Rozprawy doktorskie na temat „Human model”

Theories of Extended Mind have evolved in waves to reach the present state of disagreement with regard to whether or not external artefacts become part of the mind when used for memory purposes. A four-step approach has been used to address and resolve this disagreement. First, a new component for models of mind which provides a saliency function is provided. This saliency function corresponds to computational elements found necessary in large mainframe computer systems for handling rich data environments. Second, there is introduced a Computational Model of Memory containing the new component which models the operation of human memory. The Computational Model of Memory contains four interoperative elements including the new component, short-term memory, long-term memory, and a cross-reference associator. Third, the work of Marcin Milkowski is drawn upon to obtain a general method of assessing a computational model’s well-formedness, and the method is applied to prove the adequacy of the Computational Model of Memory. According to Milkowski’s schema, the model satisfies most criteria for a well-formed computational model, including in particular a separation between conceptual elements of the model, and constitutive elements of the model, which while explicitly related, are required to subsist at separate logical conceptual levels. Fourth, the Computational Model of Memory is applied to outstanding arguments in Extended Mind to clarify and resolve several of these arguments. The model serves to highlight where the nature of the disagreement depends upon a category error of reference, and further resolves a key disagreement by demonstrating that the mind may treat external artefacts as an alternative realizable constitutive element of short-term and long-term memory.

Thesis (MBA)--Stellenbosch University, 2004.
ENGLISH ABSTRACT: As the world's economic landscape undergoes a fundamental shift from industrial economy in which plant and equipment are the core assets, to the 'new' economy which places a high premium on people and intangible assets traditional accounting systems are becoming less effective. Intellectual Capital has become the indispensable component of corporate value. The significant rise in the market-to-book ratio of listed companies is testimonial of this fact. By focusing on physical and cash assets, and remaining oblivious to Intellectual Capital, conventional accounting methods are missing a very crucial point. The exclusion of Intellectual Capital from financial performance reports results in information deficiency for both internal and external stakeholders of organizations. Measurement and reporting of Intellectual Capital has thus become imperative. However, it is the Human Capital component (of Intellectual Capital) that should be the prime concern of business leaders and other stakeholders. People are the true agents in business; all the other assets, whether tangible or intangible, are the result of human actions and ultimately depend on people for their continued existence. Measurement and reporting of Human Capital is therefore of the essence. Measurement of Human Capital is not simple and straightforward. Development of methodologies for valuation of Human Capital is a daunting challenge. In spite of its difficulty, measurement of Human Capital has to be vigorously pursued; the stakes are just too high for the challenge to be shunned. This study proposes a system for valuation of Human Capital. "Valuation" may conjure expectations of financial measurement; however, despite concerted efforts by the accounting profession, currency-based valuation of people has received very little, if any, appreciation in industry. The model put forward here, is a non-monetary Human Capital Index.
AFRIKAANSE OPSOMMING: Met die fundamentele verskuiwing van die ekonomiese landskap van die wêreld, vanaf 'n industriële ekonomie met produksie-aanlegte en toerusting as primêre bates, tot die nuwe ekonomie wat 'n hoë premie op mense en ontasbare bates plaas, het konvensionele rekeningkundige stelsels toenemend ondoeltreffend geraak. Intellektuele kapitaal het 'n onontbeerlike onderdeel van korporatiewe waarde geword. Die betekenisvolle premie wat die markwaarde bo die batewaarde van genoteerde maatskappye geniet, lewer bewys van hierdie tendens. Deur te fokus op fisiese en monetêre bates, en nie intellektuele bates in ag te neem nie , verontagsaam konvensionele rekeningkundige stelsels 'n kern beginsel. Die uitsluiting van intellektuele kapitaal as deel van prestasie verslagdoening lei tot 'n gebrekkige inligtingsbasis vir beide interne en eksterne belangegroepe van die organisasie. Meting van, en verslagdoening oor intellektuele kapitaal, het dus 'n noodsaaklikheid geword. Dit is egter die menslike hulpbron komponent van intellektuele kapitaal wat die primêre oorweging by sakeleiers en ander belanghebbendes behoort te wees. Mense is die werklike rolspelers in organisasies. AI die ander bates, tasbaar of ontasbaar, is die gevolg van menslike aktiwiteit, en hang uiteindelik van mense af vir hul voortgesette bestaan. Daarom is dit van die uiterste belang dat daar 'n proses is wat menslike bates evalueer en verslag doen. Die meting van menslike kapitaal is nie eenvoudig en voor die hand liggend nie. Die ontwikkeling van metodes om menslike kapitaal te assesseer is 'n besondere uitdaging. Ten spyte van die probleme moet die assessering van menslike kapitaal daadwerklik nagestreef word; hierdie saak is te belangrik om te ontwyk. Hierdie studie stel 'n model voor om waardebepaling van menslike kapitaal te doen. So 'n waardebepaling mag verwagtinge van 'n finansiële metingsbasis skep; tog, ten spyte van doelgerigte pogings deur die rekeningkundige professie, het 'n monetêre waardebepaling van mense weinig, indien enige, aanvaarding in die sakewêreld ontvang. Die model wat hier voorgestel word, is 'n nie-monetêre menslike kapitaal indeks.

Thesis (Ph. D.)--Hong Kong University of Science and Technology, 2003.
Includes bibliographical references (leaves 124-134). Also available in electronic version. Access restricted to campus users.

Thesis: S.M. in Technology and Policy, Massachusetts Institute of Technology, School of Engineering, Institute for Data, Systems, and Society, Technology and Policy Program, 2017.
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (pages 101-107).
This thesis presents an investigation of human-model interaction in relation to model-centric decision-making. Models are abstractions, or simplifications, of reality that humans use to augment their ability to make sense of the world, anticipate future outcomes, and make decisions. This thesis focuses on models that aid decision-making in the design and operation of technological systems. Model-centric engineering is transforming traditional engineering towards a paradigm of comprehensive, integrated model use throughout the lifecycle of complex systems. This model-centric shift aims to increase the efficiency and efficacy of system decision-making. Without appropriately considering and designing for the human element, however, model-centric engineering will fail to achieve its desired results. Enabling effective human-model interaction, therefore, is crucial for realizing the value that models and model-centric engineering practice can provide. Advances in model technology and computational resources have been steadily made, however, the many facets of the human-model interaction experience remain relatively unexplored. Through empirical and qualitative methods, this thesis presents an exploration of human-model interaction in an effort to identify decision-making challenges, and appropriate mitigations, for individuals in model-centric environments. Learning from existing literature and past situations with similar considerations is a useful place to start in investigating the human aspects. Two analogy case studies reveal relevant individual and organizational challenges that may affect human-model interaction and decision-making within model-centric environments. An expert interview-based study yields empirical insight from thirty experts into sociotechnical factors that influence the trust and use of models by various types of actors within the model-centric decision-making process. Additionally, as automation, autonomy, and artificial intelligence (AI) will likely play key roles in successful model-centric engineering, relevant literature-based considerations are presented for how the capabilities of AI and autonomy may relate to a model-centric context. This cumulative research is ultimately distilled into twenty-nine descriptive and prescriptive heuristics for enabling effective human-model interaction and model-centric decision-making. These heuristics emerged from the voice of the experts interviewed, as well as from case studies and literature analyzed. Policy considerations based on this investigation are discussed, along with a suggested strategy of planned adaption for model-centric policymaking. Overall, this research aims to generate grounded theory to motivate and guide future research and development for enabling effective human-model interaction and model-centric decision-making.
by Erling Shane German.
S.M. in Technology and Policy

La propuesta de una vejiga artificial es un obstáculo a trasponer. El cáncer de vejiga está entre los casos más frecuentes de enfermedades oncológicas en Estados Unidos y Europa. Ese cáncer es considerado un problema médico importante una vez que esa enfermedad presenta altas tasas de re-ocurrencia, muchas veces llevando a la remoción del órgano. La solución más sofisticada para remplazar ese órgano es la vejiga ileal, que consiste en una neovejiga hecha de tejido intestinal del enfermo. Desafortunadamente, esa solución presenta no solo problemas mecánicos funcionales, descritos en la literatura como problemas de vaciado y fuga, peo también problemas de orden biológica (como ejemplo pérdida ósea, debido a la absorción por el intestino de substancias que necesitan ser eliminadas del organismo). A través de la solicitación de la comunidad urológica del Hospital Clínico de Barcelona y con su experiencia en modelos numéricos para estructuras biomédicas, el Centro de Métodos Numéricos en Ingeniería (CIMNE) ha tenido la iniciativa de proporcionar actividad investigadora de la mecánica de la vejiga urinaria y de la simulación de interacción fluidoestructura para reproducir el llenado y vaciado de ese órgano con la orina. La simulación de la vejiga humana por el Método de los Elementos Finitos (FEM) y un completo entendimiento de la mecánica de ese órgano y de su interacción con la orina dará la posibilidad de proponer mejora en la geometría y de analizar materiales para la solución artificial en caso de remplazamiento de la vejiga. Para lograr ese objetivo, primeramente procedemos a una revisión bibliográfica de los modelos matemáticos del aparato urinario y un estudio comprehensivo de la fisiología y dinámica de la vejiga. Presentamos una revisión de las principales estructuras urológicas, riñón, uréter y uretra. Las estructuras anexas también son consideradas para entender las condiciones de contorno del problema estudiado. Posteriormente, proponemos el modelo constitutivo para estudiar la vejiga urinaria humana. El comportamiento del musculo detrusor durante llenado y vaciado de la vejiga con orina, su habilidad de retención de orina a baja presión debe ser correctamente representada por medio de la implementación de un modelo constitutivo no-lineal. El modelo matemático necesita representar las variables mecánicas que gobiernan ese órgano, y también las propiedades de la orina. El comportamiento no-lineal de tejidos vivos es implementado y validado con ejemplos de la literatura. La propiedad quasi-incompressible de la orina y las ecuaciones Navier-Stokes son consideradas para análisis del fluido. Para representar la geometría de la vejiga, implementamos un modelo computacional 3D a partir de imágenes de tomografía computadorizada de un cadáver adulto. Los datos son tratados para considerar las condiciones de contorno. Hemos construido dos modelos de malla: un mallado con tetrahedos de cuatro nodos y otro mallado con elementos de membrana de tres nodos. El esquema utilizado para calcular la interacción fluido-estructura debe ser adecuado para materiales de densidad muy parecidas. El análisis numérico de llenado y vaciado de la vejiga humana es validada con tests urodinámicos estandarizados. La parte final de la tesis, presentamos una simulación de una neo-vejiga, siendo el primer paso para representar numéricamente materiales artificiales para remplazamiento de la vejiga.
The proposal of an artificial bladder is still a challenge to overcome. Bladder cancer is among the most frequent cases of oncologic diseases in United States and Europe. It is considered a major medical problem once this disease has high rates of reoccurrence, often leading to the extirpation of this organ. The most refined solution to replace this organ is the ileal bladder, which consists of a neobladder made of the patient’s intestinal tissue. Unfortunately this solution presents not only functional mechanical problems, described on the literature as voiding and leaking problems, but also biological ones (i.e. bone loss, given the absorption by the intestine of substances that should be eliminated from the organism). Urged by the urological community of the Hospital Clinic de Barcelona and backgrounded by its experience in the numerical simulation of biomedical structures, the Center of Numerical Methods in Engineering (CIMNE) had the initiative to provide the research of the mechanics of the urinary bladder and the simulation of fluid structure interaction (FSI) to account for the filling and voiding of this organ with urine. The Finite Element Method (FEM) simulation of the real bladder and the comprehensive understanding of the mechanics of this organ and its interaction with urine will give the possibility to propose geometrical improvements and study suitable materials for an artificial solution to address the cases on which the bladder needs to be removed. To reach this goal, first we proceeded to the bibliographic review of mathematical models of the urinary apparatus and to a comprehensive study of the physiology and dynamics of the bladder. A review of the major urological structures, kidney, ureter and urethra, takes place. To consider boundary conditions other surrounding structures to the urinary system are also studied. In the second part of the thesis, we propose the numerical model to study the human urinary bladder. The behavior of the detrusor muscle during filling and voiding of the bladder with urine and its ability to promote the storage of urine under low pressure need to be accurately represented, requiring the implementation of a non-linear constitutive model. The mathematical model needs to be capable to simulate the mechanical variables that govern this organ and the properties of the urine. The nonlinear behavior of living tissues is implemented and validated with examples from the literature. The quasi-incompressibility property of urine and the navierstokes equations for the fluid are taken into account. The geometry of the bladder needs to be taken into account, and the implementation of a 3D computational model obtained from the computerized tomography of a cadaver male adult is considered. The data has been treated to consider boundary conditions. Two models have been conceived: one meshed with four nodes tetrahedral and another meshed with shell elements. FSI must work for the simulation of filling and voiding of the bladder. Due to the close densities of the materials the scheme used to solve fluid-structure needs to be carefully selected. The proposed numerical model and the filling and voiding analysis are finally validated with standardized urodynamic tests. The final part of the thesis, the simulation of a neobladder is presented, being the first step to simulate numerically artificial materials for bladder replacement.

La comprensio automatica de les accions humanes observades en sequencies d'imatges es molt important en el area de recerca de la Vision per Computador, amb aplicacions a gran escala en la vigilancia de video, analisi del moviment huma, interficies de realitat virtual, robots de navegacio, aixi com per al reconeixement, indexacio, i recuperacio de video. Aquesta tesi presenta una serie de tecniques per resoldre el problema del reconeixement de les accions humanes en video. El nostre primer enfocament cap a aquesta tema es basa en la optimitzacio d'un model probabilistic de les parts del cos utilitzant una Hidden Markov Model (HMM). Aquest enfocament es basa en un strong model, capac de distingir entre accions similars considerant nomes les parts del cos que tenen les majors aportacions en la realitzacio de certes accions, per exemple en cames per caminar i correr, o en bracos per a accions com boxa i aplaudir. El nostre seguent enfocament es basa en l'observacio de que el reconeixement d'accions es pot realitzar usant nomes informacio visual, ii la postura humana desenvolupada durant una accio, analitzant la la informacio d'uns quants frames en lloc d'examinar la sequencia completa. En aquest metode, les accions es representen mitjançant un model Bag-of-key-poses per poder capturar la variacio de la postura humana durant el desenvolupament d'una accio. Per fer front al problema del reconeixement de l'accio en escenes complexes, tot seguit es proposa una aproximacio model free basada en l'analisi de punts d'interes espai-temporals (STIPs) que disposen de molta informacio local. Amb aquesta finalitat, s'ha desenvolupat un nou detector de STIPs que es basa en el mecanisme de inhibicio del camp receptiu utilitzat en l'escorça primaria, en particular en l'orientacio selectiva visual de les neurones. A mes, hem estes el nostre reconeixement d'accions basat en STIPs selectius a sistemes multi-camera. En aquest cas, els STIPs selectius de cada punt de vista es combinen mitjançant les dades 3D reconstruïts per formar STIPs selectius 4D (espai 3D + temps). A la part final d'aquesta tesi, ens dediquem al reconeixement continu d'esdeveniments visuals (CVER) en bases de dades de videos de seguretat enormes, amb un gran conjunt de dades. Aquest problema es extremadament difícil a causa de l'alta escalabilitat de les dades, a les dificultats de l'entorn real en que es aplcia ja una variabilitat en escena molt ampli. Per abordar aquests problemes, les regions en moviment son detectades a partir d'una tecnica anomenada max margin generalized Hough transformation, que s'utilitza per aprendre aquella distribucio de característiques voltant d'una accio per reconeixer hipotesis que despres es verifiquen per Bag-of-words mes un classificador lineal. Hem validat les nostres tecniques en diversos conjunts de dades de vídeo vigilància que constitueixen l'estat de l'art actual en aquest tema. Els resultats obtinguts demostren que hem millorat la precisio en la deteccio d'accions humanes en video.
La comprension automatica de las acciones humanas observadas en secuencias de imagenes es muy importante en el area de investigacion de la Vision por Computador, con aplicaciones a gran escala en la vigilancia de video, analisis del movimiento humano, interfaces de realidad virtual, robots de navegacion, asi como para el reconocimiento, indexacion, y recuperacion de video. Esta tesis presenta una serie de tecnicas para resolver el problema del reconocimiento de las acciones humanas en video. Nuestro primer enfoque hacia esta tema se basa en la optimizacion de un modelo probabilístico de las partes del cuerpo utilizando una Hidden Markov Model (HMM). Este enfoque se basa en un strong model, capaz de distinguir entre acciones similares considerando solo las partes del cuerpo que tienen las mayores aportaciones en la realizacion de ciertas acciones, por ejemplo en piernas para caminar y correr, o en brazos para acciones como boxeo y aplaudir. Nuestro siguiente enfoque se basa en la observacion de que el reconocimiento de acciones se puede realizar usando solo informacion visual, i.e. la postura humana desarrollada durante una accion, analizando la la informacion de unos cuantos frames en lugar de examinar la secuencia completa. En este metodo, las acciones se representan mediante un modelo Bag-of-\textit{key-poses} para poder capturar la variaci\'{o}n de la postura humana durante el desarrollo de una accion. Para hacer frente al problema del reconocimiento de la accion en escenas complejas, a continuacion se propone una aproximacion model free basada en el analisis de puntos de interes espacio-temporales (STIPs) que disponen de mucha informacion local. Para este fin, se ha desarrollado un nuevo detector de STIPs que se basa en el mecanismo de inhibici\'{o}n del campo receptivo utilizado en la corteza primaria, en particular en la orientacion selectiva visual de las neuronas. Ademas, hemos extendido nuestro reconocimiento de acciones basado en STIPs selectivos a sistemas multi-camara. En este caso, los STIPs selectivos de cada punto de vista se combinan mediante los datos $3$D reconstruidos para formar STIPs selectivos 4D (espacio 3D + tiempo). En la parte final de esta tesis, nos dedicamos al reconocimiento continuo de eventos visuales (CVER) en bases de datos de videos de seguridad enormes, con un gran conjunto de datos. Este problema es extremadamente dificil debido a la alta escalabilidad de los datos, a las dificultades del entorno real en el que se aplcia y a una variabilidad en escena muy amplio. Para abordar estos problemas, las regiones en movimiento son detectadas a partir de una tecnica llamada max margin generalized Hough transformation, que se utiliza para aprender aquella distribucion de caracteristicas entorno a una accion para reconocer hipotesis que luego se verifican por Bag-of-words mas un clasificador lineal. Hemos validado nuestras tecnicas en varios conjuntos de datos de video vigilancia que constituyen el estado del arte actual en este tema. Los resultados obtenidos demuestran que hemos mejorado la precision en la deteccion de acciones humanas en video.
Automatic understanding of human activity and action is a very important and challenging research area of Computer Vision with wide scale applications in video surveillance, motion analysis, virtual reality interfaces, robot navigation and recognition, video indexing, content based video retrieval, HCI, health care, choreography and sports video analysis etc. This thesis presents a series of techniques to solve the problem of human action recognition in video. First approach towards this goal is based on the a probabilistic optimization model of body parts using hidden markov model (HMM). This strong model based approach is able to distinguish between similar actions by only considering the body parts having major contributions to the actions, for example legs for walking and jogging; arms for boxing and clapping. Next approach is based on the observation that the action recognition can be done using only the visual cue, i.e. human pose during the action, even with the information of few frames instead of examining the whole sequence. In this method, actions are represented by a Bag-of-key-poses model to capture the human pose variation during an action. To tackle the problem of recognizing the action in complex scenes, we propose a model free approach which is based on the Spatio-temporal interest point (STIP) and local feature. To this end, a novel STIP detector is proposed which uses a mechanism similar to that of the non-classical receptive field inhibition that is exhibited by most orientation selective neurons in the primary visual cortex. An extension of the selective STIP based action recognition is applied to the human action recognition in multi-camera system. In this case, selective STIPs from each camera view point are combined using the 3D reconstructed data, to form 4D STIPs [3D space + time] for multi-view action recognition. The concluding part of the thesis dedicates to the continuous visual event recognition (CVER) on large scale video dataset. This is an extremely challenging problem due to high scalability, diverse real environment state and wide scene variability. To address these issues, a motion region extraction technique is applied as a preprocessing step. A max-margin generalized Hough Transform framework is used to learn the feature vote distribution around the activity center to obtain an activity hypothesis which is verified by a Bag-of-words + SVM action recognition system. We validate our proposed approaches on several benchmark action recognition datasets as well as small scale and large scale activity recognition datasets. We obtain state-of-the results which shows a progressive improvement of our proposed techniques to solve human action and activity recognition in video.

Human standing posture depends on sensory inputs from vision, from touch and from the vestibular system. Humans use these inputs to determine position in the environment and to compensate for external forces resulting in constant swaying. This thesis investigated the creation of a four-degree-of-freedom robotic model of human upright stance using only hip actuation. In this model, the ankles are free so that there is no reaction torque from the ground and balance is provided by torques at the hip joint, resulting in an underactuated system. The mathematical model of the system was first derived using the Euler-Lagrange formulation. Using this model, simulations on the system were performed with different inputs to control the actuators. The robot was then designed and built. This was followed by tests done on the robot to examine its utility as a robotic model of human stance. From these tests, a new design with possible improvements was then proposed.

Human operators often employ discontinuous or "bang-bang" control strategies when performing large-amplitude acquisition tasks. The current study applies Variable Structure Control (VSC) techniques to model human operator behavior during acquisition tasks. The result is a coupled, multi-input model replicating the discontinuous control strategy. In the VSC formulation, a switching surface is the mathematical representation of the operator's control strategy. The performance of the Variable Strategy Model (VSM) is evaluated by considering several examples, including the longitudinal control of an aircraft during the visual landing task. The aircraft landing task becomes an acquisition maneuver whenever large initial offsets occur. Several different strategies are explored in the VSM formulation for the aircraft landing task. First, a switching surface is constructed from literal interpretations of pilot training literature. This approach yields a mathematical representation of how a pilot is trained to fly a generic aircraft. This switching surface is shown to bound the trajectory response of a group of pilots performing an offset landing task in an aircraft simulator study. Next, front-side and back-side landing strategies are compared. A back-side landing strategy is found to be capable of landing an aircraft flying on either the front side or back side of the power curve. However, the front-side landing strategy is found to be insufficient for landing an aircraft flying on the back side. Finally, a more refined landing strategy is developed that takes into the account the specific aircraft's dynamic characteristics. The refined strategy is translated back into terminology similar to the existing pilot training literature.
Ph. D.

The objective of this thesis is to design, analyze, and fabricate an innovative backpack energy harvester for human walking. To model human walking with backpack energy harvester, a simple dual-mass model has been developed and studied first. Dual-mass model for three types of distinct harvesters were investigated, pure damping, traditional rack pinion energy harvester and our MMR based energy harvester. A comparison in the output power and human comfort between the three types of harvesters is discussed. However, the dual-mass model could not effectively represent human walking in real situation with sinusoidal input, like M shaped Ground Reaction Force (GRF), vertical Center of Mass (COM) motion, etc. Thus, a bipedal walking model has been proposed to simulate human walking with backpack harvester. Experiments were conducted to compare power output and efficiency of MMR based backpack energy harvester with traditional rack pinion backpack energy harvester, and verify conclusions from the bipedal walking model that the proposed backpack energy harvester using mechanical motion rectifier (MMR) mechanism has larger power output than traditional backpack energy harvester at different walking speed. In human treadmill test, subjects were asked to wear the backpack frame which embedded with harvesters walking on a treadmill. Two walking speed, 3mph and 3.5mph, and four resistor values has been tested. The test results showed that the MMR based backpack energy harvester generated more power regardless of resistor values and walking speed. Up to 4.84W average power and instant power of 12.8W could be obtained while the subject walking on the treadmill at 3.5mph speed with MMR based backpack energy harvester.
Master of Science

Iron is essential for virtually all organisms, yet it can be highly toxic if not properly regulated. Only the Lyme disease pathogen Borrelia burgdorferi has evolved to not require iron (Aguirre et al., 2013).Recent findings have characterised elements of the iron metabolism network, but understanding of systemic iron regulation remains poor. To improve understanding and provide a tool for in silico experimentation, a computational model of human iron metabolism has been constructed. COPASI was utilised to construct a model that included detailed modelling of iron metabolism in liver and intestinal cells. Inter-cellular interactions and dietary iron absorption were included to create a systemic computational model. Parameterisation was performed using a wide variety of literature data. Validation of the model was performed using published experimental and clinical findings, and the model was found to recreate quantitatively and accurately many results. Analysis of sensitivities in the model showed that, despite enterocytes being the only route of iron uptake, almost all control over the system is provided by reactions in the liver. Metabolic control analysis identified key regulatory factors and potential therapeutic targets. A virtual haemochromatosis patient was created and compared to a simulation of a healthy human. The redistribution of control in haemochromatosis was analysed in order to improve our understanding of the condition and identify promising therapeutic targets. Cellular prion protein (PrP) is an enigmatic protein, implicated in disease when misfolded, but its physiological role remains a mystery. PrP was recently found to have ferric-reductase capacity. Potential sites of ferric reduction were simulated and the findings compared to PrP knockout mice experiments. I propose that the physiological role of PrP is in the chemical reduction of endocytosed ferric iron to its ferrous form following transferrin receptor-mediated uptake.

We augment a model of endogenous credit cycles by Matsuyama et al.(2016) with human capital to study the impact of human capital on the stability of central economic aggregates. Thus we offer a linkage between human capital formation and credit market instability on a macrolevel combined with an analysis of functional income distribution. Human capital is modelled as pure external effect of production following a learning-by-producing approach. Agents have access to two different investment projects, which differ substantially in their next generations spillover effects. Some generate pecuniary externalities and technological spillovers through human capital formation whereas others fail to do so and are subject to financial frictions. Due to this endogenous credit cycles occur and a pattern of boom and bust cycles can be observed. We explore the impact of human capital on the stability of the system by numerical simulations which indicate that human capital has an ambiguous effect on the evolution of the output. Depending on the strength of the financial friction and the output share of human capital it either amplifies or mitigates output fluctuations. This analysis shows that human capital is an essential factor for economic stability and sustainable growth as a high human capital share tends to make the system's stability robust against shocks.
Series: Department of Economics Working Paper Series

Aim: We have recently developed electrophysiological recordings of human visceral afferent (HVA) activity in isolated gastrointestinal tissues. The aim of the this study was 1) test the mechano- and chemosensitivity of HVAs, 2) characterise subpopulations of HVAs based on their response to mechanical stimuli, 3) test the effect of drugs that have/are in clinical trials on the mechanosensitivity (von Frey hair (VFH) probing and appendix distension) of HVAs. Methods: All experiments were performed in accordance with UK human ethics regulations [NREC09/H0704/2]. Surgically resected human ileum, colon, and appendix were obtained from consenting patients undergoing bowel resection. Tissues were pinned in a tissue bath, or cannulated (appendix), and superfused with carbongenated Krebs buffer, at 6ml/min, 32-34°C. Mesenteric nerve bundles were carefully dissected and afferent activity was recorded using suction electrodes. Tissues were tested for mechanosensitivity (VFHs, stretching, mucosal stroking, distension) and chemosensitivity (bradykinin (BK), ATP (adenosine trisphosphate), PGE2 (prostaglandin E2), serotonin (aka 5-hydroxytryptamine (5-HT)), histamine, adenosine). The receptors involved in the activation of HVAs by BK, or ATP were also investigated. The response of HVAs to VFH probing or distension was tested before and after the application of tegaserod, STa endotoxin, or a transient receptor potential vanilloid 4 (TRPV4) agonist (GSK1016790A) or antagonist (HC067047). Results and Conclusion: HVAs were characterised as mesenteric, serosal, muscular, or muscular-mucosal. HVAs were chemosensitive to all mediators. Bradykinin B2 receptors are the most important receptors involved in the activation of HVAs by BK. P2Y receptors may play an important role in the activation of HVAs by ATP. Application of tegaserod, HC067047 or STa endotoxin reduced the HVA response to mechanical stimuli. HVA recordings are feasible and practical and are suitable for both basic scientific mechanistic studies, and could potentially be used as a pre-clinical model, in conjunction with animal experiments, to help predict the efficacy of novel compounds before clinical trials.

The three chapters of this thesis explore how previous experience and mental categories shape human judgments. Chapter One provides a sampling explanation for the in-group heterogeneity effect - a tendency of people to perceive the groups they belong to as more heterogeneous than the groups to which they do not belong. It notes that because people are more likely to interact with the in-group members, they will experience more variability of the in-group than the out-group. Chapters 2 and 3 investigate how mental categories affect feature-based inferences when the category of the object, people perceive, is uncertain. In an influential paper, Anderson (1991) proposed a rational model for this task. The model proposes that the information from all the mental categories is integrated to make a prediction about unobserved features of the object. A crucial feature of this model is the conditional independence assumption – it assumes that the within-category feature correlation is zero. In prior research, this model has been found to provide a poor fit to participants’ inferences. Chapter 2 argues that the failure of Anderson’s rational model stems from the inconsistency between the task environment and the core assumption of the model. It notes that the studies reported in existing research relied on task environments without conditional independence and shows that when this assumption is satisfied, Anderson’s model performs well. Chapter 3 proposes an extension of Anderson’s model that allows mental categories to be characterized by feature correlations and shows that such general rational model provides a good fit to the existing inference data from experiments with uncertain categorization.
Els tres capítols de la tesi estudien com experiències prèvies i categories mentals poden determinar els judicis humans. El capítol 1 aporta una explicació a l’efecte heterogeni “in-group” - la tendència de la gent a percebre els grups als quals pertanyen com a més heterogenis que els grups als quals no pertanyen. L’estudi reporta que, com que la gent interactua més amb els membres del seu in-group, experimentaran més variabilitat en aquest in-group que en el seu out-grup. Els capítols 2 i 3 estudien com les categories mentals afecten les inferències basades en característiques quan la categoria de l’objecte, segons la gent, és incerta. En un article molt influent, Anderson (1991) va proposar un model racional per aquesta tasca. El model proposa que la informació de totes les categories mentals està integrada per fer una predicció sobre les característiques de l’objecte que no son observables. Una característica principal del model és el supòsit d’independència lineal - assumeix que dins de cada categoria, la correlació entre característiques és zero. En estudis previs, s’ha provat que aquest model sembla tenir un encaix pobre amb les inferències dels participants. El capítol 2 estableix que la problemàtica del model racional d’Anderson ve per la inconsistència entre el disseny de la tasca i el supòsit central del model. S’estableix que els estudis existents es basen en aquest disseny de tasca sense complir amb el supòsit d’independència lineal i es demostra que, quan aquest supòsit es compleix, el model d’Anderson funciona correctament. El capítol 3 proposa una extensió del model d’Anderson que permet caracteritzar les categories mentals amb característiques correlacionades entre sí i es prova que, aquest model racional encaixa correctament amb les dades provinent d’experiments existents en inferències sobre categories incertes.

Human pose estimation represents the skeleton of a person in color or depth images to improve a machine’s understanding of human movement. 3D human pose estimation uses a three-dimensional skeleton to represent the human body posture, which is more stereoscopic than a two-dimensional skeleton. Therefore, 3D human pose estimation can enable machines to play a role in physical education and health recovery, reducing labor costs and the risk of disease transmission. However, the existing datasets for 3D pose estimation do not involve fast motions that would cause optical blur for a monocular camera but would allow the subjects’ limbs to move in a more extensive range of angles. The existing models cannot guarantee both real-time performance and high accuracy, which are essential in physical education and health recovery applications. To improve real-time performance, researchers have tried to minimize the size of the model and have studied more efficient deployment methods. To improve accuracy, researchers have tried to use heat maps or point clouds to represent features, but this increases the difficulty of model deployment. To address the lack of datasets that include fast movements and easy-to-deploy models, we present a human kinetic dataset called the Kivi dataset and a hybrid model that combines the benefits of a heat map-based model and an end-to-end model for 3D human pose estimation. We describe the process of data collection and cleaning in this thesis. Our proposed Kivi dataset contains large-scale movements of humans. In the dataset, 18 joint points represent the human skeleton. We collected data from 12 people, and each person performed 38 sets of actions. Therefore, each frame of data has a corresponding person and action label. We design a preliminary model and propose an improved model to infer 3D human poses in real time. When validating our method on the Invariant Top-View (ITOP) dataset, we found that compared with the initial model, our improved model improves the mAP@10cm by 29%. When testing on the Kivi dataset, our improved model improves the mAP@10cm by 15.74% compared to the preliminary model. Our improved model can reach 65.89 frames per second (FPS) on the TensorRT platform.

The work is devoted to Human-computer interaction and its main goal is to get closer to the field. The first part describes two main areas, namely Cognitive Science and Cogni-tive Psychology, from which HCI is based on. The second part deals specifically with Human-computer interaction, the history of the origins, aspects of human factor, ele-ments of interaction and modeling of interaction together with examples from everyday life.

Erythropoeitic protoporphyria (EPP) is a member of the prophyria disease class and is caused by abnormal function of the enzyme ferrochelatase (Fech). In humans it has variable penetrance, but primarily leads to toxicity in the skin and liver to varying degrees. Here I have investigated the nature of EPP progression using the Fechm1PAS mouse model. This mouse contains a point mutation in the fech gene which results in reduced Fech activity to < 7% of wild type, with resultant loss of haem, anaemia and hepatic cholestasis. The phenotypic progression of Fechm1PAS/m1PAS mice was established using pathology and clinical biochemistry from 18 days gestation to 32 weeks of age. Pathological changes were found from 4 weeks with biochemical and differential gene expression (DGE) analysis showing intraheptic cholestasis from birth. Genomic data from cDNA microarrays was derived and analysed with the DGE by phenotypic anchoring. DGE was observed in all processes responsible for cell protection and epigenetic regulation. DGE analysis have led me to hypothesise that the porphyria leads to a chronic reactive oxygen species (ROS) attack, causing DNA damage, eventually leading to hepatocarinoma. This was indicated by changes in the GSH, cytochrome P450, circadian rhythm and methylation pathways. DGE in these processes included downregulation of DNA methyltransferases (Dnmt1, Dnmt6a and Dnmt6b), and upregulation of cytochrome oxidase (Cox and Por) and GSH metabolism transcription factors (Gclc and Gclm). The findings made here contribute further to the understanding of EPP progression and the relationship between phenotype and DGE in EPP.

The studies presented in this thesis focus on human and canine models for autoimmune disease, with the main aim to gain new knowledge about disease mechanisms and to further evaluate the dog as a model for autoimmune disease. Autoimmune Polyendocrine Syndrome type 1 (APS-1) is a hereditary human multiorgan disease caused by mutations in the autoimmune regulator (AIRE) gene. Hallmarks of APS-1 are chronic mucocutaneous candidiasis caused by Candida albicans, together with the autoimmune endocrine disorders hypoparathyroidism and adrenocortical failure. Many human diseases have an equivalent disease in dogs. Because humans share environment, and in part life style with the dogs they provide an interesting model for further genetic studies. Immune responses to Candida albicans in APS-1 patients displayed an increased secretion of the proinflammatory cytokine IL-17A and similar results were also found in AIRE deficient mice. Anticytokine autoantibodies to IL-17A, IL-17F and IL-22 were detected in APS-1 patients, and a radioligand binding assay for measuring these autoantibodies was developed and evaluated. In the canine studies we investigated whether canine diabetes mellitus could serve as a model for human autoimmune diabetes mellitus. Furthermore, we investigated type I IFN responses in Nova Scotia duck tolling retriever dogs with a systemic autoimmune disease resembling human SLE. Four assays were used in search for signs of humoral autoimmunity in diabetic dogs. However, no evidence for a type 1 diabetes-like phenotype in dogs was found. Sera from Nova Scotia duck tolling retrievers suffering from steroid-responsive meningitis arteritis elicited an increased expression of IFN-inducible genes in the canine MDCK cell line. This suggests that these dogs have an IFN signature, as seen in human SLE.

For sleep classification, automatic electroencephalogram (EEG) interpretation techniques are of interest because they are labour saving, in contrast to manual (visual) methods. More importantly, some automatic methods, which offer a less subjective approach, can provide additional information which it is not possible to obtain by manual analysis. An extensive literature review has been undertaken to investigate the background of automatic EEG analysis techniques. Frequency domain and time domain methods are considered and their limitations are summarised. The weakness in the R & K rules for visual classification and from which most of the automatic systems borrow heavily are discussed. A new technique - model based dynamic analysis - was developed in an attempt to classify the sleep EEG automatically. The technique comprises of two phases, these are the modelling of EEG signals and the analysis of the model's coefficients using dynamic systems theory. Three techniques of modelling EEG signals are compared: the implementation of the non-linear prediction technique of Schaffer and Tidd (1990) based on chaos theory; Kalman filters and a recursive version of a radial basis function for modelling and forecasting the EEG signals during sleep. The Kalman filter approach produced good results and this approach was used in an attempt to classify the EEG automatically. For classifying the model's (Kalman filter's) coefficients, a new technique was developed by a state-space approach. A 'state variable' was defined based on the state changes of the EEG and was shown to be correlated with the depth of sleep. Furthermore it is shown that this technique may be useful for automatic sleep staging. Possible applications include automatic staging of sleep, detection of micro-arousals, anaesthesia monitoring and monitoring the alertness of workers in sensitive or potentially dangerous environments.

Introduction. Ototoxicity refers to medication-caused auditory and/or vestibular system dysfunction resulting in hearing loss or dysequilibrium. The potential damage that antibiotics eardrops can produce when placed directly into the middle ear in some cases is still unknown.
Objectives. To determine the safety of use of the new ciprofloxacin/dexamethasone otic drops in patients without an intact tympanic membrane.
Materials and methods. Ciprodex/dexamethasone eardrops were tested in an animal and human model. The animal part was performed in 13 adult chinchillas; Auditory Brainstem Response (ABR) was used. For the human part, twenty subjects were enrolled in the study; Distortion Products Otoacoustic Emissions (DPOAE) testing was used.
Results. Animal Part: after the tube insertion ABR threshold mean value was 19.6+/-13.3 dB for all the animals. On the last evaluation (day 60), the mean threshold was 19+/-13 dB for the experimental ears, and 13.7+/-12.2 dB for the control ears, this overall analysis showed no significant effect (p-value = 0.661). Human Part: the mean thresholds for the pre-treatment test were 4.87+/-6,34 dB for the DP value and -0.87+/-7.93 dB for the Ns value. In the post-treatment evaluation the mean thresholds were 3.48+/-4.40 dB for the DP value and -8.02+/-7.57 dB for the Ns value.
Conclusions. The use of CiprodexTM eardrops seems to be safe when instilled in ears without an intact tympanic membrane.

Research in Sentence Processing is concerned with discovering the mechanism by which linguistic utterances are mapped onto meaningful representations within the human mind. Models of the Human Sentence Processing Mechanism (HSPM) can be divided into those in which such mapping is performed by a number of limited modular processes and those in which there is a single interactive process. A further, and increasingly important, distinction is between models which rely on innate preferences to guide decision processes and those which make use of experience-based statistics. In this context, the aims of the current thesis are two-fold: - To argue that the correct architecture of the HSPM is both modular and statistical - the Modular Statistical Hypothesis (MSH). - To propose and provide empirical support for a position in which human lexical category disambiguation occurs within a modular process, distinct from syntactic parsing and guided by a statistical decision process. Arguments are given for why a modular statistical architecture should be preferred on both methodological and rational grounds. We then turn to the (often ignored) problem of lexical category disambiguation and propose the existence of a pre-syntactic Statistical Lexical Category Module (SLCM). A number of variants of the SLCM are introduced. By empirically investigating this particular architecture we also hope to provide support for the more general hypothesis - the MSH. The SLCM has some interesting behavioural properties; the remainder of the thesis empirically investigates whether these behaviours are observable in human sentence processing. We first consider whether the results of existing studies might be attributable to SLCM behaviour. Such evaluation provides support for an HSPM architecture that includes this SLCM and allows us to determine which SLCM variant is empirically most plausible. Predictions are made, using this variant, to determine SLCM behaviour in the face of novel utterances; these predictions are then tested using a self-paced reading paradigm. The results of this experimentation fully support the inclusion of the SLCM in a model of the HSPM and are not compatible with other existing models.

Approved for public release; distribution is unlimited
Current fiscal constraints are driving the reduction of system life cycle cost (LCC). A key objective of HSI is the reduction of operational cost and the improvement of operational performance. This thesis seeks to develop a Human Systems Integration (HSI) Synthesis Model for Ship Design. This model is based on the premise that ship design characteristics interact with the domains of HSI. The thesis begins with an historical overview of ship architecture and technology and their interactions with the domains of HSI. The HSI Synthesis Model for Ship Design was developed using the Framework of Naval Postgraduate Schools Systems Engineering Ship Synthesis Model. Quantitative data analysis was conducted using Offshore Patrol Vessel (OPV) design data from Information Handling Services (IHS) Janes database. The data analyzed included 35 ships from 21 nations. Multiple regression analysis consisted of nine independent ship design variables and a response variable of manpower. Data analysis revealed that ship length and ship draught were statistically significant. The proposed HSI Synthesis Model accounted for 49 per cent of the variance of crew complement. This thesis lays the foundation for future qualitative and quantitative analysis of the interaction between ship design characteristics and HSI domains. Additionally, it provides an initial HSI model that can be expanded upon by including additional HSI domains and, ultimately, may lead to a viable design tool for HSI practitioners and systems engineers.

A finite-element model of a two-dimensional slice of human atrial tissue for the study of the electrograms and propagation of action potentials is presented. Action potential propagation is described by a reaction-diffusion model coupled with the complex Courtemanche et al. atrial cell model. The effects of recording electrode size and location on electrograms are presented. Action potential propagation as a result of atrial fibrillation ablation therapy is also modeled by defining a lesion area with decreased electrical conductivity. The effect of electrical conductivity and geometry of the lesion was also studied. It is shown that the success rate of atrial fibrillation ablation therapy is primarily correlated with the creation of continuous ablation lines and also the formation of fibrous tissue.

The purpose of this thesis was to develop a 3-dimensional (3D) reconstruction algorithm which can more accurately generate a 3D shape model of a human face from a single 2-dimensional (2D) image. Like most present day approaches to building 3D human facial models, the proposed algorithm stemmed from the statistical approach pioneered by Blanz and Vetter [13], who used a morphable 3D model which was con gured to match a 2D image. With current 3D shape model reconstruction using statistical modelling, the accuracy of the reconstruction is highly dependent on the nature of the source data and the behaviour of the feature extraction algorithm. The most commonly applied representation algorithm for 3D modelling is principal component analysis (PCA). However, there are inherent problems in using principal components as a feature space. Firstly, the principal components only represent the directions of the maximum variance in the source data and therefore, may be better at representing general trends rather than subtle individual shape di erences. Secondly, the principal components in PCA are orthogonal, hence they are better suited for representing multivariate, Gaussian source data. Although PCA usually provides a good basis for reconstruction, it is not necessarily more accurate when analysing data that provides low variance directions. The reconstruction approach proposed in this thesis used two sets of information computed by both, PCA and independent component analysis (ICA) to recover the geometric information from a single image. The reconstruction was conducted in PCA and ICA feature spaces in succession. This method allowed both the global trends and the local subtle features to be preserved and well represented. To reinforce the robustness of the algorithm, kernel canonical correlation analysis( KCCA) was applied to determine the relationship between texture information and depth. A quantitative analysis was then applied to test the performance of the approach. The results showed that the proposed algorithm generated more accurate results than reconstructing 3D models with PCA or ICA alone or in succession. However, by examining the performance of the proposed hybrid reconstruction algorithm, we have found that the Z-dimensional shape error was signi cantly larger than the other two dimensions. This was due to the limited shape information contained within the 2D image. Subsequently, we used local binary patterns (LBP) to encode the texture information of the 2D image. Based on the LBP codes and their corresponding depth values to the shape models used in the training shape model database, we used kernel canonical analysis (KCCA) to train a depth value predictor. We then added the depth value predictor into the hybrid shape reconstruction algorithm to predict the depth values of the landmarks of the objective face. The depth values and the existing 2D information of the landmarks from the input image were then used to generate the nal 3D shape reconstruction. The enhanced shape reconstruction algorithm based on the above methodology was then tested using the Binghamton Human Face Database [147]. The results showed that the enhanced shape reconstruction algorithm greatly reduced the Z-dimensional shape error while maintaining the same accuracy in the other two dimensions. Thus the overall accuracy of the reconstruction operation has been signi cantly improved from the hybrid shape reconstruction algorithm.

Recent adaptationist accounts of human mental and physical health have reinvigorated the debate over the evolution of human intelligence. In the tradition of strong inference the current study was developed to determine which hypothesis; Rushton’s (2000) differential K theory, or Miller’s (2000a) fitness indicator model (F), better accounts for general intelligence ('g') in an undergraduate university population (N = 194). Due to the lengthy administration time of the test materials a newly developed 18-item short form of the Ravens Advanced Progressive Matrices (APM-18; Sefcek, Miller, & Figueredo, 2007) was used. There was a significant positive relationship between K and F (r = .31, p < .001), however no significant relationships were found between 'g' and either K or F (for each, r = -.06, p ≥ .05). While contrary to both hypotheses, these results may be explained in relation to antagonistic pleiotropy and a potential failure to derive within species comparisons directly from between species comparisons.

We consider a new approach to digital human simulation, using Model Predictive Control (MPC). This approach permits a virtual human to react online to unanticipated disturbances that occur in the course of performing a task. In particular, we predict the motion of a virtual human in response to two different types of real world disturbances: impulsive and sustained. This stands in contrast to prior approaches where all such disturbances need to be known a priori and the optimal reactions must be computed off line. We validate this approach using a planar 3 degrees of freedom serial chain mechanism to imitate the human upper limb. The response of the virtual human upper limb to various inputs and external disturbances is determined by solving the Equations of Motion (EOM). The control input is determined by the MPC Controller using only the current and the desired states of the system. MPC replaces the closed loop optimization problem with an open loop optimization allowing the ease of implementation of control law. Results presented in this thesis show that the proposed controller can produce physically realistic adaptive simulations of a planar upper limb of digital human in presence of impulsive and sustained disturbances.

Skin is the largest, and arguably, the most important organ of the body. It is a complex and multi-dimensional tissue, thus making it essentially impossible to fully model in vitro in conventional 2-dimensional culture systems. In view of this, rodents or pigs are utilised to study wound healing therapeutics or to investigate the biological effects of treatments on skin. However, there are many differences between the wound healing processes in rodents compared to humans (contraction vs. re-epithelialisation) and there are also ethical issues associated with animal testing for scientific research. Therefore, the development of skin equivalent (HSE) models from surgical discard human skin has become an important area of research. The studies in this thesis compare, for the first time, native human skin and the epidermogenesis process in a HSE model. The HSE was reported to be a comparable model for human skin in terms of expression and localisation of key epidermal cell markers. This validated HSE model was utilised to study the potential wound healing therapeutic, hyperbaric oxygen (HBO) therapy. There is a significant body of evidence suggesting that lack of cutaneous oxygen results in and potentiates the chronic, non-healing wound environment. Although the evidence is anecdotal, HBO therapy has displayed positive effects on re-oxygenation of chronic wounds and the clinical outcomes suggest that HBO treatment may be beneficial. Therefore, the HSE was subjected to a daily clinical HBO regime and assessed in terms of keratinocyte migration, proliferation, differentiation and epidermal thickening. HBO treatment was observed to increase epidermal thickness, in particular stratum corneum thickening, but it did not alter the expression or localisation of standard epidermal cell markers. In order to elucidate the mechanistic changes occurring in response to HBO treatment in the HSE model, gene microarrays were performed, followed by qRT-PCR of select genes which were differentially regulated in response to HBO treatment. The biological diversity of the HSEs created from individual skin donors, however, overrode the differences in gene expression between treatment groups. Network analysis of functional changes in the HSE model revealed general trends consistent with normal skin growth and maturation. As a more robust and longer term study of these molecular changes, protein localisation and expression was investigated in sections from the HSEs undergoing epidermogenesis in response to HBO treatment. These proteins were CDCP1, Metallothionein, Kallikrein (KLK) 1 and KLK7 and early growth response 1. While the protein expression within the HSE models exposed to HBO treatment were not consistent in all HSEs derived from all skin donors, this is the first study to detect and compare both KLK1 and CDCP1 protein expression in both a HSE model and native human skin. Furthermore, this is the first study to provide such an in depth analysis of the effect of HBO treatment on a HSE model. The data presented in this thesis, demonstrates high levels of variation between individuals and their response to HBO treatment, consistent with the clinical variation that is currently observed.