consensus optimization problem. The dissertation then solves a more general problem of unsupervised consensus over noisy, directed recommendation graphs that have uncertainty over the set of "out" edges that emanate from a vertex. The proposed consensus rank is essentially the pagerank over the expected edge-set, where the expectation is computed over the distribution that achieves the most agreeable consensus. This consensus is measured geometrically by a suitable Bregman divergence between the consensus rank and the ranks induced by item specific distributions Real world deployed ranking methods need to be resistant to spam, a particularly sophisticated type of which is link-spam. A popular class of countermeasures "de-spam" the corrupted webgraph by removing abusive pages identified by supervised learning. Since exhaustive detection and neutralization is infeasible, there is a need for ranking functions that can, on one hand, attenuate the effects of link-spam without supervision and on the other hand, counter spam more aggressively when supervision is available. A family of non-linear, iteratively defined monotonic functions is proposed that propagates "rank" and "trust" scores through the webgraph. It relies on non-linearity, monotonicity and Schurconvexity to provide the resistance against spam. / text

This essay will examine the social organization in the archipelago of Tonga. To accomplish this purpose, archaic societies and structures such as ancient Egypt and Hawaii will be examined to obtain a closer understanding of the building blocks which constitute a chiefdom or kingdom. Since Patrick V. Kirch (2010) recently redefined the Hawaiian archipelago as a kingdom this essay will examine if the same can be done with the chiefdom of Tonga since it have similarities to the Hawaiian social organization.To do this the essay will examine and describe both the geographical organization of the states and the social organization. Social organization in this essay will be defined as an archaic state organization that requires several levels of administration to rule the society. / Kandidatuppsats

No description available.

One of the key features driving the growth and success of the Social Web is large-scale participation through user-contributed content – often through short text in social media. Unlike traditional long-form documents – e.g., Web pages, blog posts – these short text resources are typically quite brief (on the order of 100s of characters), often of a personal nature (reflecting opinions and reactions of users), and being generated at an explosive rate. Coupled with this explosion of short text in social media is the need for new methods to organize, monitor, and distill relevant information from these large-scale social systems, even in the face of the inherent “messiness” of short text, considering the wide variability in quality, style, and substance of short text generated by a legion of Social Web participants. Hence, this dissertation seeks to develop new algorithms and methods to ensure the continued growth of the Social Web by enhancing how users engage with short text in social media. Concretely, this dissertation takes a three-fold approach: First, this dissertation develops a learning-based algorithm to automatically rank short text comments associated with a Social Web object (e.g., Web document, image, video) based on the expressed preferences of the community itself, so that low-quality short text may be filtered and user attention may be focused on highly-ranked short text. Second, this dissertation organizes short text through labeling, via a graph- based framework for automatically assigning relevant labels to short text. In this way meaningful semantic descriptors may be assigned to short text for improved classification, browsing, and visualization. Third, this dissertation presents a cluster-based summarization approach for extracting high-quality viewpoints expressed in a collection of short text, while maintaining diverse viewpoints. By summarizing short text, user attention may quickly assess the aggregate viewpoints expressed in a collection of short text, without the need to scan each of possibly thousands of short text items.

Research Doctorate - Doctor of Philosophy (PhD) / There is a strong connection between directed graphs and the shifts of finite type which are an important family of dynamical systems. Higher-rank graphs (or k-graphs) and their C*-algebras were introduced by Kumjian and Pask to generalise directed graphs and their C*-algebras. Kumjian and Pask showed how higher-dimensional shifts of finite type can be associated to k-graphs, but did not discuss how one might associate k-graphs to k-dimensional shifts of finite type. In this thesis we construct a family of 2-graphs A arising from a certain type of algebraic two-dimensional shift of finite type studied by Schmidt, and analyse the structure of their C*-algebras. Graph algebras and k-graph algebras provide a rich source of examples for the classication of simple, purely infinite, nuclear C*-algebras. We give criteria which ensure that the C*-algebra C*(A) is simple, purely infinite, nuclear, and satisfies the hypotheses of the Kirchberg-Phillips Classification Theorem. We perform K-theory calculations for a wide range of our 2-graphs A using the Magma computational algebra system. The results of our calculations lead us to conjecture that the K-groups of C*(A) are finite cyclic groups of the same order. We are able to prove under mild hypotheses that the K-groups have the same order, but we have only numerical evidence to suggest that they are cyclic. In particular, we find several examples for which K1(C*(A)) is nonzero and has torsion, hence these are examples of 2-graph C*-algebras which do not arise as the C*-algebras of directed graphs. Finally, we consider a subfamily of 2-graphs with interesting combinatorial connections. We identify the nonsimple C*-algebras of these 2-graphs and calculate their K-theory.

Research Doctorate - Doctor of Philosophy (PhD) / Directed graphs are combinatorial objects used to model networks like fluid-flow systems in which the direction of movement through the network is important. In 1980, Enomoto and Watatani used finite directed graphs to provide an intuitive framework for the Cuntz-Krieger algebras introduced by Cuntz and Krieger earlier in the same year. The theory of the C*-algebras of directed graphs has since been extended to include infinite graphs, and there is an elegant relationship between connectivity and loops in a graph and the structure theory of the associated C*-algebra. Higher-rank graphs are a higher-dimensional analogue of directed graphs introduced by Kumjian and Pask in 2000 as a model for the higher-rank Cuntz-Krieger algebras introduced by Robertson and Steger in 1999. The theory of the Cuntz-Krieger algebras of higher-rank graphs is relatively new, and a number of questions which have been answered for directed graphs remain open in the higher-rank setting. In particular, for a large class of higher-rank graphs, the gauge-invariant ideal structure of the associated C*-algebra has not yet been identified. This thesis addresses the question of the gauge-invariant ideal structure of the Cuntz-Krieger algebras of higher-rank graphs. To do so, we introduce and analyse the collections of relative Cuntz-Krieger algebras associated to higher-rank graphs. The first two main results of the thesis are versions of the gauge-invariant uniqueness theorem and the Cuntz-Krieger uniqueness theorem which apply to relative Cuntz-Krieger algebras. Using these theorems, we are able to achieve our main goal, producing a classification of the gauge-invariant ideals in the Cuntz-Krieger algebra of a higher-rank graph analogous to that developed for directed graphs by Bates, Hong, Raeburn and Szymañski in 2002. We also demonstrate that relative Cuntz-Krieger algebras associated to higher-rank graphs are always nuclear, and produce conditions on a higher-rank graph under which the associated Cuntz-Krieger algebra is simple and purely infinite.

La mobilité existe de longue date dans la fonction publique mais demeure limitée par des obstacles relatifs à sa mise en oeuvre et à sa finalité. Depuis le début des années1990, l'amplification de la mobilité tente de dépasser certaines de ces considérations en procédant à une diffusion protéiforme des mouvements. En ce sens, la mobilité voulue se développe tant au sein de la fonction publique qu'entre celle-ci et le secteur privé. De même, la mobilité imposée connaît un net regain d'intérêt à travers les réorganisations administratives et le besoin d'une gestion plus souple de l'emploi public. Toutefois, qu'elle soit imposée ou voulue, la mobilité sert toujours les intérêts de l'administration en permettant sa modernisation. En effet, constituant une notion indéterminable, tant par des critères précis que par des effets donnés sur un agent, sa caractéristique principale est d'accompagner et parfois d'impulser le changement. Elle est ainsi un instrument perturbateur confortant l'hybridation du droit de la fonction publique et induisant le métissage de la fonction publique. / Mobility exists for a long time in the public service but remains limited due to obstacles relative to its implementation and to its purpose. Since the beginning of 1990s, the development of the mobility tries to exceed some of these considerations by proceeding to a protean distribution of the movements. By this way, the deliberate mobility develops as well as within the public service as between this one and the private sector. In the sameway, the compulsory mobility knows a real renewed interest through the administrative reorganizations and the need for a more flexible management of the public employment.However, whether it is compulsory or wanted, mobility always serves the interests of the administration by allowing its modernization. Indeed, constituting an indeterminable concept, as well by precise criteria as by effects given on an agent, its main characteristic is to accompany and sometimes to impulse the change. It is thus a disturbing instrument consolidating the hybridization of the law of the public service and leading to the interbreeding of the public service.

Using 12 years of monthly sweep-net data from 9-12 permanent sampling stations, we evaluated the variability and stability of the dragonfly assemblage in Bays Mountain Lake (northeastern Tennessee, USA). To do this, we adopted the view that a stable assemblage (i.e. one capable of recovering quickly from disturbances) should have low variability (i.e. high persistence of taxa, relatively constant densities, and high rank concordance), except with disturbances more intense and frequent than those in this system. Moreover, a stable assemblage should contain populations that exhibit density dependence and should tend to remain within a restricted range of densities (boundedness), shifting toward a narrow density interval between generations (attraction). To test some specific predictions derived from these views, we analyzed 12-year sequences of larval population sizes just before the onset of emergence for the 13 dominant dragonfly taxa in the lake. Most but not all of the 13 dominant taxa persisted during the 12-year period. Variabilities of taxon densities, measured as standard deviations across generations of log-transformed population sizes, were representative of the broad range for other invertebrates but somewhat higher than those of terrestrial vertebrates. There were fewer than three significant abundance trends over the 12-year period, and rank concordance between generations was high (W=0.716). Density dependence was detected among some of the dragonfly density sequences by five different methods. Using techniques presented in the companion paper, we found strong indications of both boundedness and attraction in the whole assemblage. We conclude tentatively that an assemblage consisting of at least 11 of the 13 dominant dragonfly taxa at Bays Mountain Lake has low-to-moderate variability and is stable, but that the complete 29-species assemblage is probably not stable at the scale of this single lake. We emphasize the need for coupling such long-term descriptive analyses with studies of responses to experimental disturbances.

Indiana University-Purdue University Indianapolis (IUPUI) / Web search has become a part of everyday life for hundreds of millions of users around the world. However, the effectiveness of a user's search depends vitally on the quality of search result ranking. Even though enormous efforts have been made to improve the ranking quality, there is still significant misalignment between search engine ranking and an end user's preference order. This is evident from the fact that, for many search results on major search and e-commerce platforms, many users ignore the top ranked results and click on the lower ranked results. Nevertheless, finding a ranking that suits all the users is a difficult problem to solve as every user's need is different. So, an ideal ranking is the one which is preferred by the majority of the users. This emphasizes the need for an automated approach which improves the search engine ranking dynamically by incorporating user clicks in the ranking algorithm. In existing search result ranking methodologies, this direction has not been explored profoundly. A key challenge in using user clicks in search result ranking is that the relevance feedback that is learnt from click data is imperfect. This is due to the fact that a user is more likely to click a top ranked result than a lower ranked result, irrespective of the actual relevance of those results. This phenomenon is known as position bias which poses a major difficulty in obtaining an automated method for dynamic update of search rank orders. In my thesis, I propose a set of methodologies which incorporate user clicks for dynamic update of search rank orders. The updates are based on adaptive randomization of results using reinforcement learning strategy by considering the user click activities as reinforcement signal. Beginning at any rank order of the search results, the proposed methodologies guaranty to converge to a ranking which is close to the ideal rank order. Besides, the usage of reinforcement learning strategy enables the proposed methods to overcome the position bias phenomenon. To measure the effectiveness of the proposed method, I perform experiments considering a simplified user behavior model which I call color ball abstraction model. I evaluate the quality of the proposed methodologies using standard information retrieval metrics like Precision at n (P@n), Kendall tau rank correlation, Discounted Cumulative Gain (DCG) and Normalized Discounted Cumulative Gain (NDCG). The experiment results clearly demonstrate the success of the proposed methodologies.