Дисертації з теми "Representation of words"

This thesis examines the various approaches which have been taken to investigate the concrete/abstract word distinction both in normal subjects and in patients who, as a result of brain damage, have an impairment of lexical semantic representations. The nature of the definition task as a tool for assessing the semantic representations of concrete and abstract terms was examined. It was found that definitions for abstract words differed from those of concrete words only in style, not in semantic content. The metalinguistic demands of the definition task therefore make it inappropriate for assessing the semantic representations of concrete and abstract terms in patients with any form of language impairment. The performance of four patients with semantic impairments was examined using a variety of tasks designed to assess concrete and abstract word comprehension. While some of the data can be accommodated within the framework of several theories, no single theory can adequately account for the patterns of performance in all four patients. An alternative model of semantic memory is therefore proposed in which concreteness and frequency interact at the semantic level. Jones' Ease of Predication Hypothesis, which states that the difference between concrete and abstract terms can be explained in terms of disproportionate numbers of underlying semantic features (or "predicates") was also investigated. It was found that the ease of predication variable does not accurately reflect either predicate or feature distributions, and is simply another index of concreteness. As such, the validity of this concept as the basis of theories of semantic representation should be questioned. Models based on the assumption of a "richer" semantic representation for concrete words (e.g.: Plaut & Shallice, 1993) are therefore undermined by these data. The possibility that concrete and abstract concepts can be accessed from their most salient predicates and/or features was examined in a series of semantic priming experiments. It was concluded that it is not possible to prime either concrete or abstract concepts from their constituent parts. Significant facilitation only occurred for items in which the prime and target were synonymous and therefore map onto concepts which share almost identical semantic representations. In summary, it is apparent that no current theory of semantic representation can adequately account for the range of findings with regard to the concrete/abstract word distinction. The most plausible account is some form of distributed connectionist model. However, such models are based on unsubstantiated assumptions about the nature of abstract word representations in the semantic network. Alternative proposals are therefore discussed.

The representation of a Stuttgart court festival in a fascinating book of prints has received no art historical attention. The cultural production of German lands in a complex and obscure time described by one historian as being particularly bereft of "textbook facts", has not elicited much scholarly interest. In the seventeenth century before confessional disputes within the Holy Roman Empire of the German Nation turned into armed conflict, small German territorial courts modelled themselves on and assumed the courtly style of the larger European courts. The Stuttgart baptism of 1616 presents an interesting case study of the use of a courtly spectacle by a secondary court at a time of great instability. The baptism festival served as a stage to display an alliance of some German Protestant princes that held a promise of international support for the Protestant cause. The Wurttemberg court commissioned lengthy texts and a large number of engravings to represent the event. This study will address the contributions made by printed images to the festival program. The key documents for this study are the texts which complement and at times diverge from the visual representation. The differences between the visual and textual material will serve to locate the function of the visual representation of a festival held at a time of impending conflict. The triumphal procession format of the engravings discloses a strategy of disenfranchisement of a powerful parliament while it serves to assert the rank of the court within and outside the German empire. The complex amalgams of imagery that are interspersed in the paper procession allude, I suggest, to the problems presented to the Wurttemberg court by an uneasy alliance of Protestant courts within the empire. The engravings served to encode references to problematic issues such as the survival of the Holy Roman Empire, the rights of Protestant territorial princes to form an alliance and the hopes for outside help for the Protestant cause.
Arts, Faculty of
Art History, Visual Art and Theory, Department of
Graduate

Most research in word recognition uses words that already exist in the reader's lexicon, and it is therefore of interest to see whether newly learned words are represented and processed in the same way as already known words. For example, are newly learned words immediately represented in a special form of lexical memory, or is there a gradual process of assimilation? As for L2 language learners, are newly learned words incorporated into the same processing system that serves L1, or are they represented quite independently?The current study examines this issue by testing for the existence of the Prime Lexicality Effect (PLE) observed in masked priming experiments (Forster & Veres, 1998). Strong form priming was found with nonword primes (e.g., contrapt-CONTRACT), but not with word primes (e.g., contrast-CONTRACT). This effect is generally assumed to result from competition between the prime and the target. So if the readers had been trained to treat "contrapt" as a new word, would it now function like a word and produce much weaker priming? Elgort (2007) demonstrated such an effect with unmasked primes with L2 bilinguals. The current study investigates the PLE in both L1 and L2 bilinguals under different training conditions. When the training program involves mere familiarization (learning to type the words), a PLE was found with visible primes, but not with masked primes, which suggests that unmasked PLE is not the best indicator of lexicalization. In the case of "real" acquisition where the new word is given a definition and a picture of the object it refers to, and learning is spread over two weeks, a clear PLE was obtained. However, when the same experiment was carried out on Chinese-English bilinguals using the same English materials, completely opposite results were obtained. The learning enhanced priming, rather than reducing it, suggesting that the L2 lexicon might differ qualitatively from the L1 lexicon. The implications of these results for competitive theories of lexical access are discussed, and alternative explanations are considered.

The focus of this study is revelation at the limits of communication. It considers the way in which (biblical) apocalyptic literature prominently figures the interconnection between liminality, revelation, and representation. The methodology asserts an indissoluble association between theology, philosophy and literature. As such it is interdisciplinary. A preliminary theory (and theology) of liminality interweaves the theological and philosophical contributions of, amongst others, Karl Barth, Graham Ward, Jürgen Moltmann and Jacques Derrida, thereby initiating a revised perspective on the constitution of literary apocalyptic text production and interpretation. Theorising the limen begins to describe the Trinitarian economy at work in Christian apocalyptic processing of scripture. I begin with the idea that revelation (apokalypsis) is the experience of the limen itself (in a coincidence of opposites). Thus the limen (as an actively divine space) incorporates that which stands on both sides, in vertical and horizontal, linear and cyclical, spatial and temporal movements. I then propose that apocalyptic literature re-presents this complex economy in which the end is rehearsed simultaneously as limit, threshold, and rupture. Theologically, this complicates inter-relational notions of ‘apocalyptic’ and eschatology, and stimulates a debate on a metaphysics of violence in communication (between God, man and Creation). I conclude that, at the extreme limit of human understanding (where words fail), those with faith in God’s love are opened out to revelation in the apocalyptic textual performance of the liminal economy, and thus to hope and forgiveness. Stressing the importance of reading apocalyptically, I begin to demonstrate the relationship between Christian-canonical narratives and the broader western literary canon, the critical process having invited an exploration of those literary characteristics (of tone, mode and genre) shared by (biblical, modern and postmodern) texts. An important principle in the literary analyses is the association between apocalyptic text production and hermeneutics. Christopher Rowland’s description of a ‘visionary mode’ explains how this process works. Thus the preliminary theory leads into a close reading of recent Russian and American works by Mikhail Bulgakov and Thomas Pynchon. These are compared to, and worked through, Mark’s and John’s gospels and the Book of Revelation. The interpretative approach widens the often self-limiting study of apocalyptic literature, and broadens theological debate on revelation. Thus it begins to show how the rhetoric of apocalyptic makes belief compelling.

In two studies, the author attempts to develop a general symbol theoretical approach to knowledge representation. The first study, Modeling change - the frame problem, critically examines the - so far unsuccessful - attempts to solve the notorious frame problem. By discussing and analyzing a number of related problems - the prediction problem, the revision problem, the qualification problem, and the book-keeping problem - the frame problem is distinguished as the problem of finding a representational form permitting a changing, complex world to be efficiently and adequately represented. This form, it is argued, is dictated by the metaphysics of the problem world, the fundamental form of the symbol system we humans use in rightly characterizing the world. In the second study, Pictures and words, the symbol theoretical approach is made more explicit. The subject Is the distinction between pictorial (non-linguistic, non-propositional, analogical, "direct") representation and verbal (linguistic, propositional) representation, and the further implications of this distinction. The study focuses on pictorial representation, which has received little attention compared to verbal representation. Observations, ideas, and theories in AI, cognitive psychology, and philosophy are critically examined. The general conclusion is that there is as yet no cogent and mature theory of pictorial representation that gives good support to computer applications. The philosophical symbol theory of Nelson Goodman is found to be the most thoroughly developed and most congenial with the aims and methods of AI. Goodman's theory of pictorial representation, however, in effect excludes computers from the use of pictures. In the final chapter, an attempt is made to develop Goodman's analysis of pictures further turning it into a theory useful to AI. The theory outlined builds on Goodman's concept of exemplification. The key idea is that a picture is a model of a description that has the depicted object as its standard model. One consequence Is that pictorial and verbal forms of representation are seen less as competing alternatives than as complementary forms of representation mutually supporting and depending on each other.
digitalisering@umu

The present thesis addresses machine learning in a domain of naturallanguage phrases that are names of universities. It describes two approaches to this problem and a software implementation that has made it possible to evaluate them and to compare them. In general terms, the system's task is to learn to 'understand' the significance of the various components of a university name, such as the city or region where the university is located, the scienti c disciplines that are studied there, or the name of a famous person which may be part of the university name. A concrete test for whether the system has acquired this understanding is when it is able to compose a plausible university name given some components that should occur in the name. In order to achieve this capability, our system learns the structure of available names of some universities in a given data set, i.e. it acquires a grammar for the microlanguage of university names. One of the challenges is that the system may encounter ambiguities due to multi meaning words. This problem is addressed using a small ontology that is created during the training phase. Both domain knowledge and grammatical knowledge is represented using decision trees, which is an ecient method for concept learning. Besides for inductive inference, their role is to partition the data set into a hierarchical structure which is used for resolving ambiguities. The present report also de nes some modi cations in the de nitions of parameters, for example a parameter for entropy, which enable the system to deal with cognitive uncertainties. Our method for automatic syntax acquisition, ADIOS, is an unsupervised learning method. This method is described and discussed here, including a report on the outcome of the tests using our data set. The software that has been implemented and used in this project has been implemented in C.

This research studies the phonological structure and representation of first words in Hijazi Arabic. It investigates the representational nature of early words and the developmental stages of their syllable and word internal structure within the framework of the Prosodic Theory (McCarthy & Prince 1986, 1990). The issues raised relate to the relation between child and adult phonology, and whether the subjects follow a universal path or influenced by their language-specific phonology (Ferguson & Farewell, 1975; Vihman, 1991; Fikkert, 1994; Demuth, 1995; Ota, 2003; Lleo, 2006). The discussion has been accompanied by considering child-adult differences and cross-linguistic comparisons between child Arabic and child Germanic languages (English, Dutch), child Romance languages (French, Spanish, and Catalan), and child Japanese. Cross-sectional spontaneous data were collected from twenty two monolingual children (aged from 1;0 to1;9) living in Jeddah, Saudi Arabia by recording their speech using the object-naming technique in near natural settings and analyzed using a qualitative approach. The results show that the subjects’ segmental inventories are very limited at the onset of speech, which determines the shape of their early words. The subjects often employ phonological processes (e.g. reduplication, consonant harmony, substitution, truncation) when their templates cannot accommodate all the segmental material of the target words. The data provide evidence that the phonological structure of their early words has the same organizational units of adult phonology and governed by its prosodic principles. The subjects go through similar stages of prosodic word development to those reported in the literature: a minimal word stage (1;0-1;6), where their outputs display bimoraic and disyllabic forms, followed by a maximal stage (between 1;7-1;9), where more complex structures are produced. Despite the universality of many aspects of word acquisition in child Arabic, the study emphasizes the importance of investigating the impact of the ambient language and the role of language specific phonologies.

This thesis focuses on the ways in which cloth and clothwork are represented in Old French romance in order to highlight how they relate to feminine voice, expression, and identity. By focusing mainly on medieval romance from the twelfth and thirteenth centuries, the field of research is narrowed to a period in which vernacular literature was redefining literacy. On the basis that literacy is not confined to the ability to read and write in Latin, clothwork is presented as a medium of literate expression, that being a form of readable knowledge or communication not codified in written word or language, and in the works of such authors as Marie de France, Chrétien de Troyes, and Jean Renart, amongst others, the presentation of clothwork fits this classification. My research focuses on gendered performance and gendered objects highlighting the divide between masculinity and femininity in materiality. Beginning with a contextualised and historical understanding of feminine clothwork, authority, and gendered biases in the Middle Ages in France, the Virgin Mary's associations with clothwork leads into an exploration of how the identities of women are tied to the cloth that they work or possess. From this basis, feminine voice in clothwork comes to the forefront of discussion as seemingly inaudible women make themselves heard through the use of needles and thread, telling their stories in cloth and tapestry. Throughout this study, an exploration of mother-daughter relationships is highly significant to the comprehension of feminine education and tradition in clothwork. The chansons de toile included in Le Roman de la Rose ou de Guillaume de Dole by Jean Renart underline the dichotomy and tension between oral and written culture, tying feminine voice to feminine clothwork and exploring the representation of this in the written text. Finally, Christine de Pizan's intimation of the importance of feminine tasks and brilliance concludes this study in order to better understand the ways in which the literature of the Late Middle Ages and Renaissance departs from the medieval presentation of clothwork as a typically feminine activity underlying and encapsulating a woman's identity and expressive power.

This thesis argues that non-argumentational strategies of persuasion in the major-length speeches of Jesus in the Gospel of Mark account for a considerable part of the author's intended persuasive effect. Gospel criticism has typically read these speeches as deposits of possible authentic words of Jesus, as evidence of Christology or as links in the plot of Mark's narrative while rhetorical criticism of the Gospels has focused upon Jesus' shorter utterances as argumentation within episodes. However, the speeches are not consistently logical or argumentational and they stand out in the larger narrative as extended stretches of Jesus' voice.
Treating the four longest units of Jesus' speech as rhetorical units in their own right (Mk 4:11--32; 9:39--50; 12:1--11; 13:5--37), the thesis attempts to show how these texts might have worked to persuade Mark's audience towards change in action or attitude. Although neither argumentation at the rational-conceptual level nor the context of the narrative are denied, attention is focused on the rhetorical force of the speeches themselves and the features designed to affect the audience at the level of feeling or emotion. Persuasive goals are discernable in the speeches not only in what Jesus argues but also in the way he talks.

The thesis investigates parallels between the poetic theories and practices of William Wordsworth and Ezra Pound with reference primarily to The Prelude and the Cantos and to the prose writings of both poets. It argues that their work evolves in response to a common set of difficulties centred on the problem of the subject's relationship to language. Particular attention is given to the division revealed between what current criticism designates the enunciation and the enounced and for which Jacques Lacan provides the related terms the Symbolic and the Imaginary. The thesis explores the different strategies adopted to negotiate a set of related oppositions revealed by the writing process: between writing and experience; writing self and written self; thought and language; unconscious and conscious. Wordsworth's theory of 'spontaneous overflow' is contrasted with Pound's of writing as a unified act, but both are seen as the means whereby the writing process itself is rendered central to the concerns of their poetry. Chapter 1 considers Pound's responses to Wordsworth, and locates the argument of the thesis within a critical and theoretical context. Chapter 2 shows how in a strikingly similar manoeuvre both poems immediately throw open the question of the poet's relationship to language and to the empirical world. Chapter 3 returns to earlier versions of each poem to illustrate the formative role such problems had in their initial development. Chapter 4 considers how in both cases a poetics articulated as a 'search for sincere self-expression' develops strategies to overcome the division between self and language. Chapters 5 and 6 investigate with what success each poem negotiates a position for the poet in relationship to language. Chapter 7 draws together and broadens the arguments of the thesis, suggesting how The Prelude and the Cantos implicate us in different ideologies of reading.

La localització de paraules en el camp de anàlisis de documents es pot definir com el reconeixement de patrons encarregat de localitzar i recuperar una paraula específica dins d'una col·lecció d'imatges sense transcriure explícitament el corpus sencer. El seu ús és particularment interessant quan s'aplica a escenaris on el reconeixement òptic de caràcters funciona malament o no es pot utilitzar en absolut. Aquesta tesi se centra en un escenari d'aquest tipus, detectar paraules en documents manuscrits històrics que han estat escrits per un sol autor o per diversos autors amb una cal·ligrafia similar. Aquest problema requereix d'una signatura visual que sigui robusta contra artefactes de les imatges, flexible per adaptar-se a les variacions del traç i que sigui eficient per recuperar la informació de manera ràpida. Per a això, hem desenvolupat un conjunt de mètodes de localització de paraules que, en la seva base, utilitzen la coneguda representació Bag-of-Visual-Words (BoVW). Aquesta representació ha guanyat popularitat entre la comunitat d'anàlisi d'imatges de documents per caracteritzar paraules manuscrites en tasques no supervisades. Tanmateix, la majoria d'enfocaments en aquest camp es basen en una configuració bàsica de BoVW i ignoren les codificacions complexes i les representacions espacials. Determinem quines configuracions de BoVW proporcionen el millor increment de rendiment. A continuació, estenem la localització de paraules de sistemes on aquestes estan pre-segmentades a un on no utilitzem cap tipus de segmentació. L'enfocament proposat selecciona regions sobreposades del document com a candidates i les caracteritza amb una signatura BoVW. La localització s'aconsegueix comparant la imatge de consulta amb les signatures dels candidats i retornant les ubicacions que tenen un consens més alt. Aquest és un enfocament senzill però potent que requereix una signatura compacta. Primer projectem la signatura BoVW en un espai de temes semàntics i després la comprimim encara més mitjançant un producte de quantificadors. La signatura resultant requereix només unes dotzenes de bytes, cosa que ens permet indexar milers de pàgines en un ordinador de sobretaula estàndard. També estudiem com combinar diferents modalitats d'informació per tal de crear un sistema on les paraules s'indexa mitjançant una modalitat d'informació i les consultes mitjançant una altra. Considerem tres modalitats d'informació diferents: visual, textual i àudio. La nostra proposta és crear un espai de característiques latents on les característiques relacionades semànticament es projectin sobre els mateixos temes latents. Creant així un nou espai on la informació de diferents modalitats es pugui comparar. Els diccionaris que s'utilitzen per codificar les signatures BoVW es creen generalment mitjançant un algorisme de no supervisat i requereixen provar diversos paràmetres per determinar quina configuració és la millor per a una col·lecció de documents determinada. Proposem un algorisme d'agrupament semàntic que permet estimar els paràmetres a partir de dades. Atès que la recopilació de dades anotades és costosa, fem servir imatges de paraules generades sintèticament. El diccionari resultant proporciona un bon rendiment a les col·leccions de documents que utilitzen el mateix estil de text. També proposem l'ús d'un diccionari addicional per aproximar els descriptors i reduir la complexitat de codificació del descriptor a sub-lineal. Finalment, ens centrem en el problema de la dimensionalitat de les signatures. Proposem una nova signatura on cada element representa la probabilitat que un determinat símbol tingui una determinada ubicació dins la imatge de la paraula. Aquesta signatura és extremadament compacta i combinada amb tècniques de compressió, pot representar la imatge d'una paraula amb només uns quants bytes.
La detección de palabras se puede definir como el reconocimiento de patrones cuyo objetivo es localizar y recuperar una palabra clave específica dentro de una colección de imágenes de documentos sin transcribir explícitamente todo el corpus. Su uso es particularmente interesante cuando se aplica en escenarios donde el reconocimiento óptico de caracteres funciona mal o no se puede utilizar en absoluto. Esta tesis se centra en tal escenario, la detección de palabras en documentos históricos escritos a mano que han sido escritos por un solo autor o por varios autores con una caligrafía similar. Este problema requiere una signatura visual que sea robusta a los artefactos de la imagen, flexible para adaptarse a las variaciones del trazo y eficiente para recuperar información de manera rápida. Para ello, hemos desarrollado un conjunto de métodos de detección de palabras que, en su base, utilizan la conocida representación de Bag-of-Visual-Words (BoVW). Esta representación ha ganado popularidad entre la comunidad de análisis de imágenes de documentos para caracterizar palabras escritas a mano en tareas de detección no supervisadas. Sin embargo, la mayoría de los enfoques en este campo se basan en una configuración básica de BoVW y no tienen en cuenta la codificación compleja y las representaciones espaciales. Determinamos qué configuraciones de BoVW proporcionan el mejor impulso de rendimiento al sistema de detección. Luego, ampliamos la detección de palabras de sistemas donde estas están pre-segmentadas a uno sin segmentación. El enfoque propuesto selecciona regiones solapadas del documento como candidatas y las caracteriza con BoVW. La localización se logra comparando la imagen de consulta con las regiones candidatas y y devolviendo las ubicaciones que proporcionan un mayor consenso. Este es un enfoque simple pero poderoso que requiere una signatura compacta. Primero proyectamos la signatura BoVW en un espacio de temas semánticos y luego lo comprimimos aún más usando un producto de cuantificadores. La signatura resultante solo requiere unas pocas docenas de bytes, lo que nos permite indexar miles de páginas en una ordenador de sobremesa común. También estudiamos cómo combinar diferentes modalidades de información para crear un sistema donde las palabras se indexa usando una modalidad de información y las consultas se usando otra. Consideramos tres modalidades de información diferentes: visual, textual y audio. Nuestra propuesta es crear un espacio de características latentes donde las características que están relacionadas semánticamente se proyectan sobre los mismos temas. Creando así un nuevo espacio de características donde la información procedente de diferentes modalidades se puede comparar. Los diccionarios utilizados para codificar las signaturas BoVW se crean generalmente utilizando un algoritmo de agrupación no supervisado y requieren probar múltiples parámetros para determinar qué configuración es la mejor para una determinada colección de documentos. Proponemos un algoritmo de agrupamiento semántico que permite estimar el mejor parámetro a partir de los datos. Dado que recopilar datos anotados es costoso, utilizamos imágenes de palabras generadas sintéticamente. El diccionario resultante ofrece un buen rendimiento en colecciones de documentos que utilizan el mismo estilo de escritura. También proponemos el uso de un diccionario adicional para aproximar descriptores y reducir la complejidad de codificación del descriptor a sub-lineal. Finalmente, nos enfocamos en el problema de la dimensionalidad de las signaturas. Proponemos una nueva signatura de probabilidad de símbolo donde cada elemento representa la probabilidad de que un determinado símbolo esté presente en una determinada ubicación de la imagen de la palabra. Esta signatura es extremadamente compacta y combinada con técnicas de compresión puede representar la imágenes de una palabras usando solo unos pocos bytes.
Word spotting can be defined as the pattern recognition tasked aimed at locating and retrieving a specific keyword within a document image collection without explicitly transcribing the whole corpus. Its use is particularly interesting when applied in scenarios where Optical Character Recognition performs poorly or can not be used at all. This thesis focuses on such a scenario, word spotting on historical handwritten documents that have been written by a single author or by multiple authors with a similar calligraphy. This problem requires a visual signature that is robust to image artifacts, flexible to accommodate script variations and efficient to retrieve information in a rapid manner. For this, we have developed a set of word spotting methods that on their foundation use the well known Bag-of-Visual-Words (BoVW) representation. This representation has gained popularity among the document image analysis community to characterize handwritten words in unsupervised manner. However, most approaches on this field rely on a basic BoVW configuration and disregard complex encoding and spatial representations. We determine which BoVW configurations provide the best performance boost to the spotting system. Then, we extend the segmentation-based word spotting, where word candidates are given a priori, to segmentation-free spotting. The proposed approach seeds the document images with overlapping word location candidates and characterizes them with a BoVW signature. Retrieval is achieved comparing the query and candidate signatures and returning the locations that provide a higher consensus. This is a simple but powerful approach that requires a more compact signature than in a segmentation-based scenario. We first project the BoVW signature into a reduced semantic topics space and then compress it further using Product Quantizers. The resulting signature only requires a few dozen bytes, allowing us to index thousands of pages on a common desktop computer. The final system still yields a performance comparable to the state-of-the-art despite all the information loss during the compression phases. We also study how to combine different modalities of information in order to create a query-by-X spotting system where, words are indexed using an information modality and queries are retrieved using another. We consider three different information modalities: visual, textual and audio. Our proposal is to create a latent feature space where features which are semantically related are projected onto the same topics. Creating thus a new feature space where information from different modalities can be compared. The codebooks used to encode the BoVW signatures are usually created using an unsupervised clustering algorithm and, they require to test multiple parameters to determine which configuration is best for a certain document collection. We propose a semantic clustering algorithm which allows to estimate the best parameter from data. Since gather annotated data is costly, we use synthetically generated word images. The resulting codebook is database agnostic, i.e. a codebook that yields a good performance on document collections that use the same script. We also propose the use of an additional codebook to approximate descriptors and reduce the descriptor encoding complexity to sub-linear. Finally, we focus on the problem of signatures dimensionality. We propose a new symbol probability signature where each bin represents the probability that a certain symbol is present a certain location of the word image. This signature is extremely compact and combined with compression techniques can represent word images with just a few bytes per signature.
Universitat Autònoma de Barcelona. Programa de Doctorat en Informàtica

Research on the deciphering of nonsense words within the context of text, a story, or tale was conducted at various schools and day-care centres in the Stirling area of Scotland in 1985-1988. Three experiments were conducted, in which large samples of primary school children aged 4-12 were tested. The experiments resembled Werner and Kaplan's (1950) "Word-Context Task, " in which isolated sentences in a series with one nonsense word in each sentence were presented to school children. The children were asked to answer questions about the meanings of these words. The results were not in line with the rapid word learning that experience suggests happens in young children, it was not until after age 9 that the children started to give approximately correct answers, and prior to age 11 the answers did not meet up with proper adult definitions. It has been pointed out, however (Donaldson, 1978), that because these sentences were not supported by any relation to immediate context and behaviour, and because the children were required to process utterances as pure isolated language - an unnatural situation for language acquisition - the "Word-Context Task" may have given an unrealistic picture of the child's ability to acquire language naturally. In the three word-leaming studies at Stirling University in 1985-1988, in order to account for a more natural presentation, the sentences with the nonsense word were embodied in the context of a story. Children were thought to fare better (than the children in the Wemer & Kaplan study) when listening to such a story, especially if the basic theme was of interest. A methodological tool, refined in the work of Dockrell (1981), in which the full meaning of a term involves having worked out the sense, reference, and denotation of the term. was applied in each of the test batteries that followed the presentation of the story. In these tests, the children were tested on both their comprehension and production of the new term in question. Drawings were used in order to try to tap the children's denotation of the new term, and to facilitate young children's approach to the demands of the study. As regards word meaning in general. Martin Joos (1972) had argued that the common blunder was that an odd word must have an odd sense--the odder, the better. He argued that one should define words in such a fashion as to make them contribute least to the total message derivable from its passage where it is housed, rather than, e. g., defining it according to some presumed etymology of semantic history. He called this concept "a tacit principle", and argued that word learners and word users would sense the intuitive familiarity of the conveyed meaning of words and text. Words are, according to this principle, "mysterious" in their environment, their meanings are not worked out deliberately, intentionally; rather, one should make the mysterious item maximally supportive and supported in its situation, in order that redundancy would result in proper connotation of the distributed meaning. Context and knowledge of contexts reveal meaning; the text is processed holistically, and so are the instantaneous meanings of the words of which it is composed. Thus, Joos maintained that in deciphering an unknown word, the wisest course is to assume the "least meaning" consistent with the context. Tasks such as Werner and Kaplan's "Word-Context Task" (1950), force subjects to infer aspects of meaning that go well beyond this "least" meaning and, as Joos pointed out, this leads notably to errors from which recovery is difficult. In the studies at Stirling University, attempts were made to determine if different types of learning would result in different types of responses. The dichotomy, intentional/incidental or analytic/holistic was worked out into experimental and control conditions, as based on Aveling's pioneering experiment (1911, 1912) into the general and particular aspects of encoded stimuli. Later, Lee Brooks (1978) worked with the dichotomies intentional/incidental in his Lepton experiments and argued that the more complex a behaviour is (speaking or writing, for example), the more likely it is to be learned implicitly. He pointed out, however, that the dichotomies explicit/implicit, analytic/non analytic, and deliberate vs. intuitive processes need to be elaborated and not taken as a strict division. In the three experiments at Stirling, children of primary school age (ages 4 to 12) were presented with a "word-context" task and their understanding of the unknown word was probed under different conditions. In the control condition a control word was probed, but in the experimental condition the child's understanding of the target word was fully tested. All the children listened to a short story displayed by a video or read from a tape in which the unknown word occurred in several different contexts, the unknown word in each story denoted an unfamiliar natural kind. During the story's display, children in the control condition were, at certain intervals, asked questions about the story's theme. Children in the experimental group were, at these same intervals, shown a sample of objects, to one of which the unknown word referred, and they were asked to hand these objects to the experimenter as she requested the objects, or they were asked direct questions about the meaning of the target word and about other words in the story. After hearing the story, all subjects were tested on their comprehension and production of the unknown word, together with other words, and a scoring procedure based on a technique developed by Dockrell (1981) was applied. This procedure necessitated the full meaning of the term covering aspects of the sense, reference., and denotation of the new term (cf. Lyons, 1977a). The results indicate that children younger than those tested in the Wemer and Kaplan's "Word-Context Task" (ages 8.6 to 13.6) could decipher the full meaning of the new term. But individual differences within age groups showed greater differences than existed between age groups. All in all, the results indicate that working out the full meaning of a new term is a lengthy process indeed (Campbell & Dockrell, 1986), even though a sense of the given semantic domain may often be established quite early in the learning process. Performance styles also differ from younger children to older ones. The results indicate that there were significant age differences between the children in the first and second experiments, but that such differences were lacking in the third experiment, and that control subjects in the three studies seldom gave poorer responses than did experimental subjects and often did better. However, the results must be interpreted in the light of learning and recovery from error occurring, within the experimental subjects in the course of deciphering. If the initial scores of the experimental subjects on the target word as obtained during encoding are compared with the first scores obtained from the control subjects after they had heard the whole story, there is a significant difference in scores between the conditions in favour of the control subjects in all age groups. This is consistent with Joos's assumption that an interference concerning the meaning of a word that occurs too early in the learning task and not enough information of contextual cues will lead the children in the experimental groups astray in their guesses when asked too early for answers on the new word's meaning. But implied in Joos's Axiom is the likelihood for recoveries from errors, and the strategies children use in order to work them out need to be explored further. Much individual variation was found among the children's responses in the age groups. These differences were indeed more significant than were the differences between age groups.

The purpose of this research paper is to examine how Syrian refugees are visually represented in two of the largest newspapers in Sweden: Dagens Nyheter (Today’s news) and Svenska Dagbladet (The Swedish daily paper). Their choice of images on the Syrian refugee crisis will be analysed, to see what sort of message they send by using those specific images. The reason for this study is because previous studies have implied that the media has an important role in crafting national perceptions of refugees. Therefore, it is intriguing to analyse the images the newspapers published during the fall of 2015 to see how Syrian refugees were represented during that most active and chaotic period in Sweden and if the visual representation of Syrian refugees can influence cultures of hostility in Sweden. The study has been linked with previous studies research of visual representations of refugees (us and them, identifiable victim effect) and the severe consequences it can lead to (dehumanization, moral panic) if used improperly, and three main theories that will be the foundation to custom and analyse the images (agenda setting, framing and representation theory). These were all applied and used when operating the method (quantitative content analysis) to collect information and data and also later when analysing and discussing the results. The result of the study showed that the newspapers images did not only focus on negative aspects of events and stories surrounding Syrian refugees. Dagens Nyheter and Svenska Dagbladet did not publish many negative images, but after the results, it is discovered that they are following that specific path on some aspects and that can be the start of a pattern that will be used more often in the future. While it is established that hostile and unwelcoming attitudes in the West have risen, the question remains how far its been developed in Sweden and how long it will be until the Swedish media promotes it as well. Key words: Syrian Refugee Crisis, Visual representation, Framing, Agenda setting and Stereotypes

Conselho Nacional de Desenvolvimento CientÃfico e TecnolÃgico
Neste trabalho analisam-se as representaÃÃes sociais (RS) sobre o espanhol/lÃngua estrangeira (E/LE), construÃdas e partilhadas, via discurso, por alunos da rede pÃblica cearense de ensino. Para tanto, partimos do pressuposto de que hà RS sobre o espanhol, caracterÃsticas a esse grupo social, que podem depender da escola ou dos seguintes fatores: (1) a modalidade educacional (regular e a profissionalizante) e o (2) contato prÃvio ou paralelo com referida lÃngua ao Ensino MÃdio. Assim, delimita-se um campo associativo geral, constituÃdo por lexias associadas ao termo indutor âlÃngua espanholaâ, evocadas atravÃs de testes de associaÃÃo livre de palavras, a partir do qual figura-se um nÃcleo central das representaÃÃes sobre o objeto investigado, evidenciando-as, portanto. A pesquisa baseia-se, sobretudo, nos trabalhos de Moscovici (1978, 2003), Abric (1994, 1994a, 1994b, 2001), Doise (1992), Jodelet (2001), Harrà (2001) e Sà (1996, 1998, 2002), com relaÃÃo à Teoria das RepresentaÃÃes Sociais; em Moita Lopes (2011) e Guy Cook (2013), no que diz respeito à LinguÃstica Aplicada; em Saussure (2007), Vilela (1979), Pottier (1978) e van Dijk (2003), no tocante ao lÃxico, categoria linguÃstica utilizada como vetor de representaÃÃes; alÃm dos documentos parametrizadores do ensino PCNEM (2000), OCEM (2006) e REM-Seduc/CE (2008), quanto Ãs discussÃes sobre ensino e aprendizagem de lÃngua. A pesquisa conta com a colaboraÃÃo de 24 estudantes, distribuÃdos em quatro grupos focais, selecionados atravÃs da aplicaÃÃo de um questionÃrio sondagem. AlÃm dos testes, os dados foram gerados por meio de entrevistas semiestruturadas realizadas aos grupos. A partir da anÃlise, constatou-se que a base das representaÃÃes construÃdas e partilhadas pelo grupo sobre a lÃngua gravitavam em torno de trÃs principais thematas: âespanhol à CULTURAâ, âespanhol à AQUISIÃÃOâ e âespanhol à LUGARâ. Desse modo, evidenciaram-se representaÃÃes sobre lÃngua espanhola como cultura, cultura espanhola, lÃngua massificada, lÃngua fÃcil, âarmadilhaâ, como saber, sistema/cÃdigo linguÃstico, instrumento de comunicaÃÃo, elo, lÃngua atraente e âlÃngua novaâ. Por fim, pudemos afirmar, porÃm, que as representaÃÃes nÃo sà orientam atitudes e percepÃÃes sobre a lÃngua, podendo se constituir como um fator condicionante da aprendizagem, como tambÃm o inverso à verdadeiro, ou seja, a forma como a aprendizagem do objeto se dà tambÃm condiciona e orienta percepÃÃes e atitudes em relaÃÃo a este.
In this research, we analyze the social representations (SR) of the Spanish/ foreign language (S/FL) which are build and shared via discourse by students from the public education Network of CearÃ. For that we assume that there's RS about the Spanish language, which are common in this social group, that may depend on the school or on the following factors: (1) the educational mode (regular and professional) and (2) previous contact or parallel contact with the language in High School. Thus, we delimitate a general associative field that is composed by lexias associated with the inductor term âSpanish languageâ, evoked through free association tests of words, from which figures a central core of the representations related to the investigated object that, consequently, highlights them. The research is based on the works by Moscovici (1978, 2003), Abric (1994, 1994a, 1994b, 2001), Doise (1992), Jodelet (2001), Harrà (2001) and Sà (1996, 1998, 2002), regarding the Social Representations Theory; in Moita Lopes (2011) and Guy Coock (2013), regarding Applied Linguistics; in Saussure (2007), Vilela (1979), Pottier (1978), van Dijk (2003) regarding the lexical aspect, a linguistic category perceived as a vector of representations besides the documents PCNEM (2000), OCEM (2006), REM-SEDUC (2008) as for the discussions related to teaching and language learning. This research has the support of 24 students distributed into four focal groups, which were chosen through the application of a questionnaire. Besides the tests, the data was also generated by semi-structural interviews conducted in the groups. Based on the analysis, it was found that the basis of the representations about the language built and shared by the group gravitates around three main thematas: âSpanish is CULTUREâ, âSpanish is ACQUISITIONâ and âSpanish is PLACEâ. Thus, it was possible to identify representations of the Spanish language as culture, Spanish culture, massified culture, easy culture, âa trapâ, that is, system/ linguistic code, communication tool, link, attractive language and ânew languageâ. Ultimately, we can argue, however, that the representations do not only orientate the attitude and perceptions about the language, since they can also be a conditioning learning factor, but also that the inverse is a fact, which means that the way the object is learned also determines and orientates the perceptions and attitude towards it.

The thesis is divided into three sections. The first section discusses the manuscript Marcianus graecus 524, the second looks at the Greek text of the dedicatory epigrams on works of art from the same manuscript, and the third puts these texts in their context. In the first part, the compilation of the manuscript is analysed. I suggest that the manuscript was copied mainly by one individual scribe living in Constantinople at the end of the thirteenth century. He copied the quires individually, but at some point he put all these quires together, added new quires, and compiled an anthology of poetry. The scribe’s connection to the Planudean School and the Petra monastery in Constantinople is discussed. Although their relationship remains inconclusive, the manuscript provides evidence regarding the literary interests of late-thirteenth-century intellectuals. The second part contains thirty-five unpublished dedicatory epigrams on works of art. New readings are offered for the text of previously published epigrams. The third section analyses the dedicatory epigrams on works of art in their context. The first chapter of this section discusses the epigrams as Gebrauchstexte, i.e. texts with a practical use. The difference between epigrams intended to be inscribed and epigrams intended to be performed is highlighted. In the next chapter of this part, La poésie de l’objet, the composition of the dedicatory epigrams is discussed. The conventional character of the epigrams suggests that the poetics express the ritual aspect of the epigram. The last chapter considers the texts from a more pragmatic angle. After a short discussion of the objects on which the epigrams were written, the mechanisms of the twelfth-century art market are presented based on evidence taken mainly from the epigrams. At the end of this part, conclusions are drawn on the understanding of these texts in the twelfth century.

'Shifting Eyes', presents a re-reading of Freud’s structural theories of the self, through the semiotics of C.S. Peirce. In place of the self split between unconscious representations and the syntax of speech, Peirce’s general sign theory provides an evolutionary account of symbol development within a trichotomy of sign-object relations, icon, index and symbol, as opposed to interpretations of Freud using the linguistic sign which reify the split subject and assimilate unconscious processes to the tropes of language. Peirce’s sign-interpretant relation, is used to re-describe Freud’s account of the shift from narcissism to object relations, from the primary iconic dyad to the subject constructed through the symbol of sexual difference. One class of icons, the hypoicon, is evaluated as a representation of the subject, since the hypoicon, unlike the symbol, does not uphold contradiction. Metaphor, as hypoiconic Third, is compared with Freud’s account of the structure of identification, both in terms of ego development and dream formation. The second part of the thesis uses these concepts to interpret the work of author H.D. and artist-writer Claude Cahun. H.D.’s œuvre - poetry, novels, memoirs and autobiography - lay bare the structure of the subject through the semiotics of the text, in particular transference and the act of naming. The poetry demonstrates the boundary between ego and world, myth and ideals of the ego, as the semiotics of identification. Cahun’s photographic self-portraits raise questions of the relation between body-image and narcissism, ideals and the subject of sexual difference. The last chapter concentrates on 'Aveux non avenus', (1930a) a work which integrates text and image using the principles of collage, juxtaposing photomontages with fragments of dream, fantasy, polemic and fiction as an extension of self-representation. In conclusion, the signifying self, as hypoiconic Third, is related to the body, re-posing the question of desire.

Submitted by Renata Lopes (renatasil82@gmail.com) on 2018-11-22T17:34:06Z No. of bitstreams: 1 renatobrazalbertinosouza.pdf: 1758176 bytes, checksum: bc18bf464a0293c532187d08d77ed586 (MD5)
Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2018-11-23T12:50:40Z (GMT) No. of bitstreams: 1 renatobrazalbertinosouza.pdf: 1758176 bytes, checksum: bc18bf464a0293c532187d08d77ed586 (MD5)
Made available in DSpace on 2018-11-23T12:50:40Z (GMT). No. of bitstreams: 1 renatobrazalbertinosouza.pdf: 1758176 bytes, checksum: bc18bf464a0293c532187d08d77ed586 (MD5) Previous issue date: 2018-10-10
Esta dissertação foi desenvolvida no Programa de Mestrado Profissional em Administração Pública da Faculdade de Administração e Ciências Contábeis da Universidade Federal de Juiz de Fora e buscou compreender como a representação social de greve dos funcionários da UFJF tem impactado o movimento grevista. As greves no setor público são fenômenos polêmicos, podendo ser consideradas como causadoras de transtornos à sociedade ou como um direito inalienável. É possível verificar na instituição a vivência de greves quase todos os anos, tendo algumas uma longa duração. A pesquisa procurou compreender os conceitos e as percepções relacionadas aos movimentos grevistas. Para tanto, apoiou-se na Teoria das Representações Sociais, que se revela promissora para a obtenção de um entendimento mais profundo dos fenômenos sociais. Identificou-se a representação das greves sob a perspectiva dos Técnico-administrativos em Educação, que constituem o pessoal não docente da universidade. Foi utilizada a técnica de evocação de palavras para a coleta de dados. A pesquisa revelou que a representação social das greves comporta, significativamente, a ideia de conotação positiva luta por direitos, demonstrando que os servidores as compreendem como movimentos reivindicatórios. Contudo, ao mesmo tempo, envolvem relevantes conotações negativas, como a transformação em uma simples oportunidade de folga por parte dos servidores, além dos prejuízos naturalmente causados à comunidade acadêmica devido à descontinuidade dos serviços prestados em função da paralisação das atividades laborativas. Os principais problemas identificados referem-se à carência de diálogo e interação entre os servidores e à falta de informações adequadas dos mesmos acerca das questões que envolvem as greves. Como possível estratégia de ação, sugere-se, especialmente, a promoção do diálogo e interação entre os funcionários, por meio da criação de ambientes de discussão das questões concernentes às relações de trabalho de maneira institucionalizada pela administração da UFJF. Também se recomenda o fornecimento de informações adequadas aos servidores e ainda a adoção de ações de formação pessoal e profissional, relacionadas à conscientização, treinamento e desenvolvimento dos mesmos. Tais ações podem ser implementadas tanto pela administração da universidade como pelo sindicato representativo da categoria.
This dissertation was developed in the Professional Master's Program in Public Administration of the Faculty of Administration and Accounting Sciences of the Federal University of Juiz de Fora and sought to understand how the social representation of UFJF employees' strike has impacted the strike movement. Strikes in the public sector are controversial phenomena, and can be considered as causing disruption to society or as an inalienable right. It is possible to verify in the institution the existence of strikes almost every year, some of which have a long duration. The research sought to understand the concepts and perceptions related to the strike movements. To that end, it relied on the Theory of Social Representations, which proves to be promising for a deeper understanding of social phenomena. It was identified the representation of the strikes from the perspective of the Technical-administrative in Education, who constitute the non-teaching personnel of the university. The word evocation technique was used to collect data. The research revealed that the social representation of strikes comprise, significantly, the idea of positive connotation fighting for rights, demonstrating that the servants understand them as protest movements. However, at the same time, they involve relevant negative connotations, such as the transformation into a simple opportunity to take time off on the part of the servers, in addition to the damages naturally caused to the academic community due to the discontinuity of the services provided due to the work stoppage. The main problems identified relate to the lack of dialogue and interaction among the servers and the lack of adequate information on the issues surrounding the strikes. As a possible strategy for action, it is suggested, especially, the promotion of dialogue and interaction among employees, through the creation of environments for discussion of issues concerning labor relations in an institutionalized manner by the UFJF administration. It is also recommended to provide adequate information to the servers and also the adoption of personal and professional training actions, related to the awareness, training and development. Such actions may be implemented by both the university administration and the union representative of the category.

In this study, the newspapers such as Hü
rriyet, Milliyet, Radikal, Posta, AkSam, Sabah, Cumhuriyet, Star, Gü
neS, Takvim, Dü
nya, Gö
zcü
and Yeni Asir are considered liberal or left leaning. The newspapers such as Akit, Milli Gazete, Yeni Asya, Anayurt, Tü
rkiye, Yeni Safak, Dü
nden Bugü
ne Tercü
man, Vakit and Zaman are considered conservative or Islamic newspapers. The period chosen for this analysis is between 28.12.1998 and 15.06.2006, and all the news items published in this period are scrutinised in the archive of the Parliament Library, where there is a special category for the news on homosexuality and homosexuals. First, the news items of the newspapers mentioned above are categorised as presented in the tables prepared. Through these tables, a general perspective on the press coverage in Turkey is revealed by the analysis of the evaluations of the general features of this coverage on homosexuality and homosexuals. Second, an extensive textual analysis is carried out by analysing the news texts regarding three common events covered in most of the newspapers mentioned above to reflect a more accurate and detailed account of the representation of homosexuality and homosexuals in the Turkish Daily Press. Moreover, in-depth interviews are conducted with homosexuals to find out their point of view on the portrayal of themselves in the Turkish Daily Press. Finally, the findings of the analysis of the press coverage of homosexuality and homosexuals, and the findings of the interviews are compared.

This thesis is an exloration and exposition of a highly efficient shallow neural network algorithm called word2vec, which was developed by T. Mikolov et al. in order to create vector representations of a language vocabulary such that information about the meaning and usage of the vocabulary words is encoded in the vector space structure. Chapter 1 introduces natural language processing, vector representations of language vocabularies, and the word2vec algorithm. Chapter 2 reviews the basic mathematical theory of deterministic convex optimization. Chapter 3 provides background on some concepts from computer science that are used in the word2vec algorithm: Huffman trees, neural networks, and binary cross-entropy. Chapter 4 provides a detailed discussion of the word2vec algorithm itself and includes a discussion of continuous bag of words, skip-gram, hierarchical softmax, and negative sampling. Finally, Chapter 5 explores some applications of vector representations: word categorization, analogy completion, and language translation assistance.

Image representation is in the heart of many computer vision algorithms. Different computer vision tasks (e.g. classification, detection) require discriminative image representations to recognize visual categories. In a nutshell, the bag-of-visual-words image representation is the most successful approach for object and scene recognition. In this thesis, we mainly revolve around this model and search for discriminative image representations. In the first part, we present a novel approach to incorporate spatial information in the BoVW method. In this framework, we present a simple and efficient way to infuse spatial information by taking advantage of the orientation and length of the segments formed by pairs of similar descriptors. We introduce the notion of soft-similarity to compute intra and inter visual word spatial relationships. We show experimentally that, our method adds important discriminative information to the BoVW method and complementary to the state-of-the-art method. Next, we focus on color description in general. Differing from traditional approaches of invariant description to account for photometric changes, we propose discriminative color descriptor. We demonstrate that such a color description automatically learns a certain degree of photometric invariance. Experiments show that the proposed descriptor outperforms existing photometric invariants. Furthermore, we show that combined with shape descriptor, the proposed color descriptor obtain excellent results on four challenging data sets.Finally, we focus on the most accurate color representation i.e. multispectral reflectance which is an intrinsic property of a surface. Even with the modern era technological advancement, it is difficult to extract reflectance information without sophisticated instruments. To this end, we propose to use the display of the device as an illuminant while the camera captures images illuminated by the red, green and blue primaries of the display. Three illuminants and three response functions of the camera lead to nine response values which are used for reflectance estimation. Results show that the accuracy of the spectral reconstruction improves significantly over the spectral reconstruction based on a single illuminant. We conclude that, multispectral data acquisition is potentially possible with consumer hand-held devices such as tablets, mobiles, and laptops

In recent decades, various software systems have gradually become the basis of our society. Programmers search existing code snippets from time to time in their daily life. It would be beneficial and meaningful to have better solutions for the task of semantic code search, which is to find the most semantically relevant code snippets for a given query. Our approach is to introduce tree representations by multi-modal learning. The core idea is to enrich semantic information for code snippets by preparing data of different modalities, and meanwhile ignore syntactic information. We design one novel tree structure named Simplified Semantic Tree and then extract RootPath representations from that. We utilize RootPath representation to complement the conventional sequential representation, namely the token sequence of the code snippet. Our multi-modal model receives code-query pair as input and computes similarity score as output, following the pseudo-siamese architecture. For each pair, besides the ready-made code sequence and query sequence, we extra one extra tree sequence from Simplified Semantic Tree. There are three encoders in our model, and they respectively encode these three sequences as vectors of the same length. Then we combine the code vector with the tree vector for one joint vector, which is still of the same length, as the multi-modal representation for the code snippet. We introduce triplet loss to ensure vectors of code and query in the same pair be close at the shared vector space. We conduct experiments in one large-scale multi-language corpus, with comparisons of strong baseline models by specified performance metrics. Among baseline models, the simplest Neural Bag-of-Words model is with the most satisfying performance. It indicates that syntactic information is likely to distract complex models from critical semantic information. Results show that our multi-modal representation approach performs better because it surpasses baseline models by far in most cases. The key to our multi-modal model is that it is totally about semantic information, and it learns from data of multiple modalities.
Under de senaste decennierna har olika programvarusystem gradvis blivit basen i vårt samhälle. Programmerare söker i befintliga kodavsnitt från tid till annan i deras dagliga liv. Det skulle vara fördelaktigt och meningsfullt att ha bättre lösningar för uppgiften att semantisk kodsökning, vilket är att hitta de mest semantiskt relevanta kodavsnitten för en given fråga. Vår metod är att introducera trädrepresentationer genom multimodal inlärning. Grundidén är att berika semantisk information för kodavsnitt genom att förbereda data med olika modaliteter och samtidigt ignorera syntaktisk information. Vi designar en ny trädstruktur med namnet Simplified Semantic Tree och extraherar sedan RootPath-representationer från det. Vi använder RootPath-representation för att komplettera den konventionella sekvensrepresentationen, nämligen kodsekvensens symbolsekvens. Vår multimodala modell får kodfrågeställningar som inmatning och beräknar likhetspoäng som utgång efter den pseudo-siamesiska arkitekturen. För varje par, förutom den färdiga kodsekvensen och frågesekvensen, extrager vi en extra trädsekvens från Simplified Semantic Tree. Det finns tre kodare i vår modell, och de kodar respektive tre sekvenser som vektorer av samma längd. Sedan kombinerar vi kodvektorn med trädvektorn för en gemensam vektor, som fortfarande är av samma längd som den multimodala representationen för kodavsnittet. Vi introducerar tripletförlust för att säkerställa att vektorer av kod och fråga i samma par är nära det delade vektorn. Vi genomför experiment i ett storskaligt flerspråkigt korpus, med jämförelser av starka baslinjemodeller med specificerade prestandametriker. Bland baslinjemodellerna är den enklaste Neural Bag-of-Words-modellen med den mest tillfredsställande prestanda. Det indikerar att syntaktisk information sannolikt kommer att distrahera komplexa modeller från kritisk semantisk information. Resultaten visar att vår multimodala representationsmetod fungerar bättre eftersom den överträffar basmodellerna i de flesta fall. Nyckeln till vår multimodala modell är att den helt handlar om semantisk information, och den lär sig av data om flera modaliteter.

On the moving word task 3- to 4-year-olds judge that a face-up written word changes in what it 'says' when moved from alongside a matching object to a nonmatching object. Bialystok (2000) interprets these errors as symptomatic of a misconception that words' meanings are flexible, and argues that surmounting this misconception is a prerequisite to reading. On the false word task 3- to 4-year-olds judge that a face-down word changes in what it says in line with a change to its referent. Thomas and colleagues (1999) argue that children make these errors because they assume that words physically change. Two experiments compared 3- to 5-year-olds' performance on face-up and facedown versions of the two tasks. Consistent with Bialystok's interpretation, children made errors on both tasks for face-up words, and consistent with Thomas and colleagues' interpretation, children were more likely to make errors for face-down words on the false word task and the moving word task employed in Experiment 2. Six experiments compared 3- to 5-year-olds' performance on the face-up moving word task with modified tasks. Children were less likely to treat face-up words' meanings as flexible on three new tasks: in the reverse task a word was moved from alongside a nonmatching object to a matching object; in the word exchange task two words were exchanged from matching objects to nonmatching objects; and when asked to make print say something different in the card tum task, children responded as if they assumed that the word's physical form needed to be changed rather than its pictorial context. The findings suggest that errors for face-up words are not indicative of a misconception, instead, they are maximised by demands intrinsic to the task. It is argued that, as with other representations, young children assume that words' meanings are fixed, but this understanding is fragile and can be overridden. With increasing familiarity with written words, children become better able to resist making errors for face-up words.

The focus of the current thesis is to examine the linguistic and non-linguistic factors that affect bilingual performance with different language tasks. The aim of the investigation is twofold: on the one hand it examines the role of translation similarity in processing words in the two languages. Specifically, if processing of cognates (translations that share form) and non-cognates (translations with no form overlap) differ as a function of orthographic and phonological overlap between Greek and English. This pair of languages is not commonly researched in bilingual studies and the benefit of investigating them stems from their partial only similarity. Greek and English differ phonologically and orthographically although they are both alphabetic and have an etymological relationship. This makes it an interesting pair because their partial similarity in both codes is not as usual as in more commonly studied language pairs, where there is either only phonological (e.g., Chinese - English cognates) or more extensive orthographic and phonological similarity (with identical alphabets, e.g., Spanish-English cognates). By studying languages with partial similarity, it is possible to compare translations with various degrees of orthographic and phonological similarity. The current study also examines whether adult and children bilinguals with minimal usage of the second language and tested in their first language environment would show differences in processing translations that appear in their first or second language. Of particular interest here is that two groups of adults and children are tested that have either Greek or English as their first language, which provides the opportunity to examine group effects cross-linguistically and as function of age while ensuring that language dominance and testing environment are controlled.

This work highlights some important factors for consideration when developing word vector representations and data-driven conversational systems. The neural network methods for creating word embeddings have gained more prominence than their older, count-based counterparts.However, there are still challenges, such as prolonged training time and the need for more data, especially with deep neural networks. Shallow neural networks with lesser depth appear to have the advantage of less complexity, however, they also face challenges, such as sub-optimal combination of hyper-parameters which produce sub-optimal models. This work, therefore, investigates the following research questions: "How importantly do hyper-parameters influence word embeddings’ performance?" and "What factors are important for developing ethical and robust conversational systems?" In answering the questions, various experiments were conducted using different datasets in different studies. The first study investigates, empirically, various hyper-parameter combinations for creating word vectors and their impact on a few natural language processing (NLP) downstream tasks: named entity recognition (NER) and sentiment analysis (SA). The study shows that optimal performance of embeddings for downstream \acrshort{nlp} tasks depends on the task at hand.It also shows that certain combinations give strong performance across the tasks chosen for the study. Furthermore, it shows that reasonably smaller corpora are sufficient or even produce better models in some cases and take less time to train and load. This is important, especially now that environmental considerations play prominent role in ethical research. Subsequent studies build on the findings of the first and explore the hyper-parameter combinations for Swedish and English embeddings for the downstream NER task. The second study presents the new Swedish analogy test set for evaluation of Swedish embeddings. Furthermore, it shows that character n-grams are useful for Swedish, a morphologically rich language. The third study shows that broad coverage of topics in a corpus appears to be important to produce better embeddings and that noise may be helpful in certain instances, though they are generally harmful. Hence, relatively smaller corpus can show better performance than a larger one, as demonstrated in the work with the smaller Swedish Wikipedia corpus against the Swedish Gigaword. The argument is made, in the final study (in answering the second question) from the point of view of the philosophy of science, that the near-elimination of the presence of unwanted bias in training data and the use of foralike the peer-review, conferences, and journals to provide the necessary avenues for criticism and feedback are instrumental for the development of ethical and robust conversational systems.

This thesis presents work on learning representations of text and Knowledge Bases by taking into consideration their respective structures. The tasks for which the methods are developed and evaluated on are: Short-text classification, Word Sense Induction and Disambiguation, Knowledge Base Completion with linked text corpora, and large-scale Knowledge Base Question Answering. An introductory chapter states the aims and scope of the thesis, followed by a chapter on technical background and definitions. In chapter 3, the impact of dependency syntax on word representation learning in the context of short-text classification is investigated. A new definition of context in dependency graphs is proposed, which generalizes and extends previous definitions used in word representation learning. The resulting word and dependency feature embeddings are used together to represent dependency graph substructures in text classifiers. In chapter 4, a probabilistic latent variable model for Word Sense Induction and Disambiguation is presented. The model estimates sense clusters using pretrained continuous feature vectors of multiple context types: syntactic, local lexical and global lexical, while the number of sense clusters is determined by the Integrated Complete Likelihood criterion. A model for Knowledge Base Completion with linked text corpora is presented in chapter 5. The proposed model represents potential facts by merging subgraphs of the knowledge base with text through linked entities. The model learns to embed the merged graphs into a lower dimensional space and score the plausibility of the fact with a Multilayer Perceptron. Chapter 6 presents a system for Question Answering on Knowledge Bases. The system learns to decompose questions into entity and relation mentions and score their compatibility with queries on the knowledge base expressed as subgraphs. The model consists of several components trained jointly in order to match parts of the question with parts of a potential query by embedding their corresponding structures in lower dimensional spaces.

This thesis examines the representation of time and history in the literature of Joseph Conrad. It explores the importance of nineteenth-century Polish Romantic philosophies of history on Conrad's literary development, arguing that the Polish response to the Hegelian tradition of historiography in nineteenth-century Europe influenced.Conrad's interpretation of history and time. After investigating Conrad's early career in the context of the philosophy of history and the philosophy of time, the thesis analyses Conrad's major works, Nostromo (1904), The Secret Agent (1907), and Under Western Eyes (1911) in light of Conrad's writing on the subject of Poland.These novels treat the question of the nation and history. Conrad juxtaposes his belief in an inherited Polish national identity, derived from Herder and Rousseau, with a sceptical questioning of modem nationalism. Nostromo presents the creation of the modem nation state of Sulaco; The Secret Agent explores the subject of 'foreigners' and nationality in England; while Under Western Eyes constitutes a sys~ematic attempt to undermine Russian national identity. The importance of the subject of the nation to Conrad's work ensures he is an author who examines critically the forces of nationalism and identity that troubled Europe throughout the nineteenth century and in the period before the First World War.

Submitted by Gustavo Marchisotti (marchisotti@terra.com.br) on 2014-11-25T16:12:30Z No. of bitstreams: 1 Dissertacao_Final & Definitiva.pdf: 1891620 bytes, checksum: b110a698ad08c8eacc0c239564e378ae (MD5)
Approved for entry into archive by Janete de Oliveira Feitosa (janete.feitosa@fgv.br) on 2014-11-25T18:34:06Z (GMT) No. of bitstreams: 1 Dissertacao_Final & Definitiva.pdf: 1891620 bytes, checksum: b110a698ad08c8eacc0c239564e378ae (MD5)
Approved for entry into archive by Marcia Bacha (marcia.bacha@fgv.br) on 2014-12-12T16:25:51Z (GMT) No. of bitstreams: 1 Dissertacao_Final & Definitiva.pdf: 1891620 bytes, checksum: b110a698ad08c8eacc0c239564e378ae (MD5)
Made available in DSpace on 2014-12-12T16:30:33Z (GMT). No. of bitstreams: 1 Dissertacao_Final & Definitiva.pdf: 1891620 bytes, checksum: b110a698ad08c8eacc0c239564e378ae (MD5) Previous issue date: 2014-11-13
Este estudo visa desenvolver uma investigação exploratória e quali-quantitativa, a cerca da representação social do Cloud Computing, na visão dos profissionais de TI brasileiros. Objetiva expor quais as percepções dos usuários da área de TI a respeito do paradigma computacional Cloud Computing. Para suportar o estudo teórico, foram coletados dados empíricos, por meio de questionários online respondidos por 221 profissionais da área de TI. Com o uso da técnica de evocação de palavras e da teoria da representação social (TRS), os dados coletados foram sumarizados. Após o tratamento dos dados mediante o uso da técnica do quadro de quatro casas de Vergès, obteve-se como resultado, a identificação do núcleo central e do sistema periférico da representação social do Cloud Computing. Por fim, os dados foram analisados utilizando-se as análises implicativa e de conteúdo, de forma a que todas as informações fossem abstraídas para melhor interpretação do tema. Obteve-se como resultado, que o núcleo central da representação social do Cloud Computing é composto pelas seguintes palavras 'Nuvem', 'Armazenamento', 'Disponibilidade', 'Internet', 'Virtualização' e 'Segurança'. Por sua vez, as palavras identificadas como parte do sistema periférico da representação social do Cloud Computing foram: 'Compartilhamento', 'Escalabilidade' e 'Facilidade'. Os resultados permitem compreender qual percepção dos profissionais de TI a respeito deste paradigma tecnológico e sua correlação com o referencial teórico abordado. Tais informações e percepções podem auxiliar a tornar o não familiar em familiar, ou seja, compreender como o Cloud Computing é representado, visto e, finalmente, reconhecido pelos profissionais da área de TI.

Many current psycholinguistic theories view the mental lexicon as a listing of (only) unpredictable sound-meaning correspondences (primarily words and morphemes). Under this view, regular complex word forms and syntactic structures are built with rules during language production, obviating the need for storage of complex but regular linguistic material. This type of model conflicts with recent experimental evidence that suggests that lexical memory may in fact consist of a more heterogeneous set of linguistic units, including complex word forms and multi-word expressions that in theory could be constructed via rules. Storage of such material seems to be driven largely by frequency. The present research consisted of two experiments designed to investigate whether semantically transparent noun and adjective phrases are stored as single lexical units when they are very frequent. Results from the two tasks (grammaticality judgment and speech production) supported the notion that frequent word combinations can come to be stored holistically in lexical memory. It was argued that usage-based models of lexical memory (vs. dominant generative theories) best account for such data. Data from language acquisition, aphasia, and corpus studies were offered as complementary evidence in support of the more general claim that a large component of linguistic competence is knowledge of lexical co-occurrence patterns. Finally, it was hypothesized that an exemplar-based model of lexical memory best captures the range of available data.

L’objectif du travail de thèse est de proposer un modèle de représentation des écritures dans les images de documents du patrimoine sans recourir à une transcription des textes. Ce modèle, issu d’une étude très complète des méthodes actuelles de caractérisation des écritures, est à la base d’une proposition de scénario de recherche par similarité de mots, indépendante du scripteur et ne nécessitant pas d’apprentissage. La recherche par similarité proposée repose sur une structure de graphes intégrant des informations sur la topologie, la morphologie locale des mots et sur le contexte extrait du voisinage de chaque point d’intérêt. Un graphe est construit à partir du squelette décrit en chaque point sommet par le contexte de formes, descripteur riche et compact. L’extraction de mots est assurée par une première étape de localisation grossière de régions candidates, décrites par une séquence déduite d’une représentation par graphes liée à des critères topologiques de voisinage. L’appariement entre mots repose ensuite sur une distance dynamique et un usage adapté du coût d’édition approximé entre graphes rendant compte de la nature bi-dimensionnelle de l’écriture. L’approche a été conçue pour être robuste aux distorsions de l’écriture et aux changements de scripteurs. Les expérimentations sont réalisées sur des bases de documents manuscrits patrimoniaux exploitées dans les compétitions de word-spotting. Les performances illustrent la pertinence de la proposition et ouvrent des voies nouvelles d’investigation dans des domaines d’applications autour de la reconnaissance de symboles et d’écritures iconographiques
As more and more documents, especially historical handwritten documents, are converted into digitized version for long-term preservation, the demands for efficient information retrieval techniques in such document images are increasing. The objective of this research is to establish an effective representation model for handwriting, especially historical manuscripts. The proposed model is supposed to help the navigation in historical document collections. Specifically speaking, we developed our handwriting representation model with regards to word spotting application. As a specific pattern recognition task, handwritten word spotting faces many challenges such as the high intra-writer and inter-writer variability. Nowadays, it has been admitted that OCR techniques are unsuccessful in handwritten offline documents, especially historical ones. Therefore, the particular characterization and comparison methods dedicated to handwritten word spotting are strongly required. In this work, we explore several techniques that allow the retrieval of singlestyle handwritten document images with query image. The proposed representation model contains two facets of handwriting, morphology and topology. Based on the skeleton of handwriting, graphs are constructed with the structural points as the vertexes and the strokes as the edges. By signing the Shape Context descriptor as the label of vertex, the contextual information of handwriting is also integrated. Moreover, we develop a coarse-to-fine system for the large-scale handwritten word spotting using our representation model. In the coarse selection, graph embedding is adapted with consideration of simple and fast computation. With selected regions of interest, in the fine selection, a specific similarity measure based on graph edit distance is designed. Regarding the importance of the order of handwriting, dynamic time warping assignment with block merging is added. The experimental results using benchmark handwriting datasets demonstrate the power of the proposed representation model and the efficiency of the developed word spotting approach. The main contribution of this work is the proposed graph-based representation model, which realizes a comprehensive description of handwriting, especially historical script. Our structure-based model captures the essential characteristics of handwriting without redundancy, and meanwhile is robust to the intra-variation of handwriting and specific noises. With additional experiments, we have also proved the potential of the proposed representation model in other symbol recognition applications, such as handwritten musical and architectural classification

Le sujet étudié dans cette thèse est celui de l’encodage, de la mémorisation et du rappel des mots écrits. Nous avons abordé ces trois étapes du point de vue de la modélisation connexioniste. Notre méthodologie générale a été d’identifier un certain nombre de contraintes, d’étudier de manière formelle et empirique les systèmes qui pouvaient les satisfaire, avant d’appliquer ces résultats à la modélisation des résultats expérimentaux. La manière dont nous encodons l’ordre des lettres dans un mot est nécessairement très particulière: nous distinguons parfaitement entre deux anagrammes comme «douaniers» et « dinosaure», mais sommes plus aisément trompés par les trasnpositions. La signature de ce format est actuellement très étudiée. L’étude comportementale et l’imagerie cérébrale en révêlent déjà certaines propriétés : la première suggère notemment que l’ordre est encodé de manière relative et non absolue, tandis que la deuxième pointe vers un code distribué et hiérarchique, qui serait computé dans l’aire cérébrale dite de la forme visuelle des mots. Pour rendre compte de ces derniers résultats, nous étudions une technique connexionniste portant le nom de représentations holographiques, dont le format distribué particulier est propice à la construction de structures combinatoires. Nous dérivons formellement certaines de leurs propriétés, avant de traduire les codes existants dans ce nouveau format. Nous montrons alors que l’usage des représentations holographiques améliore globalement le pouvoir explicatif des codes existants, dans certains cas en exhibant des propriétés émergentes rendant compte d’effets comportementaux bien connus. Les études computationelles du système lexical sont aussi remarquablement bien contraintes par des années d’observations comportementales, et par les études d’imagerie cérébrales dont la base de données s’étoffe extrèmement rapidement. L’approche classique conçoit le système lexical comme l’intéraction d’au moins trois systèmes, Orthographe, Phonologie et Sémantique (OPS). Cependant la nature des représentations supportées par ces systèmes, ainsi que la façon dont ils communiquent est très débatue. De plus, deux aspects fondamentaux du problème sont presque toujours passés sous silence: la capacité du système à apprendre dans des conditions bruitées et non-supervisées, ainsi que l’origine de l’effet de fréquence. Un modèle connectioniste supervisé connu pour sa faculté de résistance au bruit est le réseau de Hopfield. Nous investiguons le comportement de ce réseau en situation réaliste du point de vue cognitif, c'est-à-dire en l’exposant à des représentations bruitées et corrélées pendant l’apprentissage. Ce travail aboutit à de nouveaux résultats concernant les capacités de généralisation des réseaux de Hopfield, ainsi qu’à la formulation d’une méthode simple de conditionnement du système. Nous appliquons ensuite ces résultats dans un nouveau modèle de reconnaissance visuelle des mots, dans lequel l’effet de fréquence s’avère être une conséquence du bruit pendant l’apprentissage. En explorant et implémentant l’idée de lexiques distribués, ce modèle est à notre connaisance le premier modèle distribué capable d’accomplir la tâche de décision lexicale sans utiliser d’information sémantique. Finalement nous spéculons sur les corrélats neuronaux de notre travail. Nous établissons plusieurs correspondances, d’abord entre la procédure d’apprentissage du modèle et certaines fonctions qui ont été largement attribuées au système hippocampal, ensuite entre les dynamiques d’attracteurs à l’œuvre dans ce modèle et le large réseau d’aires corticales latéralisé à gauche qui est spécifiquement actif lors du traitement des mots écrits.

In the last few years 3D Virtual Worlds (3DVWs) became a reality. Initially considered as a new mean for social communication, triggered by the development of software and hardware technology, 3DVWs are exposing now different functionalities, experiences and acquaintances. Therefore, they achieved their popularity very fast, indicated by the number and the progression of active users. Being the representation of the user, the avatar is one of the most significant and most complex assets of a Virtual World. A short analysis of a VW content allows one to observe that from the point of view of the storage/transmission the most significant amount is represented by the VW assets. Within the set of assets, the avatars are the most complex structures, consisting of different components: geometry, images, animations, structures, etc. The overall objective of developing tools and methods for a large deployment of VW are translated into three specific ones: To propose a compression framework to enable efficient, compact transfer of avatars, and general 3D graphics assets. Specifically for avatars, the framework should be independent from the representation formalism. To propose an optimized solution making the avatars accessible on weak terminals such as mobile phones. To define a metadata model allowing avatars interoperability between different VWs. The three objectives are addressed in this thesis and for each we propose original contributions.

Au cours des dernières années les Mondes Virtuels 3D (MV3D) sont devenus une réalité. Initialement considérés comme un nouveau moyen de communication sociale, initiés par le développement logiciel et matériel, les MV3Ds révèlent diverses fonctionnalités, des expériences et des connaissances. Étant la représentation de l'utilisateur, l'avatar est l’une des ressources les plus significatives et les plus complexes d'un monde virtuel. Une courte analyse d'un contenu de MV nous informe que le stockage/transmission est la partie la plus significative des ressources du MV. Dans l’ensemble des ressources, les avatars sont les structures les plus complexes, représentés par différents composants : géométrie, images, animations, etc. Donc, en abordant les problématiques de compression, d'adaptation et d'interopérabilité liées aux avatars, nous traitons implicitement presque tout type de ressources tifs qu’on pourrait retrouver dans les MVs. L'objectif global du développement d’outils et des méthodes, pour un déploiement élargi de MV, est traduit par les trois points spécifiques suivants: Proposer un framework de compression pour permettre le transfert efficace et compact d'avatars et de ressources graphiques 3D généraux. Spécifiquement pour des avatars, le framework devrait être indépendante avec un formalisme de représentation. Proposer une solution optimisée permettant l’accessibilité aux avatars sur des terminaux à faibles ressources tels que les téléphones portables. Définir un modèle de métadonnées permettant l'interopérabilité d'avatars entre différents MVs. Les trois objectifs sont traités dans ce manuscrit et pour chacun nous proposons des contributions originales
In the last few years 3D Virtual Worlds (3DVWs) became a reality. Initially considered as a new mean for social communication, triggered by the development of software and hardware technology, 3DVWs are exposing now different functionalities, experiences and acquaintances. Therefore, they achieved their popularity very fast, indicated by the number and the progression of active users. Being the representation of the user, the avatar is one of the most significant and most complex assets of a Virtual World. A short analysis of a VW content allows one to observe that from the point of view of the storage/transmission the most significant amount is represented by the VW assets. Within the set of assets, the avatars are the most complex structures, consisting of different components: geometry, images, animations, structures, etc. The overall objective of developing tools and methods for a large deployment of VW are translated into three specific ones: To propose a compression framework to enable efficient, compact transfer of avatars, and general 3D graphics assets. Specifically for avatars, the framework should be independent from the representation formalism. To propose an optimized solution making the avatars accessible on weak terminals such as mobile phones. To define a metadata model allowing avatars interoperability between different VWs. The three objectives are addressed in this thesis and for each we propose original contributions

Jorge Semprún’s work is the fruit of an incarceration in the concentration camp of Buchenwald as a resistance fighter and his expulsion from the Partido Comunista Español in 1964. Due to these biographical circumstances, many critical literary studies to date limit the discussion of his works to the autobiographical and the realm of Holocaust studies. Together with the texts that do not fit adequately into this categories, his self-identification as a Spanish exile has up to now been neglected. The present thesis aims to provide a more global view of his oeuvre by extending the literary analyses to texts that have deserved little critical attention. In order to achieve this, it investigates the role played by memory and self-representation in a variety of works by Semprún. Aspects connected to memory such as exile and nostalgia, the Holocaust, the interplay between memory and writing, politics and collective memory, postmemory and identity are examined by means of a detailed analysis of the selected works and are discussed thematically. Differences in genre are discarded for the discussion and interconnections between the various narratives are highlighted. With the help of memory and trauma theories, we come to the conclusion that memory is the overarching principle of Semprún’s writing and that he invests it with an aesthetic and ethical value which is interpreted as the justification for his devotion to writing.

This thesis describes the representation of Christianity in the writings of John Berryman-his struggle with the faith being the most central and incessant preoccupation of his verse. Focussing on each major stage of his artistic development in tum, I demonstrate how its depiction is influenced by biographical factors, his scholarship and sources, and his evolving poetic style. In The Dispossessed the issue of faith is evident, but obscured; however, much of his unpublished verse of the period is characterised by a more transparent confessional idiom, frequently expressing his dilemma of conscience over the question of religious commitment. His failure to develop an effective poetic voice is the main reason why his religious poetry of the 1930s and 1940s remained in the private sphere. He achieved his stylistic breakthrough with Berryman's Sonnets, where the struggle with his conscience is depicted as a religious conflict, in which his adultery means a confrontation with the Law of God. Homage to Mistress Bradstreet features a more developed representation of a similar conflict; the two alternative life choices before him are personified in the characters of Anne Bradstreet and the 'poet'. Difficulties of faith continue to play a major role in The Dream Songs, where the poet, adopting the persona of Henry, directly confronts God and Christianity with the problem of evil and the historical quest for Jesus. His poetry portrays a perceived conflict between faith and reason, and an intellectual pursuit for the truth epitomised by his poem 'The Search'. However, the poet's 'conversion experience' during the composition of Love & Fame is depicted as a response to the direct intervention of God in his life. His subsequent devotional poetry is dominated by his new sense of relationship with the' God of Rescue', who increasingly becomes associated with the full Christian conception of Jesus Christ the Saviour.