In this thesis, we build upon the work of Plate by advancing the theory and utility of Holographic Reduced Representations (HRRs). HRRs are a type of linear, associative memory developed by Plate and are an implementation of Hinton’s reduced representations. HRRs and HRR-like representations have been used to model human memory, to model understanding analogies, and to model the semantics of natural language. However, in previous research, HRRs are restricted to storing and retrieving vectors of random numbers, limiting both the ability of HRRs to model human performance in detail, and the potential applications of HRRs. We delve into the theory of HRRs and develop techniques to store and retrieve images, or other kinds of structured data, in an HRR. We also investigate square matrix representations as an alternative to HRRs, and use iterative training algorithms to improve HRR performance. This work provides a foundation for cognitive modellers and computer scientists to explore new applications of HRRs. / Thesis (Master, Computing) -- Queen's University, 2010-08-10 12:50:04.004

No description available.

Individuals are generating, storing and accessing more information than ever before. The information comes from a variety of sources such as the World Wide Web, email and books. Storage media is becoming larger and cheaper. This makes accumulation of information easy. When information is kept in large volumes, retrieving it becomes a problem unless there is a system in place for managing this. This study examined the techniques that users have devised to make retrieval of their documents easy and timely. A survey of user document management techniques was done through interviews. The uncovered techniques were then used to build an expert system that provides assistance with document management decision-making. The system provides recommendations on file naming and organization, document backup and archiving as well as suitable storage media. The system poses a series of questions to the user and offers recommendations on the basis of the responses given. The system was evaluated by two categories of users: those who had been interviewed during data collection and those who had not been interviewed. Both categories of users found the recommendations made by the system to be reasonable and indicated that the system was easy to use. Some users thought the system could be of great benefit to people new to computers.

Since the terrorist attacks of September 11th, 2001, a vast discourse on contemporary terrorism has emerged within 'Western' media. This thesis analyzes the discourse of contemporary terrorism, and highlights how the postcolonialist critique of Orientalism pioneered by Edward Said is still relevant within the discourse. This is accomplished by analyzing books that have been published post-9/11 and which have been reviewed in the journal Foreign Affairs. A primary goal of this thesis is to facilitate the de-reification of the socially constructed concepts of both 'the East' and 'the West' which currently dominate representations within the discourse, as well as to highlight some of the key features of the discursive field on contemporary terrorism. The binary representation and stereotyping within the contemporary discourse provides a one-dimensional representation of the issue of terrorism, and by questioning the conformity of these representations we can critically examine one of the most important social issues within our society.

This study investigated whether or not variability in the severity of speech sound disorders is related to variability in phonological short-term memory and/or variability in the accuracy of phonological representations. The aim was to determine speech processing predictors of severity of speech sound disorders. A total of 33 children, aged three to six years of age, were assessed on measures of nonword repetition, accuracy of phonological representations, accuracy of speech production, and language. The tests administered included the Clinical Evaluation of Language Fundamentals Preschool – 2 Australian, the Diagnostic Evaluation of Articulation and Phonology, the Nonword Repetition Test (modified), and the Phonological Representation Judgement Task (modified). The relationships between the results of these tests were established using a correlation analysis. The relationship between accuracy of phonological representations and the percentage of consonants correct was found to be mediated by language. There was no significant relationship between nonword repetition and percentage consonants correct. These findings may have been the result of small sample size, age of the participants, or co-morbid language difficulties. These findings imply that variability in severity of speech sound disorders may be related to a variable not directly assessed in this study. This variable may be a constraint relating to the stored motor programs within children’s speech processing systems. Implications for future research are discussed.

Cette thèse s'inscrit dans le cadre de la correspondance de Langlands locale $p$-adique, imaginée par Breuil et établie par Colmez pour GL_2(Q_p). Soit L une extension finie de Q_p et soit V une L-représentation irréductible du groupe de Galois absolu de Q_p, de dimension 2. En utilisant la théorie des (phi,Gamma)-modules de Fontaine, Colmez associe à V une GL_2(Q_p)-représentation de Banach Pi(V), unitaire, admissible, topologiquement irréductible. On donne une nouvelle preuve, nettement plus simple, d'un théorème de Colmez, qui permet de décrire les vecteurs localement analytiques Pi(V)^an de Pi(V) en fonction du (phi,\Gamma)-module surconvergent attaché à V. Le résultat principal de cette thèse est une description simple de l'action infinitésimale de GL_2(Q_p) sur Pi(V)^an. En particulier, on montre que Pi(V)^an admet un caractère infinitésimal, que l'on peut calculer en fonction des poids de Hodge-Tate de V, ce qui répond à une question de Harris. En utilisant ces résultats, on montre aussi l'absence d'un analogue p-adique d'un théorème classique de Tunnell et Saito, répondant à une autre question de Harris. Nous étendons et précisons certains résultats de Colmez concernant le modèle de Kirillov des vecteurs U-finis de Pi(V) (U est l'unipotent supérieur de GL_2(Q_p)). En combinant cette étude avec la description de l'action infinitésimale, on obtient une démonstration simple d'un des résultats principaux de Colmez, caractérisant les représentations V telles que Pi(V) possède des vecteurs localement algébriques non nuls. Ce résultat permet de faire le pont avec la correspondance classique et est un des ingrédients clés de la preuve d'Emerton de la conjecture de Fontaine-Mazur en dimension 2. On étend nos méthodes pour démontrer l'analogue de ce résultat pour les déformations infinitésimales de V. Cela répond à une question de Paskunas et a des applications à la conjecture de Breuil-Mézard. Une autre application est l'étude du module de Jacquet de Pi(V)^an. On montre qu'il est non nul si et seulement si V est trianguline, ce qui permet de donner une preuve simple des conjectures de Berger, Breuil et Emerton. Enfin, dans un travail en collaboration avec Benjamin Schraen, nous démontrons le lemme de Schur pour les représentations de Banach et localement analytiques topologiquement irréductibles d'un groupe de Lie p-adique. Ce résultat basique n'était connu que pour des groupes de Lie commutatifs et pour GL_2(Q_p).

A central problem in artificial intelligence is to choose actions to maximize reward in a partially observable, uncertain environment. To do so, we must obtain an accurate environment model, and then plan to maximize reward. However, for complex domains, specifying a model by hand can be a time consuming process. This motivates an alternative approach: learning a model directly from observations. Unfortunately, learning algorithms often recover a model that is too inaccurate to support planning or too large and complex for planning to succeed; or, they require excessive prior domain knowledge or fail to provide guarantees such as statistical consistency. To address this gap, we propose spectral subspace identification algorithms which provably learn compact, accurate, predictive models of partially observable dynamical systems directly from sequences of action-observation pairs. Our research agenda includes several variations of this general approach: spectral methods for classical models like Kalman filters and hidden Markov models, batch algorithms and online algorithms, and kernel-based algorithms for learning models in high- and infinite-dimensional feature spaces. All of these approaches share a common framework: the model’s belief space is represented as predictions of observable quantities and spectral algorithms are applied to learn the model parameters. Unlike the popular EM algorithm, spectral learning algorithms are statistically consistent, computationally efficient, and easy to implement using established matrixalgebra techniques. We evaluate our learning algorithms on a series of prediction and planning tasks involving simulated data and real robotic systems.

Pornografijos įtaka asmenybei – tai vienas iš prieštaringiausių klausimų, atkreipusių mokslininkų dėmesį per pastaruosius dešimtmečius. Didžioji dalis tyrimų, nagrinėjančių pornografijos poveikį žmogui, tyrinėja seksualiai atviros medžiagos ryšį su elgesiu. Tačiau ne mažiau svarbu suprasti, ar pornografijos žiūrėjimas įtakoja žmogaus seksualinę Aš-koncepciją ir Aš-reprezentacijas, kadangi šios struktūros veikia ne tik žmogaus elgesį, bet ir kitus bei intrapersonalinius bei tarpasmeninius žmogaus procesus. Atliktame tyrime siekėme išsiaiškinti kaip pornografijos žiūrėjimas įtakoja seksualinės Aš koncepcijos aspektus ir žmogaus dabartinio, idealaus ir privalomojo seksualinio Aš reprezentacijas (iš savo ir reikšmingojo kito pozicijos). Tyrimo metu buvo įvertinti 70 tiriamųjų (36 vyrai ir 34 moterys), kurie turi seksualinius partnerius ir žiūri pornografinę medžiagą internete. Įvertinimui naudoti šie metodai: „Rizikingo pornografijos žiūrėjimo anketa“, „Multidimencinis Aš-koncepcijos klausimynas“ ir struktūruotas interviu. Nustatyta, kad vyrai daugiau laiko skiria pornografijos žiūrėjimui, negu moterys, nors patiriamų rizikingo pornografijos žiūrėjimo požymių stiprumas tarp abiejų lyčių nesiskiria. Nustatytas ryšys tarp rizikingo pornografijos žiūrėjimo požymių stiprumo ir šių seksualinės Aš-koncepcijos ypatumų: susirūpinimo seksu, troškimo įsitraukti į lytinius santykius, polinkio rūpintis savo seksualiniu įvaizdžiu ir galvojimo apie savo seksualinį gyvenimą, įsitikinimų, kad... [toliau žr. visą tekstą] / The impact of pornography on human personality is one of the most controversial questions, attracted most scientists’ attention over past decades. Vast majority of research, exploring pornography influence on personality, investigate relation of sexually explicit material and behavior. However, it is worth pointing out, how pornography influences sexual Self-concept and Self-representations. This research investigates the relation between consuming of sexually explicit material and sexual Self-concept. What is more, we tried to identify the impact of pornography viewing on Real, Ideal and Ought Self-representations (from self and significant other perspectives). Sample consisted of 70 subjects (34 women, 36 men), who announced to watch pornography on Internet and used to have an intimate partner. Three methods where used to evaluate the aspects of Self-concept and Self-representations in relation to pornography use: “Problematic pornography watching scale”, “Multidimensional sexual Self-concept Questionnaire” and structured interview, evaluating Real, Ideal and Ought Self-representations. It was found significant difference comparing male and female time spent on watching sexually explicit material, although no significant difference was found between gender and problematic pornography watching symptoms. It was found, that subjects, scored high on Problematic pornography watching scale tend to be preoccupied with sex and constantly think about aspects of their sexual life... [to full text]

Actions are goal-directed. It can be the goal of an action to change the environment (i.e. to produce an effect), but also to change one´s own situation in the environment (i.e. to move to a physical target). Previous research has shown that kinematics of actions directed towards physical targets are not only mere reactions to such targets. Instead, targets evoke intentional goals. Representations of such intentional goals influence action execution. However, thus far, most studies in the context of the ideomotor theory of action control have focused on the influence of anticipated action effects on action planning. The role of targets as action goals as well as the role of goal anticipations on overt action execution has mostly been neglected. In this dissertation the role of goal representations in action control was investigated. The ideomotor theory served as a theoretical framework. It was assumed that targets function as action goals similar to action effects and that action goals influence action execution by the anticipation of upcoming events. Action execution towards targets and towards effects was compared. This was done in the temporal and the spatial domain. Furthermore, goal representations were manipulated in order to evaluate their influence on action execution and to disentangle the role of physical target characteristics and the role of goal representations. The findings obtained strengthen the assumption that goal representations play an important role in action control. First, both targets and effects can be viewed as goals of an action in the temporal and spatial domain. Second, movement kinematics are shaped by the way targets are represented as action goals, rather than by physically target properties. In conclusion, as goal representations are formed before the action is actually executed they influence action execution by the anticipation of upcoming events. The ideomotor theory of action control should incorporate action targets as goals similar to action effects.

This study contends that the spaces where crime occurs in films are not neutral; they are layered with maps of meaning that we construct somewhere between the imagined and the lived. Given that popular cultural representations both shape and reflect our understandings of crime and space, a study examining where crime occurs in films was warranted but previously unrealized in the criminological literature. This study addresses this gap in the literature by considering how geographies of crime are characterized in a sample of ten recent popular crime films. Applying a qualitative content analysis approach, this study foregrounds the onscreen spaces where crimes occur in an attempt to expose and denaturalize the meanings around crime that are embedded in these backgrounds. Particular regard is given here to the twinning of crime and urbanity, the aesthetics of insecurity, and the gendering of geographies of crime.