Дисертації з теми "Representation space / Latent space"

Apprendre à éditer des images et des vidéos de visages est un domaine particulièrement actif dans la recherche académique et industrielle. Cette thèse aborde le problème de l'édition de visages dans le cas particulier des images et des vidéos à haute résolution. Dans cette thèse, nous développons des méthodes basées sur l'apprentissage profond pour effectuer l'édition d'images faciales. Plus précisément, nous explorons la tâche en utilisant les représentations latentes obtenues à partir de deux types de réseaux neuronaux profonds : les modèles basés sur l'auto-encodage et les réseaux antagonistes génératifs (GAN). Pour chaque type de méthode, nous considérons un problème spécifique d'édition d'image et proposons une solution efficace qui surpasse l'état de l'art. La thèse comprend deux parties. Dans la partie I, nous explorons les tâches d'édition d'images via l'espace latent des autoencodeurs. Nous considérons d'abord la tâche de transfert de style entre les photos, et proposons un algorithme efficace qui est construit sur une paire de réseaux basés sur des autoencodeurs. Ensuite, nous étudions la tâche d'édition de l'âge du visage pour les images à haute résolution, en utilisant une architecture d'encodeur-décodeur. Le réseau proposé encode une image de visage en représentations de caractéristiques invariantes selon l'âge, et apprend un vecteur de modulation correspondant à un âge cible. Notre approche permet une édition fine de l'âge sur des images à haute résolution dans un seul modèle unifié.Dans la deuxième partie, nous explorons la tâche d'édition via l'espace latent des modèles antagonistes génératifs (GAN). Tout d'abord, nous considérons le problème de l'édition "démêlée" (disentangled) des attributs faciaux sur des images synthétiques et réelles, en proposant un réseau de transformation latent qui agit dans l'espace latent d'un modèle GAN pré-entraîné. Nous avons également proposé un pipeline de manipulation vidéo, afin de généraliser le résultat de l'édition aux vidéos. Deuxièmement, nous étudions le problème de l'inversion du GAN - la projection d'une image réelle dans l'espace latent d'un GAN pré-entraîné. En particulier, nous proposons un encodeur feed-forward, qui encode une image donnée en un code caractéristique et un code latent en une seule passe. L'encodeur proposé s'avère plus précis et plus stable pour l'inversion d'images et de vidéos, tout en conservant de bonnes capacités d'édition
Learning to edit facial images and videos is one of the most popular tasks in both academia and industrial research. This thesis addresses the problem of face editing for the special case of high-resolution images and videos.In this thesis, we develop deep learning-based methods to perform facial image editing. Specifically, we explore the task using the latent representations obtained from two types of deep neural networks: autoencoder-based models and generative adversarial networks. For each type of method, we consider a specific image editing problem and propose an effective solution that outperforms the state-of-the-art.The thesis contains two parts. In part I, we explore image editing tasks via the latent space of autoencoders. We first consider the style transfer task between photos and propose an effective algorithm that is built on a pair of autoencoder-based networks. Second, we study the face age editing task for high-resolution images, using an encoder-decoder architecture. The proposed network encodes a face image to age-invariant feature representations and learns a modulation vector corresponding to a target age. Our approach allows for fine-grained age editing on high-resolution images in a single unified model.In part II, we explore the editing task via the latent space of generative adversarial models (GANs). First, we consider the problem of facial attribute disentangled editing on synthetic and real images, by proposing a latent transformation network that acts in the latent space of a pre-trained GAN model. We also proposed a video manipulation pipeline, to generalize the editing result to videos. Second, we investigate the problem of GAN inversion -- the projection of a real image to the latent space of a pretrained GAN. In particular, we propose a feed-forward encoder, which encodes a given image to a feature code and a latent code in one pass. The proposed encoder is shown to be more accurate and stable for image and video inversion, meanwhile, maintaining good editing capacities

Un élément clé du récent succès des modèles d'apprentissage profond de traitement du langage réside dans la capacité à apprendre des "embeddings" de mots efficaces. Ces méthodes fournissent des espaces vectoriels structurés de dimension réduite ayant des relations métriques intéressantes. Ceux-ci, à leur tour, peuvent être utilisés comme des représentations d'entrées efficaces pour traiter des tâches plus complexes. Dans cette thèse, nous nous concentrons sur la tâche d'apprentissage d'espaces "d'embedding" pour la musique polyphonique dans le domaine symbolique. Pour ce faire, nous explorons deux approches différentes.Tout d'abord, nous introduisons un modèle d'embedding basé sur un réseau convolutif avec un nouveau type de mécanisme d'attention hiérarchique auto-modulée, qui est calculé à chaque couche afin d'obtenir une vision hiérarchique de l'information musicale.Puis, nous proposons un autre système basé sur les VAE, un type d'auto-encodeur qui contraint la distribution des données de l'espace latent à être proche d'une distribution préalablement choisie. La musique polyphonique étant un type d'information complexe, le choix de la représentation d'entrée est un processus crucial. Nous introduisons donc une nouvelle représentation de données musicales symboliques, qui transforme une partition polyphonique en un signal continu.Enfin, nous montrons le potentiel de nos espaces d'embedding à travers le développement de plusieurs applications créatives utilisées pour améliorer la connaissance et l'expression musicales, à travers des tâches telles que la modification de mélodies ou l'identification de compositeurs
A key part in the recent success of deep language processing models lies in the ability to learn efficient word embeddings. These methods provide structured spaces of reduced dimensionality with interesting metric relationship properties. These, in turn, can be used as efficient input representations for handling more complex tasks. In this thesis, we focus on the task of learning embedding spaces for polyphonic music in the symbolic domain. To do so, we explore two different approaches.We introduce an embedding model based on a convolutional network with a novel type of self-modulated hierarchical attention, which is computed at each layer to obtain a hierarchical vision of musical information.Then, we propose another system based on VAEs, a type of auto-encoder that constrains the data distribution of the latent space to be close to a prior distribution. As polyphonic music information is very complex, the design of input representation is a crucial process. Hence, we introduce a novel representation of symbolic music data, which transforms a polyphonic score into a continuous signal.Finally, we show the potential of the resulting embedding spaces through the development of several creative applications used to enhance musical knowledge and expression, through tasks such as melodies modification or composer identification

This report shows how knowledge about the visual world can be built into a shape representation in the form of a descriptive vocabulary making explicit the important geometrical relationships comprising objects' shapes. Two computational tools are offered: (1) Shapestokens are placed on a Scale-Space Blackboard, (2) Dimensionality-reduction captures deformation classes in configurations of tokens. Knowledge lies in the token types and deformation classes tailored to the constraints and regularities ofparticular shape worlds. A hierarchical shape vocabulary has been implemented supporting several later visual tasks in the two-dimensional shape domain of the dorsal fins of fishes.

Thesis: S.M. in Engineering and Management, Massachusetts Institute of Technology, School of Engineering, System Design and Management Program, Engineering and Management Program, 2016.
Cataloged from PDF version of thesis.
Includes bibliographical references (pages 95-98).
In this thesis, I explore a novel algorithm to model the joint behavior of multiple correlated signals. Our chosen example is the ECG (Electrocardiogram) and ABP (Arterial Blood Pressure) signals from patients in the ICU (Intensive Care Unit). I then use the generated models to predict blood pressure levels of ICU patients based on their historical ECG and ABP signals. The algorithm used is a variant of a Hidden Markov model. The new extension is termed as the Latent State Space Copula Model. In the novel Latent State Space Copula Modelthe ECG, ABP signals are considered to be correlated and are modeled using a bivariate Gaussian copula with Weibull marginals generated by a hidden state. We assume that there are hidden patient "states" that transition from one hidden state to another driving a joint ECG-ABP behavior. We estimate the parameters of the model using a novel Gibbs sampling approach. Using this model, we generate predictors that are the state probabilities at any given time step and use them to predict a patient's future health condition. The predictions made by the model are binary and detects whether the Mean arterial pressure(MAP) is going to be above or below a certain threshold at a future time step. Towards the end of the thesis I do a comparison between the new Latent State Space Copula Model and a state of the art Classical Discrete HMM. The Latent State Space Copula Model achieves an Area Under the ROC (AUROC) curve of .7917 for 5 states while the Classical Discrete HMM achieves an AUROC of .7609 for 5 states.
by Prashan Wanigasekara.
S.M. in Engineering and Management

L'évolution rapide des modèles de sous-titrage d'images, impulsée par l'intégration de techniques d'apprentissage profond combinant les modalités image et texte, a conduit à des systèmes de plus en plus complexes. Cependant, ces modèles fonctionnent souvent comme des boîtes noires, incapables de fournir des explications transparentes de leurs décisions. Cette thèse aborde l'explicabilité des systèmes de sous-titrage d'images basés sur des architectures Encodeur-Attention-Décodeur, et ce à travers quatre aspects. Premièrement, elle explore le concept d'espace latent, s'éloignant ainsi des approches traditionnelles basées sur l'espace de représentation originel. Deuxièmement, elle présente la notion de caractère décisif, conduisant à la formulation d'une nouvelle définition pour le concept d'influence/décisivité des composants dans le contexte de sous-titrage d'images explicable, ainsi qu'une approche par perturbation pour la capture du caractère décisif. Le troisième aspect vise à élucider les facteurs influençant la qualité des explications, en mettant l'accent sur la portée des méthodes d'explication. En conséquence, des variantes basées sur l'espace latent de méthodes d'explication bien établies telles que LRP et LIME ont été développées, ainsi que la proposition d'une approche d'évaluation centrée sur l'espace latent, connue sous le nom d'Ablation Latente. Le quatrième aspect de ce travail consiste à examiner ce que nous appelons la saillance et la représentation de certains concepts visuels, tels que la quantité d'objets, à différents niveaux de l'architecture de sous-titrage
The rapid advancement of image captioning models, driven by the integration of deep learning techniques that combine image and text modalities, has resulted in increasingly complex systems. However, these models often operate as black boxes, lacking the ability to provide transparent explanations for their decisions. This thesis addresses the explainability of image captioning systems based on Encoder-Attention-Decoder architectures, through four aspects. First, it explores the concept of the latent space, marking a departure from traditional approaches relying on the original representation space. Second, it introduces the notion of decisiveness, leading to the formulation of a new definition for the concept of component influence/decisiveness in the context of explainable image captioning, as well as a perturbation-based approach to capturing decisiveness. The third aspect aims to elucidate the factors influencing explanation quality, in particular the scope of explanation methods. Accordingly, latent-based variants of well-established explanation methods such as LRP and LIME have been developed, along with the introduction of a latent-centered evaluation approach called Latent Ablation. The fourth aspect of this work involves investigating what we call saliency and the representation of certain visual concepts, such as object quantity, at different levels of the captioning architecture

GPS systems have been in use for navigational purposes for almost three decades and have found their way into location tracking systems. The principal, Scania has a fleet of 300 thousand connected vehicles sending in position information. In this paper, we make use of position information sourced from Scania’s connected trucks, which have been abstracted into stops. The stop abstractions are built by coalescing raw position information on the basis of temporal and spatial thresholds and denote locations where the truck halts. We apply unsupervised machine learning approaches to try and understand the semantics behind these stops. The features of the truck stops are projected into a low dimensional latent space using deep autoencoders, and a clustering objective is then optimized in this low dimension space. The resultant clusters are found to be representative of different types of truck stops. The characterized truck stoppages can be useful for understanding the truck usage patterns as well transport hub usage statistics.
GPS-systemen för navigation har funnits i nästan tre årtionden och de kan numera även hittas i lokaliseringssystem. Scania har en flotta bestående av 300 000 uppkopplade fordon som skickar information om deras position till Scania. I den här masteruppsatsen används positionsinformationen från de Scaniafordon som klassificeras som stillastående. Denna klacificering bygger på rå positionsinformation som baserat på tid och rum inte får variera mer än vissa tröskelvärden och de beskriver därigenom platser där lastbilar har stannat. En oövervakad maskininlärningsmetod användes för att försöka förstå semantiken bakom dessa stillaståenden. Data från lastbilarna projiceras till ett lägre dimensionellt rum med hjälp av deep autoencoders och klustringen optimeras sedan fram i denna lägre dimension. Klustringen har i denna masteruppsats visat sig respresentativ för olika anledningar till stillastående lastbilar. Detta kan vara användbart för att förstå användarmönster men även förtransportsnavets användarstatistik.

My area of research is best described as the promotion of a new methodological approach to the study of film. It is an approach that is founded upon a spatial reading, based on an exploration into abstract aesthetics and pays particular attention to the interactions between sound and image. Whilst this methodological approach can be utilised to read any film, this thesis looks particularly at the musical genre and, more specifically, the musical number and its representation of space. The need to delimit my study notwithstanding, the musical has been taken as a case study in order to demonstrate how this spatial methodology should pay attention to, and be aware of, the peculiarities and idiosyncrasies that genres encapsulate. My thesis challenges the dominance of cognitive theory by providing an approach based upon gestalt theory, making use of ‘forensic’ analysis to remove aesthetics from their narrative context. Theorists such as Rick Altman (1989, 1999) and Jane Feuer (1993) have long discussed the structural qualities of the musical genre in an attempt to delimit the musical number from that which surrounds it. A different representation of space emerges in the musical number, one that permits a deeper exploration into the negotiated relationship between sound and image. In this thesis I examine this space closely utilising a range of innovative analytical techniques including virtual reconstruction and diagrammatic notation. This ‘forensic’ analysis is considered within the overarching framework of gestalt theory: that the whole is more than the sum of its parts. This thesis studies film as an audio-visual medium and considers a range of different spatial realms in order to best understand the complex negotiations between sound and image. Previous scholars of the musical genre have largely focused upon narrative readings of either new or canonical filmic texts. I argue that these narrative readings, whilst 4 providing significant contributions to the field, are ultimately deficient as they fail to adequately explore the finer qualities of film language.

Control predictivo basado en modelos (MPC) es una metodología de control ampliamente utilizada en la industria por su habilidad para controlar procesos multivariable con restricciones en sus entradas y sus salidas. Se distinguen dos fases en la implementación de MPC: identificación y control. El propósito de esta tesis es doble: realizar contribuciones en la identificación para MPC y proponer una nueva metodología de control MPC. La respuesta en bucle cerrado de una implementación de MPC depende, en gran medida, de la capacidad de predicción del modelo; luego la identificación del modelo es un punto crucial en MPC y la parte que a menudo exige la mayor parte del tiempo del proyecto. El primer objetivo que cubre la tesis es la identificación para MPC. Puesto que un modelo es una aproximación del comportamiento de un proceso, dicha aproximación se puede hacer teniendo en cuenta el fin que se le va a dar al modelo. En MPC, el modelo se utiliza para realizar predicciones dentro de una ventana futura, luego la identificación para MPC (MRI) tiene en cuenta dicho uso del modelo y considera los errores de predicción dentro de dicha ventana para el ajuste de los parámetros del modelo. En esta tesis, se cubren tres temas dentro de MRI. Primero se define MRI y las distintas formas de abordarlo. Luego se compara en términos de MRI el ajuste de un modelo con múltiples entradas y múltiples salidas con el ajuste de varios modelos con múltiples entradas y una salida concluyendo que el ajuste de un único modelo con múltiples entradas y múltiples salidas proporciona mejores resultados en términos de MRI para horizontes de predicción lo suficientemente grandes. Por último, se propone el algoritmo PLS-PH para implementar MRI con modelos paramétricos en el caso de correlación en los datos de identificación. PLS-PH es un método de optimización numérica por búsqueda lineal basado en PLS (mínimos cuadrados parciales). Se muestra en un ejemplo como PLS-PH es capaz de proporcionar mejores modelos que las técnicas convencionales de MRI en modelos paramétricos en el caso de correlación en los datos de identi ficación. Una vez obtenido el modelo se puede formular el controlador predictivo. En esta tesis se propone LV-MPC, un controlador predictivo para procesos continuos que implementa la optimización en el espacio de las componentes principales.
Laurí Pla, D. (2012). MPC: Relevant Identification and Control in the Latent Variable Space [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/15178
Palancia

The ultimate goal of any DSM is a scalable and accurate representation of lexical semantics. Recent developments due to Bullinaria & Levy (2012) and Caron (2001) indicate that the accuracy of such models can be improved by redistribution of weight on the principal components. However, this method is poorly understood and barely replicated due to the computational expensive dimension reduction and the puzzling nature of the results. This thesis aims to explore the nature of these results. Beginning by reproducing the results in Bullinaria & Levy (2012) we move onto deepen the understanding of these results, quantitatively as well as qualitatively, using various forms of the BLESS test and juxtapose these with previous results.  The main result of this thesis is the verification of the 100% score on the TOEFL test and 91.5% on a paradigmatic version of the BLESS test. Our qualitative tests indicate that the redistribution of weight away from the first principal components is slightly different between word categories and hence the improvement in the TOEFL and BLESS results. While we do not find any significant relation between word frequencies and weight distribution, we find an empirical relation for the optimal weight distribution. Based on these results, we suggest a range of further studies to better understand these phenomena.
Målet med alla semantiska fördelningsmodeller (DSMs) är en skalbaroch precis representation av semantiska relationer. Nya rön från Bullinaria & Levy (2012) och Caron (2001) indikerar att man kan förbättra prestandan avsevärt genom att omfördela vikten ifrån principalkomponenterna med störst varians mot de lägre. Varför metoden fungerar är dock fortfarande oklart, delvis på grund av höga beräkningskostnader för PCA men även på grund av att resultaten strider mot tidigare praxis. Vi börjar med att replikera resultaten i Bullinaria & Levy (2012) för att sedan fördjupa oss i resultaten, både kvantitativt och kvalitativt, genom att använda oss av BLESS testet. Huvudresultaten av denna studie är verifiering av 100% på TOEFL testet och ett nytt resultat på en paradigmatisk variant av BLESStestet på 91.5%. Våra resultat tyder på att en omfördelning av vikten ifrån de första principalkomponenterna leder till en förändring i fördelningensins emellan de semantiska relationerna vilket delvis förklarar förbättringen i TOEFL resultaten. Vidare finner vi i enlighet med tidigare resultat ingen signifikant relation mellan ordfrekvenser och viktomfördelning. Utifrån dessa resultat föreslår vi en rad experiment som kan ge vidare insikt till dessa intressanta resultat.

Visual information is processed by the brain in a large number of functional sites across a network of anatomically separate areas. In order to guide coherent behaviour, visual attention is required to select and integrate information regarding the spatial and perceptual attributes of separate objects from the numerous areas involved in their representation. The empirical work reported in this thesis investigates the role of spatial information in guiding this process and considers the different types of representation that may be involved. Using an experimental paradigm designed to disambiguate priming in egocentric and allocentric coordinates, the thesis contrasts the predictions of location and object-based models of attention across a series of experiments that manipulate the way attention is oriented to the location or identity of objects in the visual scene. Initial chapters investigate the distinction between exogenous and endogenous attention and its implication for the coordinate frame in which selection occurs. Subsequent chapters investigate the role of non-spatial attributes such as colour differentiation and grouping in determining the nature of spatial representation underlying shifts of attention as well as spatial-temporal constraints on object-based priming. The results across the thesis are inconsistent with the distinction imposed by space and object-based models of attention and instead support a more flexible account in which attentional mechanisms activate representations that combine non-spatial and spatial information about localised objects at a number of levels of spatial description.

Extensible Markup Language (XML) is a multi-purpose text-based format, used for storage, transmission and manipulation of data. XML documents are often held in main memory and processed via standard interfaces such as the Document Object Model (DOM). However, XML is inherently verbose, and the in-memory representation of XML documents by existing DOM implementations is up to ten times larger than the file size. This is a problem for machines with limited memory, such as mobile devices, where processing even moderately-sized XML documents requires more memory than is available. We focus on in-memory representations of XML documents for situations where space is limited and where rapid processing time is important. We propose a compact representation of XML documents that uses succinct or highly space-efficient data structures, that allows XML processing to be executed efficiently. Succinct data structures use space that approaches the information-theoretic lower bound on the space that is required to represent the data, and support operations upon the representation in constant time. In the context of XML documents, we study and improve succinct representations for ordinal trees by adding features that make them more suitable for use in XML documents. We explore fast and space-efficient representations of the textual data of XML documents. Our basic approach is to concatenate all the textual data in the XML document into a single string, and extract individual textual values by computing the appropriate substring of the concatenated string. Computing the substring requires us to store offsets into the text. The storage of the offsets is surprisingly expensive, if stored naively (as 32 or 64-bit integer values). We give a succinct representation and provide data-aware representations (adapted from work on inverted indices in information retrieval), and show their close connection. We describe Succinct DOM (SDOM), which is a DOM implementation that has low, stable and predictable memory usage. We show, via an experimental evaluation, that SDOM is extremely fast. A variant, SDOM-CT, applies BZip-based compression to textual and attribute data, and its space usage is comparable with “query-friendly” XML compressors. Some of these compressors support navigation and/or querying (e.g. subpath queries) of the compressed file. SDOM-CT does not support querying directly, but remains extremely fast: it is several orders of magnitude faster for navigation than query-friendly XML compressors that support navigation (and only a few times slower than popular DOM implementations such as the Apache Foundation’s Xerces-C).

PONTIFÍCIA UNIVERSIDADE CATÓLICA DO RIO DE JANEIRO
COORDENAÇÃO DE APERFEIÇOAMENTO DO PESSOAL DE ENSINO SUPERIOR
FUNDAÇÃO DE APOIO À PESQUISA DO ESTADO DO RIO DE JANEIRO
PROGRAMA DE EXCELENCIA ACADEMICA
BOLSA NOTA 10
Esta tese é um estudo sobre memórias dos espaços escolares vividos por professores. Procura identificar as condições e dimensões do trabalho docente a partir do espaço escolar e memórias sobre a jornada educativa de dezessete professores de Ensino Superior e Ensino Médio de um município de Minas Gerais As memórias foram colhidas por meio de entrevista semiestruturada. Essas memórias de um tempo e espaço vividos pelos professores entrevistados formam uma narrativa do passado que se entrelaça no presente, criando significados na forma de sua atividade docente. Dessa forma, o fio condutor de todo o trabalho são as memórias do espaço escolar. Tanto o espaço, quanto o trabalho docente são construções sociais interdependentes, que são influenciadas pela trajetória de vida, cultura e formação do sujeito. As análises são fundamentadas em estudos sobre arquitetura/espaço (Bachelard, Viñao-Frago, Escolano), cultura visual (Hernández), espaço/relações de poder (Foucault), memória (Bergson, Bosi) e estudos sobe docentes (Tardiff, Raymond). O espaço escolar causa impactos objetivos e subjetivos em seus usuários. Nesse sentido, o espaço é produtor de identidades, símbolos e relações que exercem influencia no trabalho docente. Dessa forma, a partir das falas dos professores pode-se perceber que: embora as ideias e práticas pedagógicas sejam ligadas ao contexto político, social e tecnológico, esse último é o mais percebido e, quanto ao uso do espaço, o professor sente-se aprisionado por imposições superiores como decisões da direção escolar, exames nacionais e conteúdo da disciplina.
This thesis is a study of memories of school spaces experienced by teachers. The goal is to identify the conditions and dimensions of teachers work from the scholar environment and memories of the educational journey of seventeen teachers working in higher education and in high school in a city in Minas Gerais. The memories were collected through semi-structured interviews. These memories of a time and space experienced by the teachers interviewed form a narrative of the past that blends with the present time, creating meanings in his teaching activities. Thus, the common thread of this work are the memories of school space. Both the space, as the teaching are interdependent social constructs, which are influenced by the life, culture and formation of the subject. The analyzes are based on studies of architecture / space (Bachelard, Vinao-Frago, Escolano), visual culture (Hernández), space / power relations (Foucault), memory (Bergson, Bosi) and studies about teachers (Tardiff, Raymond). The school environment causes subjective and objective impacts on their inhabitants. In this sense, space produces identity, symbols and relationships which influence on teaching. Thus, from the teachers speeches can be seen that: although the ideas and teaching practices are linked to political, social and technological context, the latter is the most noticed and, for the use of space, the teacher feels trapped by higher decisions of the school board, national tests and content of the discipline.

Brain structures that support spatial cognition by encoding one’s position and direction have been extensively studied for decades. In the majority of studies, neural substrates have been investigated on a horizontal two-dimensional plane, whereas humans and other animals also move vertically in a three-dimensional (3D) world. In this thesis, I investigated how 3D spatial information is represented in the human brain using functional MRI experiments and custom-built 3D virtual environments. In the first experiment, participants moved on flat, tilted-up or tilted-down pathways in a 3D lattice structure. Multivoxel pattern analysis revealed that the anterior hippocampus expressed 3D location information that was similarly sensitive to the vertical and horizontal axes. The retrosplenial cortex and posterior hippocampus represented direction information that was only sensitive to the vertical axis. In the second experiment, participants moved in a virtual building with multiple levels and rooms. Using an fMRI repetition suppression analysis, I observed a hierarchical representation of this 3D space, with anterior hippocampus representing local information within a room, while retrosplenial cortex, parahippocampal cortex and posterior hippocampus represented room information within the wider building. As in the first experiment, vertical and horizontal location information was similarly encoded. In the last experiments, participants were placed into a virtual zero-gravity environment where they could move freely along all 3 axes. The thalamus and subiculum expressed horizontal heading information, whereas retrosplenial cortex showed dominant encoding of vertical heading. Using novel fMRI analyses, I also found preliminary evidence of a 3D grid code in the entorhinal cortex. Overall, these experiments demonstrate the capacity of the human brain to implement a flexible and efficient representation of 3D space. The work in this thesis will, I hope, serve as a stepping-stone in our understanding of how we navigate in the real – 3D – world.

Previous research has shown that in two‐dimensional environments, ‘place cells’ and ‘grid cells’ in the hippocampal formation comprise two units of O’Keefe and Nadel’s (1978) ‘cognitive map’, coding for the rat’s present location and for metric distance information, respectively. Since the world is three‐dimensional, an important question is whether the cognitive map is also volumetric. To explore this issue, place and grid cells were recorded from rats as they shuttled up and down between two ends of a spiral staircase (the ‘helical track’), allowing the same horizontal locations to be sampled at different vertical levels. Using this novel paradigm, it was possible to investigate whether place and grid cell receptive fields are globular or planar. The first experiment demonstrated that place fields extended in the vertical, as well as horizontal, dimension, suggesting that they were globular. However, the vertical extent was larger than the horizontal extent, suggesting either a coarser representation of height, or contextual modulation of fields in the vertical dimension, in the absence of metric vertical distance. Both possibilities imply that the cognitive map is anisotropic (not uniform in all dimensions). The second set of experiments involved probe trials that showed that both distal and local cues influence place fields in the vertical dimension. The third experiment demonstrated that grid cells produced several subfields on the helical track that, similar to the place fields, were vertically elongated. However, they were more elongated than place fields, and showed no vertical periodicity, suggesting the lack of metric vertical information. Overall, these observations suggest that three‐dimensional space is anisotropically represented in the rat brain as a contextually modulated flat map, in which only the current navigation plane is metrically represented. Due to the complexity of a truly three‐dimensional representation, an anisotropic representation is likely in all surface‐navigating animals including humans.

Submitted by Cristian Gutierrez (cristian@ift.unesp.br) on 2016-11-25T17:35:07Z No. of bitstreams: 1 Cristian_Gutierrez_PhD_Thesis.pdf: 2056100 bytes, checksum: 98402a9e05e7c393491419def7ff3ca9 (MD5)
Approved for entry into archive by Felipe Augusto Arakaki (arakaki@reitoria.unesp.br) on 2016-11-30T13:24:29Z (GMT) No. of bitstreams: 1 gutierrezgomez_cl_dr_ift.pdf: 2056100 bytes, checksum: 98402a9e05e7c393491419def7ff3ca9 (MD5)
Made available in DSpace on 2016-11-30T13:24:29Z (GMT). No. of bitstreams: 1 gutierrezgomez_cl_dr_ift.pdf: 2056100 bytes, checksum: 98402a9e05e7c393491419def7ff3ca9 (MD5) Previous issue date: 2016-10-27
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)
O trabalho apresentado nessa tese foi dedicado em explorar soluções de estado ligado para a equação de Bethe-Salpeter, obtidas diretamente no espaço de Minkowski. Para isso, consideramos um procedimento que combina a representação integral de Nakanishi para a amplitude Bethe-Salpeter, desenvolvido por N. Nakanishi na década de sessenta, em conjunto com a projeção da amplitude de Bethe-Salpeter no plano nulo, também conhecida como a projeção na frente de luz. Este método, além de permitir calcular as energias de ligação, que são acessíveis a partir de cálculos bem conhecidos no espaço Euclidiano, permite que se obtenha a amplitude Bethe-Salpeter no espaço de Minkowski e a função de onda de valência na frente de luz. A verificação da validade desse procedimento foi confirmada através de comparação da amplitude de Bethe-Salpeter obtida diretamente no espaço Euclidiano com a amplitude correspondente derivada da equação de Bethe-Salpeter, usando a representação integral de Nakanishi, uma vez a rotação de Wick é realizada. O sucesso dessa abordagem, quando aplicado ao problema do estado ligado de duas partículas escalares trocando uma outra partícula escalar no estado fundamental, assim como o estudo correspondente no limite de energia zero, nos motivou a ampliar a aplicação do procedimento para o estudo de outros problemas de interesse. Em particular, o método foi estendido para o estudo de sistemas com duas dimensões espaciais e uma temporal (2+1), considerando o interesse crescente que surgiu em Física da matéria condensada, onde podemos destacar o caso de elétrons de Dirac no grafeno. Nessa análise preliminar, nos restringimos ao modelo escalar que nos permitiu acessar as principais dificuldades que deverão ser enfrentadas ao estudar o problema do estado ligado entre dois férmions. Dessa forma, este tratamento pode ser considerado como um primeiro passo para a implementação de um método mais realístico em um problema fermiônico. Os cálculos anteriores que consideramos em nossos estudos foram realizados através da aproximação de escada para o kernel de interação irredutível para os estados de onda-s. Portanto, uma das extensões que exploramos nesta tese foi o efeito de se introduzir a contribuição de ordem seguinte no kernel de interação, conhecida como a contribuição de escada-cruzada (cross-ladder). Os efeitos nas energias de ligação e na função de onda na frente de luz é foram analisados de forma detalhada, através dos resultados apresentados. Um estudo particularmente interessante, que foi extensivamente estudado nesta tese, se refere ao problema do espectro da equação Bethe-Salpeter para o estado ligado escalar-escalar. O espectro de estados excitados foi obtido com a abordagem da representação integral Nakanishi, sendo comparado com o obtido no espaço Euclidiano. Além disso, as raçoes excitado/fundamental do espectro relativístico foram reduzidas para às não-relativístico através da escolha de energias de ligação pequenas e considerando a massa do bóson trocado sendo próxima de zero. A função de onda de valência na frente de luz e a função de onda no parâmetro de impacto são apresentadas mostrando as principais características dos estados excitados conhecidos da estrutura não relativística. Na análise do espectro, também são estudadas as amplitudes de momentum-transverso para o estado fundamental e o primeiro estado excitado, que podem ser obtidos, de forma equivalente, no espaço de Minkowski assim como no espaço Euclidiano. Finalmente, focamos o estudo nos fatores de forma eletromagnéticos elásticos na abordagem da Bethe-Salpeter. Consciente de que o cálculo correto dos fatores de forma deve ser feito no espaço de Minkowski, o fator de forma elástico foi calculado levando-se em consideração a aproximação de impulso padrão. Além disso, foi também estudado o efeito da contribuição de ordem superior no fator de forma.
The work presented in this thesis was dedicated in exploring bound-state solutions of the Bethe-Salpeter equation directly in the Minkowski space. For that, we consider a method that combines the Nakanishi integral representation for the Bethe-Salpeter amplitude, developed by Noboru Nakanishi in the sixties, together with the projection of the Bethe-Salpeter amplitude onto the null-plane, also known as the light-front projection. This approach, besides of allowing to compute the binding energies, which are accessible from the usual Euclidean calculation, enables to obtain the Bethe-Salpeter amplitude in the Minkowski space and the light-front wave function. The feasibility of such an approach is further verified by comparing the Bethe-Salpeter amplitude obtained directly in the Euclidean space with the corresponding amplitude obtained by solving the Bethe-Salpeter equation, using the Nakanishi integral representation, once the Wick rotation is performed to this latter. The success of the approach when applied to study the bound state problem of two-scalar particles exchanging another scalar particle in the ground state, as well as the corresponding study at the zero-energy limit, has encouraged us to extend this method to another interesting problems. In particular, we start by extending the method to study problems in (2+1) dimensions due to the increasing interest in the condensed-matter physics, like the study of Dirac electrons in graphene. In this initial examination we restrict to the scalar model, which enables us to access to the main difficulties that we will face when studying the fermion-fermion bound state problem. Hence, this calculation can be considered as the first step towards the implementation of the method to real fermionic problems. The previous calculations have been performed by considering the ladder approximation for the irreducible interacting kernel for s-wave states. Therefore, one of the extensions that is explored in this thesis is the effect of introducing the next contribution in the interacting kernel, known as the scalar-scalar cross-ladder contribution. The effects in the eigenvalues and the light-front wave functions are analyzed in detail, by considering the computed results. A particular interesting subject, extensively studied in this thesis, is concerned to the spectrum of the Bethe-Salpeter equation for the scalar-scalar bound-state problem. The spectrum of excited states obtained with the Nakanishi integral representation approach is compared with that obtained in the Euclidean calculation. Besides, the ratio energies excited/ground of the relativistic spectrum is reduced to the non-relativistic one by choosing small binding energies and the mass of the exchanged boson approaching to zero. The valence light-front wave function and the impact-parameter space valence wave function are displayed, revealing the main features of excited states known from the non-relativistic framework. In the analysis of the spectrum, we also studied the transverse-momentum amplitudes for the ground and first-excited state, which can be equivalently obtained in the Minkowski or Euclidean spaces. Finally, we focus on the study of electromagnetic elastic form factors within the Bethe-Salpeter approach. Aware that the correct calculation of form factors should be performed in the Minkowski space, the calculation of the elastic form factor is carried out with the standard impulse approximation and in addition the effect of the next contribution to the form factor is studied.

Orientador: Lauro Tomio
Coorientador: Tobias Frederico
Banca: Vladimir Karmanov
Banca: Kazuo Tsushima
Banca: Alfredo Takashi Suzuki
Banca: Waynei Leonardo da Silva de Paula
Resumo: O trabalho apresentado nessa tese foi dedicado em explorar soluções de estado ligado para aequação de Bethe-Salpeter, obtidas diretamente no espaço de Minkowski. Para isso, consideramos um procedimento que combina a representação integral de Nakanishi para a amplitude Bethe-Salpeter, desenvolvido por N. Nakanishi na década de sessenta, em conjunto com a projeção da amplitude de Bethe-Salpeter no plano nulo, também conhecida como a projeção na frente de luz. Este método, além de permitir calcular as energias de ligação, que são acessíveis a partir de cálculos bem conhecidos no espaço Euclidiano, permite que se obtenha a amplitude Bethe-Salpeter no espaço de Minkowski e a função de onda de valência na frente de luz. A verificação da validade desse procedimento foi confirmada através de comparação da amplitude de Bethe-Salpeter obtida diretamente no espaço Euclidiano com a amplitude correspondente derivada da equação de Bethe-Salpeter, usando a representação integral de Nakanishi, uma vez a rotação de Wick é realizada. O sucesso dessa abordagem, quando aplicado ao problema do estado ligado de duas partículas escalares trocando uma outra partícula escalar no estado fundamental, assim como o estudo correspondente no limite de energia zero, nos motivou a ampliar a aplicação do procedimento para o estudo de outros problemas de interesse. Em particular, o método foi estendido para o estudo de sistemas com duas dimensões espaciais e uma temporal (2+1), considerando o interesse cresc... (Resumo completo, clicar acesso eletrônico abaixo)
Abstract: The work presented in this thesis was dedicated in exploring bound-state solutions of the Bethe-Salpeter equation directly in the Minkowski space. For that, we consider a method that combines the Nakanishi integral representation for the Bethe-Salpeter amplitude, developed by Noboru Nakanishi in the sixties, together with the projection of the Bethe-Salpeter amplitude onto the null-plane, also known as the light-front projection. This approach, besides of allowing to compute the binding energies, which are accessible from the usual Euclidean calculation, enables to obtain the Bethe-Salpeter amplitude in the Minkowski space and the light-front wave function. The feasibility of such an approach is further verified by comparing the Bethe-Salpeter amplitude obtained directly in the Euclidean space with the corresponding amplitude obtained by solving the Bethe-Salpeter equation, using the Nakanishi integral representation, once the Wick rotation is performed to this latter. The success of the approach when applied to study the bound state problem of two-scalar particles exchanging another scalar particle in the ground state, as well as the corresponding study at the zero-energy limit, has encouraged us to extend this method to another interesting problems. In particular, we start by extending the method to study problems in (2+1) dimensions due to the increasing interest in the condensed-matter physics, like the study of Dirac electrons in graphene. In this initial examination we restrict to the scalar model, which enables us to access to the main difficulties that we will face when studying the fermion-fermion bound state problem. Hence, this calculation can be considered as the first step towards the implementation of the method to real fermionic problems. The previous calculations have been performed by considering the ladder approximation for the... (Complete abstract click electronic access below)
Doutor

Beim Graphisomorphieproblem geht es um die Frage, ob zwei Graphen bis auf Knotenumbenennungen die gleiche Struktur haben. Es ist eines der wenigen verbleibenden natürlichen Probleme, für die weder ein Polynomialzeitalgorithmus noch NP-Härte bekannt ist. Aus dieser Situation ist ein Forschungszweig erwachsen, der effiziente Isomorphiealgorithmen für eingeschränkte Graphklassen entwickelt. Der Hauptbeitrag dieser Arbeit besteht in Logspace-Algorithmen, die das Isomorphieproblem für k-Bäume, Intervallgraphen, sowie Helly- und Proper-Kreisbogengraphen lösen. Dies verbessert zuvor bekannte parallele Algorithmen und führt zu einer vollständigen Klassifikation der Komplexität dieser Probleme, da für sie auch Logspace-Härte nachgewiesen wird. Tatsächlich leisten die vorgestellten Algorithmen mehr: Im Fall der k-Bäume berechnet der Algorithmus kanonische Knotenbenennungen mit O(k log n) Platz. Eine alternative Implementation des Algorithmus kommt mit O((k+1)!n) Zeit aus – hierbei ist n die Anzahl der Knoten – und ist damit der schnellste bekannte FPT-Algorithmus für Isomorphie von k-Bäumen. Die Algorithmen für Intervall- und Kreisbogengraphen berechnen kanonische Repräsentationen – das heißt, sie weisen jedem Knoten ein Intervall (beziehungsweise einen Kreisbogen) zu, sodass diese sich genau dann schneiden, wenn die zugehörigen Knoten benachbart sind, und isomorphe Eingabegraphen das gleiche Intervallmodell (beziehungsweise Kreisbogenmodell) erhalten. Außerdem werden auch Logspace-Algorithmen angegeben, die Intervallrepräsentationen mit zusätzlichen Eigenschaften berechnen – oder erkennen, dass dies nicht möglich ist: Für die resultierenden Intervallmodelle kann gefordert werden, dass sie proper sind (also kein Intervall ein anderes enthält), dass sie unit sind (also alle Intervalle die gleiche Länge haben) oder dass die Längen der paarweisen Schnitte (und optional der einzelnen Intervalle) vorgegebenen Werten entsprechen.
The graph isomorphism problem deals with the question if two graphs have the same structure up to renaming their vertices. It is one of the few remaining natural problems for which neither a polynomial-time algorithm nor NP-hardness is known. This situation has led to a branch of research that develops efficient algorithms for special cases of the graph isomorphism problem, where the input graphs are required to be from restricted graph classes. The main contribution of this thesis comprises of logspace algorithms that solve the isomorphism problem for k-trees, interval graphs, Helly circular-arc graphs and proper circular-arc graphs. This improves previously known parallel algorithms and leads to a complete classification of the complexity of these problems, as they are also shown to be hard for logspace. In fact, these algorithms achieve more: In the case of k-trees, the algorithm computes canonical labelings in space O(k log n). An alternative implementation runs in time O((k+1)!n), where n is the number of vertices, yielding the fastest known FPT algorithm for k-tree isomorphism. The algorithms for interval and circular-arc graphs actually compute canonical representations, i.e., each vertex is assigned an interval (or arc) such that these intersect each other if and only if the corresponding vertices are adjacent, and isomorphic input graphs receive the same interval (or arc) model. This thesis also presents logspace algorithms that compute interval representations with additional properties, or detect that this is not possible: The resulting interval models can be required to be proper (no interval contains another), unit (all intervals have the same length), or to satisfy prescribed lengths for pairwise intersections (and possibly prescribed lengths of intervals).

Studio work The Postgraduate Degree Show is held from 6PthP December 2005 to 17PthP December 2005 and my work is installed in the Sculpture Studio (as a gallery space) at Building 29 of Sydney College of the Arts, Sydney. There are three installations of work, each with a series of paintings and object-models. The media I am using are oil on canvas for the paintings and wood for the models. The titles of my exhibition pieces are Project Studio (Stairs), Project Rented Room (Chair), Project Rented Room (Bed), and Project Object. Together they are entitled Project: Spanning the Space of Dislocation. The project explores the perception of space and its representation through painting and installation. The starting point is the image of familiar architectural objects to which I displace the experience of it from one site (my painting studio) to another (the gallery space) through painting as index. I use the technique of ambiguous linear forms in painting and the reflexive reading of orthogonal projective planes in installation to further extend the viewer’s perception of space and objects. The aim is to show that space has a meaningful relationship to objects and bring about a renewed awareness of habitual practice in seeing and representing space. UResearch paperU I have divided my research paper in two chapters. Chapter one explores the issue of spatial representation through ambiguity of simple linear forms and painting as index. My concern is on space being less important to objects in the distinction between space as ground and object as figure. Within this chapter I argue for an extended and a reflexive mode of seeing and representing space and objects instead of for a ground-figure contrast. By mapping my experience on a usual working site and displacing it to another space, I show that my perception of space is extended such that the boundary between the familiar and foreign (that is, space-object distinction) is blurred. Chapter two explores the method of presentation through painting and installation in a gallery and addresses the viewer’s space of perception with the work. I also discuss possible reflexive readings on the projective planes of the work which further extend the perception of it.

Master of Visual Arts
Studio work The Postgraduate Degree Show is held from 6PthP December 2005 to 17PthP December 2005 and my work is installed in the Sculpture Studio (as a gallery space) at Building 29 of Sydney College of the Arts, Sydney. There are three installations of work, each with a series of paintings and object-models. The media I am using are oil on canvas for the paintings and wood for the models. The titles of my exhibition pieces are Project Studio (Stairs), Project Rented Room (Chair), Project Rented Room (Bed), and Project Object. Together they are entitled Project: Spanning the Space of Dislocation. The project explores the perception of space and its representation through painting and installation. The starting point is the image of familiar architectural objects to which I displace the experience of it from one site (my painting studio) to another (the gallery space) through painting as index. I use the technique of ambiguous linear forms in painting and the reflexive reading of orthogonal projective planes in installation to further extend the viewer’s perception of space and objects. The aim is to show that space has a meaningful relationship to objects and bring about a renewed awareness of habitual practice in seeing and representing space. UResearch paperU I have divided my research paper in two chapters. Chapter one explores the issue of spatial representation through ambiguity of simple linear forms and painting as index. My concern is on space being less important to objects in the distinction between space as ground and object as figure. Within this chapter I argue for an extended and a reflexive mode of seeing and representing space and objects instead of for a ground-figure contrast. By mapping my experience on a usual working site and displacing it to another space, I show that my perception of space is extended such that the boundary between the familiar and foreign (that is, space-object distinction) is blurred. Chapter two explores the method of presentation through painting and installation in a gallery and addresses the viewer’s space of perception with the work. I also discuss possible reflexive readings on the projective planes of the work which further extend the perception of it.

Representation of physical space in cinema is a subject that has been investigated by scholars from different disciplines, including filmmakers. Space in cinema is formed through the overlapping of the three components' of physical space, projected space - the image of the physical space..:.. and mental space, an image or an abstract idea formed in the viewer's mind. Representation of space in cinema is therefore a spatial system structured by three interrelated components of physical; optical and psyc!tological relationships. This study investigates the nature ofinfluential factors contributing to the representation of space in Iranian cinema. To do this, the study constructs a model ofcinematic space system structured by the three components of physical space, screen space and mental space. At the centre of the model there is a shared, overlapping area, whose landscape is a complex spatial form influenced by cultural images. The study argues that this landscape is the site for the development of the viewer's mental image in the context ofIranian films. The study reviews the socio-cultural factors that have influenced the development of cinema in Iran since 1900. It then focuses on analysing the representation of space in Iranian films, identifying dominant influential factors in the landscape of the shared area which have influenced the representation ofspace. The findings of this research confirm that shared cultural images, especially traditional painting and performing arts, Taziyeh, have influenced the cinematic space of Iranian films, although the extent and direction of, influence varies and depends upon the socio-political environment in each period.

Parabolic induction is a method of constructing representations of a reductive group from representations of its parabolic subgroups. We study the parabolic induction for p-adic Banach space representation of p-adic groups. In this dissertation, using the Schneider-Teitelbaum duality, we consider the corresponding Iwasawa modules. We present the dual of parabolically induced representation in terms of the tensor product.

The ability to consistently and accurately decode end target reach goals from a patient's neural activity is necessary for the development of an optimal cognitive neural prosthetic which can assist patients with reach system dysfunction. The use of implantable microelectrode arrays to gather neural activity often leads to a spatial bias in the directional tuning of the sampled neurons due to the small subset of neurons which can be simultaneously captured by the microelectrode array. Cortical neural activity has been shown to be dynamic by several studies using brain-machine interfaces (BMI) decoding neural activity to control an external device. In this study, we investigated the capacity of the sampled neurons to undergo functional reorganization leading to an optimization of the representation of reaching space in neurons encoding end target reach goal as a result of a variable reward schedule associated with the adaptive decoding algorithm used by the BMI. We recorded neural activity using 72 electrodes implanted in the posterior parietal and premotor dorsal areas of a Rhesus Macaque (NHP). The BMI was designed in such a way as to use an adaptive Bayesian decoder to predict the intended reach goal of the NHP, and a successful trial was established when the decoder predicted the instructed reach goal. The amount of reward the primate received was proportional to the ability of the decoder to accurately predict the cued reach goal, with the primate receiving visual feedback of his performance during the task. Our findings established the presence of neurons stable across multiple recording sessions and tracked significant changes in the representation of space in the subset of neurons sampled by the microelectrode arrays, coinciding with an improvement in decode performance. These results suggest that neural changes can be actively induced by the use of reward, leading to a reorganization and optimization of the representation of space in the parietal cortex and an increase in the information available to the decoder.
La possibilité de décoder précisément et fidèlement l'intention de mouvement à partir de l'activité cérébrale d'un patient est nécessaire au développement d'une prothèse cognitive neurale optimale qui puisse subvenir aux besoins de patients atteints d'une dysfonction du système moteur. L'utilisation de matrices de microélectrodes implantables pour échantillonner l'activité neuronale mène souvent à un biais spatial dans la représentation directionnelle des neurones prélevés en raison du petit sous-ensemble de neurones qui peut être simultanément capturés par les microélectrodes. Plusieurs études utilisant des interfaces cerveau-machine (ICM) ont rapporté que l'activité neuronale corticale est dynamique suite au contrôle d'un périphérique externe. Dans cette étude, nous avons étudié la capacité des neurones prélevés à subir une réorganisation fonctionnelle conduisant à une optimisation de la représentation de l'espace dans les neurones codant l'intention de mouvement suite à la remise d'une récompense variable associée à l'algorithme adaptatif de décodage utilisé par le IMC. Nous avons enregistré l'activité neuronale en utilisant 72 électrodes implantées dans le cortex pariétal postérieur et prémoteur dorsale d'un Macaque rhésus (PNH). L'IMC a été conçu de telle manière à utiliser un décodeur adaptatif bayésien pour prédire l'intention de mouvement du PNH, avec un essai réussi établi lorsque le décodeur prédit correctement le mouvement indiqué. La valeur de la récompense reçue par le primate était proportionnelle à la capacité du décodeur à prédire avec précision le mouvement indiqué ainsi qu'un rétrocontrôle visuel de sa performance au cours de la tâche. Nous établissons la présence de neurones stables au travers de plusieurs sessions d'enregistrement et découvrons des changements importants dans la représentation de l'espace dans le sous-ensemble de neurones prélevés par les réseaux de microélectrodes, qui coïncide avec une amélioration des performances de décodage. Ces résultats suggèrent que les changements neuronaux peuvent être activement induits par l'utilisation de la récompense, ce qui conduit à une réorganisation et une optimisation de la représentation de l'espace dans le cortex pariétal et une augmentation de la quantité d'information disponible au décodeur.

The dynamics of a rodent hippocampus during active locomotion is essentially two-dimensional in its multi-dimensional space of states (understanding a reduced description in terms of short-term averaged neuronal activities). Furthermore, it is a two-dimensional model of the animal's motion in an environment. Experimental data show that this well-known hippocampal dynamical constraint (a cognitive map property: O'Keefe & Nadel, 1978) results from intrinsic mechanisms, in particular, involving integration of self-motion (path integration). Several proposals have been made regarding these mechanisms as based on a special architecture of the hippocampus and on an attractor dynamics of some kind. The main problem with an attractor interpretation is that hippocampal spatial codes observed under different behavioral conditions may be statistically independent of each other, and so have to be the underlying two-dimensional attractor structures (called here attractor maps), simultaneously stored in the same network. The present work addresses the above problem and shows that alternative, independent attractor maps can be stored simultaneously in the same network, in a number proportional to the number of neurons. The results enable the design of an attractor-map-based model of the hippocampal formation with a built-in path integration mechanism. A necessary assumption of this model is the pre-existence of a special, "multichart" architecture of the main component, presumably based on CA3. It is shown numerically that the proposed model accounts for most of known observed phenomena in the field. Some of these phenomena become especially clear when the model is approximated by a "macroscopic" version. An alternative plausible explanation, based on another version of the multichart architecture, according to which the hippocampal spatial code originates outside the hippocampus proper, is also examined. Numerical simulations show that both versions of the model are consistent with available experimental data, but can be distinguished by a feasible future experimental test. Finally, it is proposed that multiple attractor maps may constitute a universal tool used by the brain for representation of other internal cognitive models and, in particular, for hippocampal-dependent management of explicit long-term memory. The latter implication has an extreme significance for understanding the hippocampal memory function in humans.

The abilities to adapt and act autonomously in an unstructured and human-oriented environment are necessarily vital for the next generation of robots, which aim to safely cooperate with humans. While this adaptability is natural and feasible for humans, it is still very complex and challenging for robots. Observations and findings from psychology and neuroscience in respect to the development of the human sensorimotor system can inform the development of novel approaches to adaptive robotics. Among these is the formation of the representation of space closely surrounding the body, the Peripersonal Space (PPS) , from multisensory sources like vision, hearing, touch and proprioception, which helps to facilitate human activities within their surroundings. Taking inspiration from the virtual safety margin formed by the PPS representation in humans, this thesis first constructs an equivalent model of the safety zone for each body part of the iCub humanoid robot. This PPS layer serves as a distributed collision predictor, which translates visually detected objects approaching a robot’s body parts (e.g., arm, hand) into the probabilities of a collision between those objects and body parts. This leads to adaptive avoidance behaviors in the robot via an optimization-based reactive controller. Notably, this visual reactive control pipeline can also seamlessly incorporate tactile input to guarantee safety in both pre- and post-collision phases in physical Human-Robot Interaction (pHRI). Concurrently, the controller is also able to take into account multiple targets (of manipulation reaching tasks) generated by a multiple Cartesian point planner. All components, namely the PPS, the multi-target motion planner (for manipulation reaching tasks), the reaching-with-avoidance controller and the humancentred visual perception, are combined harmoniously to form a hybrid control framework designed to provide safety for robots’ interactions in a cluttered environment shared with human partners. Later, motivated by the development of manipulation skills in infants, in which the multisensory integration is thought to play an important role, a learning framework is proposed to allow a robot to learn the processes of forming sensory representations, namely visuomotor and visuotactile, from their own motor activities in the environment. Both multisensory integration models are constructed with Deep Neural Networks (DNNs) in such a way that their outputs are represented in motor space to facilitate the robot’s subsequent actions.

How do we represent numbers and make mathematical calculations? This question is the main focus of the present work and it falls in the domain of mathematical cognition, the field of knowledge concerned with the cognitive and neurological processes that underline mathematical abilities. The mental number line, with its analogue left-to-right orientation of growing numerical values, is often regarded as the best candidate to the role of mental representation of numbers. Many studies have examined the so-called mental number line taking the Spatial-Numerical Association of Response Codes (SNARC) effect as evidence for a unique connection between space and number. However, left-to-right orientation has been shown to extend to other dimensions, including duration and physical size. Such observations converge with the notion of a general magnitude system, where different magnitudes share neural and conceptual resources. This rise an important question about the nature of the information represented along the mental number line: is it exclusive to number or not? The present work is divided in 4 chapters. Chapter 1 addresses several issues in mental representations of numbers. Research in mathematical cognition has a long history and has made considerable progress over the last decades; sometimes this big volume of data makes it difficult to gain a global view of what the state-of-the-art is. For this reason, Chapter 1 will offer an overview of the different mental representations of numbers, whether innate or acquired, precise or approximate, symbolic or non symbolic. On the one hand, the most important insight gained on mental representations of numbers are listed; on the other, literature on the representation of time, space, number and intensity, related with number representation, are revised to show the similarities between these domains, and how those are indicative of common processing mechanisms. The theoretical background specific for the present work is introduced. A great body of evidences point out that the representation and processing of numbers is associated to an activation of spatial codes. One of the classical view of numerical cognition on this subject states that spatial codes are an imprescindible component of the long-term representation of numerical magnitude information. According to this idea, refereed to as number mental line hypothesis, numbers are systematically associated to spatial codes, as if numerical magnitudes were represented along a spatial continuum with small numbers to the left and large numbers to the right. Nevertheless, the origin of the association between numbers and space is not completely clear to date. Studies will be presented showing that the spatial coding of numbers is not stable and is not necessarily the result of long-term memory associations but, on the contrary, a flexible type of representation built during cognitive processing as the result of task demands and spatial coding preferences. Moreover, studies on the association of numbers with others, non-spatial, magnitudes will be reviewed. In Chapter 2 a series of three studies are presented, in all of them a numerosity production method of response was used. Participants performed approximate arithmetic task on symbolically presented numbers, they were instructed to respond by the production of a dot pattern, the set size of which was controlled by a rotating knob. Study 1 shows two experiments in which participants judged the average numerosity between two sequentially presented dot patterns. In Experiment 1, the response was given on a 0–20 numerical scale (categorical scaling), and in Experiment 2, the response was given by the production of a dot pattern of the desired numerosity numerosity production). Data showed that responses were shaped according to an averaging integration model. This suggests the linearity in the response scale of both of the response methods in the approximate arithmetic task. More important, the two operands were found to have the same influence in determining the result. These two experiments served as a validation tool of the numerosity production method of response to be applied in the sequent studies. Study 2 proposes one experiment using the numerosity production method of response to test the influence of the force required to respond on the Operational Momentum (OM) effect. The OM effect is the finding of a systematic tendency to overestimate the results of addition problems and underestimate the results of subtraction problems under conditions that prevent exact calculation. In this experiment the force required to turn the knob has been manipulated in three between subjects blocks. It has been suggested that the OM effect depend on the spatial representation of numbers; by showing that the elimination of psycho-motor feedback nullifies the difference between addition and subtraction, evidence are provided that the OM effect is modulated by information from a magnitude different from space (required force), even when such information is entirely task-irrelevant. Study 3 proposes an experiment on the comparison of four different effects classically considered examples of the automatic spatial organization of numerical information. The spatial-numerical association of response codes (SNARC), that is, the tendency to be faster in responding to small numbers on the left and to bigger number on the right. The distance effect, that is, close numbers are more difficult to compare than numbers far apart. The size congruency effect, that is, numbers are identified more rapidly as bigger or smaller than 5 if their physical size is congruent with the correct answer. And last but not least, the OM effect. Those effects have been tested together to investigate the relationship among them with an inter-individual differences approach. The presence of all the effects object of this study was verified in the participants set. Linear regression have been used to calculate the coefficient of each subject for each effect in order to test the correlation between all the effects this study take into consideration. The result of this study, even if not conclusive, point in the direction of a common representational mechanism underling the tested numerical effects (SNARC, size congruency, distance ). Moreover, the operational bias seem to have a negative correlation only with the SNARC effect, suggesting a connection between the two, but weekending the mental number line account of those effects. In Chapter 3 conclusions are drawn upon the presented experimental work taking into account different explanatory frameworks. The present research work use a relatively unknown method of response to numerical tasks; the numerosity production method of response. This method shows a wide range of applications and opens new scenarios in mathematical cognition, providing a good instrument to understand in detail the implications of action in mathematical cognition. Moreover, the experiments here presented provide clear indications for a role of non-spatial psycho-motor feedback in arithmetical calculations carried out with the numerosity production method of response, thus challenging the classical interpretation of OM as an effect derived from a purely spatial representation of numbers. Moreover, considering that the force information was presented haptically but numerical information visually, such integration across sensory modalities is consistent with the General magnitude system hypothesis suggesting that representations of magnitudes are multimodal. Study 3 comparing, at our knowledge for the first time, different effects connected to the mental number line hypothesis, provide new insight on the shared processing undergoing these classical findings of mathematical cognition. Our findings, although not conclusive, renew the question on the nature of the representation of numbers.
Come possiamo rappresentare i numeri e fare calcoli matematici? Questa domanda è l'obiettivo principale del presente lavoro e cade nel campo della cognizione matematica, il quale si interessa dei processi cognitivi e neurologici che sottendono le abilità matematiche. L'ipotesi della linea numerica mentale (MNL) prevede che i numeri siano rappresentati mentalmente sottoforma di una misura continua (analogica) con valori numerici crescenti da sinistra a destra. La MNL viene considerata uno dei migliori modelli per la rappresentazione mentale dei numeri. Molti studi hanno esaminato la MNL considerando l’effetto SNARC (Spatial-Numerical Association of Response Codes) come prova per una connessione univoca tra spazio e numero. Tuttavia, è stato dimostrato che la rappresentazione mentale di valori piccoli a sinistra e di valori più grandi esiste anche per grandezze diverse dalla numerosità, compresa la durata temporale e la grandezza fisica. Queste osservazioni convergono con l'idea di un sistema dove diverse grandezze (ad esempio tempo, spazio e numerosità) condividono risorse neurali e concettuali, definito sistema generale di elaborazione delle grandezze (GMS). Questo solleva un'importante domanda sulla natura delle informazioni rappresentate lungo la MNL: si tratta esclusivamente di informazioni numeriche? Il presente lavoro è diviso in 4 capitoli. Il capitolo 1 affronta diversi problemi riguardanti la rappresentazione mentale dei numeri. La ricerca nel campo della cognizione matematica ha una lunga storia e ha fatto notevoli progressi negli ultimi decenni; a volte questo grande volume di dati rende difficile ottenere una visione globale di quello che è lo stato dell'arte. Per questo motivo il Capitolo 1 offrirà una 1panoramica dei diversi modelli di rappresentazione mentale dei numeri, sia innati che acquisiti, precisi o approssimati, simbolici o non simbolici. Prima di tutto sono elencate le principali scoperte sulla rappresentazione mentale dei numeri; in secondo luogo verrà presentata una carrellata sulla letteratura che mostra come le rappresentazioni di tempo, spazio, intensità e numero interagiscano tra loro e probabilmente condividano meccanismi di elaborazione; questo fornirà un adeguato contesto teorico necessario alla chiara comprensione dei lavori sperimentali presentati nei capitoli successivi. Una gran quantità di risultati scientifici dimostra che la rappresentazione e l'elaborazione dei numeri siano associate all'attivazione di una rappresentazione di natura spaziale. Una delle posizioni canoniche della cognizione numerica a tal riguardo afferma che la codifica spaziale è una componente imprescindibile della rappresentazione mentale a lungo termine dei numeri. Secondo questa idea, che porta il nome di ipotesi della linea numerica mentale, i numeri sarebbero rappresentati come una linea continua con i numeri più piccoli a sinistra e quelli più grandi a destra. Tuttavia l'origine dell'associazione tra numeri e spazio non è stata ancora totalmente chiarita. Verranno presentati degli studi che dimostrano come la codifica spaziale dei numeri non sia, in effetti, stabile nè necessariamente il risultato di un'associazione a lungo termine, ma al contrario sia una rappresentazione flessibile costruita a partire dalle necessità di elaborazione delle informazioni specifiche per i compiti che ogniuno di noi si trova a svolgere quotidianamente. Inoltre saranno presi in considerazione studi sull'associazione dei numeri con grandezze prive di caratteristiche spaziali. Nel Capitolo 2 viene presentata una serie di tre studi sperimentali ed in ognuno di essi è stato impiegato un metodo di risposta basato sulla produzione di numerosità. I partecipanti hanno eseguito un compito di aritmetica approssimata su numeri presentati, a seconda dello studio, in notazione simbolica o non simbolica. In tutti gli studi presentati i partecipanti sono stati istruiti ad utilizzare un metodo di risposta caratterizzato dalla produzione di numerosità non simboliche, essi infatti fornivano la risposta al compito specifico nel quale erano impegnati attraverso la produzione, sullo schermo di un computer, di un insieme di punti la cui numerosità era controllata dalla rotazione di una manopola posta davanti ai partecipanti e connessa al computer. Un apposito programma si occupava di registrare il grado di rotazione della manopola ed aggiornare il numero di punti presentati sullo schermo. Lo studio 1 presenta due esperimenti in cui i partecipanti giudicavano la numerosità media tra due insiemi di punti presentati in sequenza. Nell'Esperimento 1 di questo studio, i partecipanti utilizzavano una scala di numerica di risposta da 0 a20 (scala categorica), mentre nell'Esperimento 2 la risposta è stata data attraverso il metodo di risposta basato sulla produzione di numerosità. I risultati di questo studio hanno mostrato come le risposte siano state fornite secondo un modello di integrazione Average. Questo suggerisce una linearità nella scala risposta per entrambi i metodi usati nel compito di aritmetica approssimativa. Più importante, i due operandi mostravano di esercitare la stessa influenza sulla risposta fornita dai partecipanti, il che esclude un effetto sequenza o recenza legata ai compiti impiegati. Questi due esperimenti sono serviti come strumento di validazione del metodo di risposta basato sulla produzione di numerosità al fine della sua applicazione negli studi successivi. Lo Studio 2 presenta un esperimento in cui il metodo di risposta basato sulla produzione di numerosità è stato utilizzato per testare l'effetto della forza necessaria a ruotare la manopola usata per portare a termine un compito di aritmetica mentale. In particolare si è verificata l'influenza della variabile Forza sull'effetto denominato Operational Momentum (OM). L'effetto OM è la tendenza sistematica a sovrastimare i risultati di addizione e a sottovalutare i risultati di sottrazioni in condizioni che impediscono un esatto conteggio. In questo esperimento la forza necessaria per ruotare la manopola è stata manipolata in tre blocchi tra i soggetti. La letteratura ha suggerito che l'effetto OM possa dipendere da una rappresentazione spaziale dei numeri; tuttavia i risultati di questo studio dimostrano che l'eliminazione di un feedback psicomotorio quale la forza richiesta per ruotare la manopola, porta all'annullamento della differenza tra addizioni e sottrazioni. I risultati di questo studio forniscono evidenze sperimentali dell'influenza di una grandezza priva di connotazioni spaziali quale la Forza su un fenomeno di aritmetica mentale come l'effetto OM. Questo risultato è particolarmente interessante considerando che la Forza fosse una variabile interamente irrilevante per lo svolgimento del compito. Lo Studio 3 presenta un esperimento sul confronto tra quattro diversi effetti classicamente considerati esempi dell'automaticità dell'attivazione di codici spaziali durante l'elaborazione di informazioni numeriche. Gli effetti che sono stati considerati in questo studio sono l'effetto SNARC, l'effetto distanza, l'effetto di congruenza delle dimensioni e l'effetto OM. L'effetto SNARC: la tendenza ad essere più veloci nel rispondere a numeri piccoli sulla sinistra e a numeri più grandi a destra. L'effetto distanza: il fatto per cui numeri vicini tra loro sono piu difficili da discriminare rispetto a numeri distanti tra loro. L'effetto di congruenza delle dimensioni: il fatto che i numeri sono identificati come maggiori o minori di 5 più rapidamente se la loro dimensione fisica è congruente con la loro grandezza numerica. Ultimo ma non meno importante, l'effetto OM. Tali effetti sono stati testati insieme per indagare i rapporti che li legano con un approccio basato sulle differenze individuali. La presenza di ognuno degli effetti è stata verificata. Al fine di valutare la correlazione tra i vari effetti in esame, è stato calcolato il coefficiente di regressione lineare di ciascun effetto su ognuno dei partecipanti. I risultati di questo studio, anche se non conclusivi, puntano in direzione di una rappresentazione mentale comune tra gli effetti numerici testati (effetto SNARC, effetto di congruenza della dimensione, effetto distanza). L'effetto OM, inoltre, sembra correlare negativamente con l'effetto SNARC, suggerendo una connessione tra i due, ma contraddicendo la teoria della linea numerica mentale. Nel capitolo 3 si traggono conclusioni sul lavoro sperimentale presentato tenendo conto di diversi quadri esplicativi. Il presente lavoro di ricerca utilizza un metodo di risposta per compiti numerici relativamente poco noto: il metodo di risposta basato sulla produzione di numerosità. Questo metodo presenta una vasta gamma di applicazioni e apre nuovi scenari nel campo della cognizione matematica, fornendo un valido strumento per comprendere nel dettaglio le implicazioni dell'azione nella cognizione matematica. Gli esperimenti qui presentati, inoltre, forniscono indicazioni chiare rispetto al ruolo del feedback psicomotorio con caratteristiche non spaziali in compiti di aritmetica mentale portati a termine attraverso un metodi di risposta basato sulla produzione di numerosità, mettendo così in discussione l'interpretazione classica dell'effetto OM come effetto derivato da una rappresentazione puramente spaziale dei numeri. Considerando che le informazioni riguardanti la forza sono state presentate attraverso un feedback tattile mentre le informazioni numeriche sono state presentate visivamente, tale integrazione tra modalità sensoriali diverse è coerente con l'ipotesi di un sistema generale per le grandezze. Lo studio 3 confrontando, a nostra conoscenza per la prima volta, diversi effetti legati all'ipotesi della linea numerica mentale, fornisce nuove informazioni sui meccanismi di elaborazioni condivisi a questi classici effetti nel campo della cognizione matematica. I nostri risultati, anche se non conclusivi, rinnovano la domanda sulla natura della rappresentazione mentale dei numeri.

Despite advances in image, video, and motion capture technologies, human interactions are frequently represented as line drawings. Intuitively, drawings provide a useful way of filtering complex, dynamic sequences to produce concise representations of interaction. They also make it possible to represent phenomena such as topic spaces, that do not have a concrete physical manifestation. However, the processes involved in producing these drawings, the advantages and limitations of line drawings as representations, and the implications of drawing as an analytic method have not previously been investigated. This thesis explores the use of drawings to represent human interaction and is informed by the prior experience and abilities of the investigator as a practising visual artist. It begins by discussing the drawing process and how it has been used to capture human activities. Key drawing techniques are identified and tested against an excerpt from an interaction between architects. A series of new drawings are constructed to depict one scene from this interaction, highlighting the contrasts between each drawing technique and their impact on the way shared spaces are represented. A second series of original drawings are produced exploring new ways of representing these spaces, leading to a proposal for a field-based approach that combines gesture paths, fields, and human figures to create a richer analytic representation. A protocol for using this approach to analyse video in practice is developed and evaluated though a sequence of three participatory workshops for researchers in human interaction. The results suggest that the field based process of drawing facilitates the production of spatially enriched graphical representations of qualitative spaces. The thesis concludes that the use of drawing to explore non-metric approaches to shared interactional space, has implications for research in human interaction, interaction design, clinical psychology, anthropology, and discourse analysis, and will find form in new new approaches to contemporary artistic practice.

Machine learning methods have been successfully used to analyse neuroimaging data for a variety of applications, including the classification of subjects with different brain disorders. However, most studies still rely on the labelling of the subjects, constraining the study of several brain diseases within a paradigm of pre-defined clinical labels, which have shown to be unreliable in some cases. The lack of understanding regarding the association between brain and behaviour presents itself as an interesting challenge for more exploratory machine learning approaches, which could potentially help in the study of diseases whose clinical labels have shown limitations. The aim of this project is to explore the possibility of using eigen-decomposition approaches to find multivariate associative effects between brain structure and behaviour in an exploratory way. This thesis addresses a number of issues associated with eigen-decomposition methods, in order to enable their application to investigate brain/behaviour relationships in a reliable way. The first contribution was showing the advantages of an alternative matrix deflation approach to be used with Sparse Partial Least Squares (SPLS). The modified SPLS method was later used to model the associations between clinical/demographic data and brain structure, without relying on a priori assumptions on the sparsity of each data source. A novel multiple hold-out SPLS framework was then proposed, which allowed for the detection of robust multivariate associative effects between brain structure and individual questionnaire items. The linearity assumption of most machine learning methods used in neuroimaging might be a limitation, since these methods will not have enough flexibility to detect non-linear associations. In order to address this issue, a novel Sparse Canonical Correlation Analysis (SCCA) method was proposed, which allows one to use sparsity constraints in one data source (e.g. neuroimaging data), with non-linear transformations of the data in the other source (e.g. clinical data).

Spacecraft designs are a result of system properties and design variables that ensure the spacecraft will operate to mission objectives. The focus of this effort is a set of global system variables for frequency, length, total mass and the ratio between the payload mass and the support structure mass. These properties will be explored to observe the behavior of the system and develop relationships that govern the trade-offs between the variables and assist mission planners in future spacecraft design. These variables will be observed in one-dimensional structures where the dominating dimension is many times larger than the other two dimensions and the system is comprised of a support and a payload member. To observe the interaction between the payload and the support, the system was varied for different system variables and observed through ABAQUS finite element software. Attempts were made to predict the system frequency through mathematical approaches. The finite element work was able to generate several approximate relationships between the system variables and the fundamental natural frequency of the system. From these relationships an approximate equation was developed for the frequency for a fixed mass ratio and load ratio as a function of the length, bending stiffness, and total mass of the system. Additional work into the changes to the system as the number of connect points is increased shows the system converging towards a frequency solution which results in a minimized dependence on the connection points. These results were then compared to those of several derived analytical models to determine if a closed-form solution could be used to predict system behavior over the same range of structural characteristics. This closed form solution proved to correlate well to analytical predictions only for the case where the support structure dominates the total system mass, and thus the structural system performs like a beam under compression. Further work is necessary to accurately predict the system frequency through an analytical approach. These insights promise to aid mission designers in objectively evaluating new structural architectures based on structural performance rather than on an unbalanced adherence to heritage or in some cases personal preference.

The act of making challenges ideas through fabrication and the laws of reality that are part of becoming. This research explores the making of physical models as a design process where that act of making 'models for'1 design intention is itself a rich field of speculation. These models for design intention are different to the models of design intention as they are less a finished and singular object, and more an instrument for thinking. The aim of this research is to explore the qualities of models for design intention through an engagement with the landscape in order to understand making as a transformative and emergent process of space, time, material, technique, and the role of the observer. Making for design, the model as idea, seeks to both test and provide opportunities for the convergence of forces and relationships to be created and emergent. Fundamental to this notion is an understanding that the act of making is itself a continual re-making process. The reciprocity invoked by this action engages a rich field of criteria which are potent because of their schizophrenic nature. This paper will discuss my research through a number of projects and esquisses that have been explored during the course of this research which demonstrate the development of my position of making as a continual re-making of space.

Advances in surface performance capture have enabled the reconstruction of real world scenes such as people and animals with a realism hard to achieve by an animator. The work presented in this thesis aims to develop techniques for interactive editing and manipulation of captured mesh sequences with the flexibility associated with conventional computer animation techniques. In particular, the application of Laplacian deformation for animation and compression of surface motion capture data is investigated. Laplacian deformation enables the manipulation of a mesh at a vertex level while maintaining its local geometric properties but lacks a mechanism for ensuring the preservation of its underlying physical structure. Motivated by this limitation, a learnt surface deformation basis constructed in the space of differential cqordinates is introduced. The incorporation of this basis into the Laplacian framework constrains the solution to the space of plausible deformations built from a set of examples, therefore preserving the structure of the mesh. The successful application of this approach to space-time editing together with a set of novel non-linear edit propagation techniques are presented. Representations for efficient storage of surface motion capture sequences, generally comprised of hundreds of frames with thousands of vertices, are investigated. A novel layered representation that exploits the articulated nature of the data is presented and compared with other compression techniques based on PCA and Laplacian deformation, with and without using the aforementioned surface deformation basis. The proposed layered representation achieves consistently high compression ratio with low maximum reconstruction errors in three test sequences from different characters.

This thesis presents results on the description of plasma waves in terms of wavepackets. The wave field is decomposed into a distribution of wavepackets in a space of position, wavevector, time, and frequency. A complex structure joining each pair of Fourier conjugate variables into a single complex coordinate allows the efficient derivation of equations of motion for the phase space distribution by exploiting its analytic properties. The Wick symbol calculus, a mathematical tool generalizing many convenient properties of the Fourier transform to a local setting, is used to derive new exact phase space equations which maintain full information on the phase of the waves and include effects nonlocal in phase space such as harmonic generation. A general purpose asymptotic expansion of the Wick symbol product formula is used to treat dispersion, refraction, photon acceleration, and ponderomotive forces. Examples studied include the nonlinear Schrödinger equation, mode conversion, and the Vlasov equation. The structure of partially coherent wave fields is understood in terms of zeros in the phase space distribution caused by dislocations in its complex phase which are shown to be correlated with the field entropy. Simulations of plasma heating by crossing electron beams are understood by representing the resulting plasma waves in phase space. The local coherence properties of the beam driven Langmuir waves are studied numerically.

This report covers a master’s thesis project done at the University Of Utah for the OpenSpace project. OpenSpace is a open-source astronomy visualization software and the thesis focus was to visualize the ever-increasing number of man-made space debris. Two different visualization methods have been used in this thesis. One was a volume rendering and it was evaluated how it works in relation to an orbital trail representation, which was the other method. If the volumetric representation would reduce cluttering, is one of the aspects that will be evaluated, as well as a more open ended exploratory question which is if the volumetric representation can provide any new insights about the data. In short, will a volumetric representation give anything that an orbital representation cannot? A volume rendering can use different types of grids. The thesis evaluates the pros and cons of a cartesian- and spherical grid, as well as the different resolution of the grid and tweaks in the transfer function. An orbital trail representation was previously implemented in OpenSpace (which will be called the individual scene graph node implementation in this report) that had its pros. One con, however, was that it did not scale very well with increasing number of data elements. Visualizing all the data sets containing each trackable piece of space debris simultaneously using this implementation causes the software to slow down significantly. An alternative implementation (which will be called single draw call implementation in this report) was therefore tested in hopes to solve this issue. To see the performance difference, tests were performed where frame time for the whole scene was measured.

My thesis interrogates the complex and indeterminate nature of Taiwanese identity as it is articulated in post- I 980s Taiwanese theatre productions. I argue that Taiwanese identity is negotiated in a 'diasporic space' that has manifestations through cultural hybridity, spatio-temporal disruption and homing in travelling. Initially, I establish the conceptual framework of diasporic space through critical investigations of the sociality of modem diaspora, post-dolonial notions of cultural difference and hybridity (Homi Bhabha) and space-time dynamics as elaborated in Foucault's conception of heterotopias. The subsequent chapters consist of performance analyses and provide dramatic illustrations of these theories as they are imbricated in diasporic space. Subsequently, I examine the appropriation of Beijing Opera aesthetics in a Taiwanese context, and argue that it engenders a hybrid identity by defying the totalising force of Chineseness. I also consider how national space and its attendant essentialist identity is attempted via a sacralised home of homogeneous constitution, thus arguing for the impossibility of identifying a stable national cultural identity due to infracultural differences in the diasporic community of. Taiwan. To fully account for the lived experience of the Taiwanese, I then explore the dialectic force of history that shapes the cultural imaginary of home and identity in ten theatrical productions. I argue that, rather than being bound to a fixed home/land, Taiwanese identity is mediated in the spatio-temporal difference between the homes in the past in China and the present in Taiwan. In addition, I examine the internal conflicts in present-day Taiwan that are unfolded through stories depicting everyday life. The Taiwanese constantly travel in and out of the present locality, and each journey in its own particularity touches upon broader cultural politics of locating home identity. Probing the various manners in which these chosen performances locate Taiwanese identity, I evaluate their achievement in presenting a multiplicity of theatrical possibilities and alternative perspectives of cultural reality that helps envision a 'new' 'diasporic' understanding of homing through travelling, inhabiting shifting moments and movements when/where identity is always being re-configured.

In this thesis, I will explore the triangular relationship between space, representational practices, and the colonial present. I will grapple with a few key research questions: how do we, as Westerners, represent the “other”? How, in turn, do we represent “ourselves”? How have these representational practices shaped the conduct of the War on Terror? And finally, how are (neo)colonial struggles over the politics of representation intricately bound up with questions of geography? By focusing my attention upon the recent invasion, and subsequent occupation of Iraq, I hope to offer a historico-geographically responsible, as well as anti-essentialist, reading of three distinct “digital spaces”: two blogs (Riverbend’s Baghdad Burning and Colby Buzzell’s My War: Killing Time in Iraq) and Multi-National Force Iraq’s YouTube channel. Here, I will argue that, to paraphrase Edward Said, broader geographical struggles, over forms, over images, and over imaginings are not only being dispersed around the globe, they are also being fractured and subsequently contested on a more micro-scale in these new digital battlegrounds. As I hope to demonstrate over the course of this thesis, social media websites such as blogs and YouTube must conceptualized not only as political, but also as antipolitical spaces, in that they both encourage and stifle critical debate on issues pertaining to late modern warfare. Furthermore, the discursive dimensions of geographical struggle must be brought into (vexed) relation with its material dimensions (i.e., armies moving across space), and it is the mutually constitutive nature of this relationship that I will emphasize in this thesis: in other words, the conduct of late modern warfare is not only influenced by, but also influences, the deployment of representational practices. Ultimately, I argue that the increasing importance of the so-called “social media” (i.e., blogs, YouTube, Facebook, Twitter, etc.) as digital spaces of (anti?)politics enables us, as critical human geographers, to produce a genuinely human geography, and to think about space, the body, and representational practices in very different ways.

This project examines the conception and imaging of the city in the late sixteenth and early seventeenth centuries. The thesis aims to chart the ways in which a spatialised reading of the metropolis most fully realised in ceremonial representations of the city informs representational strategies of the time. Chapter 1 looks at the transformations taking place during this period in the practice of land surveying, exploring the implications of the new techniques of geometrical survey for conceptions of civic space. Examining the parallels between the viewing of the estate and the reformation of the Rogationtide ceremonies of perambulating the bounds, the urban context for spatial description is analysed through a reading of John Stow's Survey of London. In Chapter 2 the resistance of the city to a strictly geometrical conception of space is traced through an analysis of early printed maps of the city and the texts of civic ceremonies. The shared interest of these cultural practices in the representation of civic space is interrogated to reveal an understanding of the city as comprising !oth built environment and social body which informs the deployment of the city as a subject of cartographic representation. The next chapter analyses the costume book in the context of a Europewide project of geographical description. The production of a clothed body capable of articulating spatial and hierarchical difference is examined in relation to the available ceremonial models for the negotiation of these intersecting axes of description and the tensions generated by this representational strategy The final chapter undertakes a reinvestigation of the Earl of Essex's rebellion, reading a wide range of materials to argue for the centrality of anxieties over the control of the civic sign to the understanding of this event.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Architecture, 2000.
Includes bibliographical references (leaves 52-54).
Although the notion of "space" in architecture is a relatively contemporary one, this research looks at the difference between the conception and representation of space and the actual material reality. With contemporary thought brought about by the modern measure, as architects formalize their ideas in representations, this paper argues that there arises a tendency to quantify and objectify the represented space and discount the experiential nature of the space. This research was initiated in reaction to this tendency to conceive of space as a given, formal static container in search of a wider notion of space as a product of interactions between various dynamics. Using small time based representational design experiments as well as specific precedents of conceptions and representations of space as running parallel points of reference; this investigation explores the element of time as one of the possible components of the various dynamics that produce space. Specifically, a non-chronological look at the modern, contemporary and pre-modern notion of time was taken to explore possible alternative conceptions and representations of space and time, contending that space is neither static or exclusive of time, nor is it a stage set for speed. In other words, this paper concludes that space and time are first and foremost products of experience.
by Hiroshi Okamoto.
S.M.