Equivalência de estímulos, um fenômeno comportamental amplamente demonstrado em humanos, tem sido encontrada com dificuldades em não-humanos, provavelmente devido a dificuldades de procedimento. Controle não programado pela posição dos estímulos tem sido apontado como um fator que interfere no desempenho de "matching-to-sample" e que leva a resultados negativos em testes de propriedades definidoras de equivalência (reflexividade, simetria e transitividade). Duas linhas de pesquisa foram exploradas: 1) Posição como estímulo em discriminações condicionais e testes de propriedades emergentes e 2) Treino e testes de relações condicionais na ausência de correlação entre a posição e a função dos estímulos. Foram conduzidos quatro experimentos com três macacos Cebus apella, uma fêmea e dois machos. No Experimento I, os sujeitos foram submetidos a treino de discriminações condicionais e testes de propriedades emergentes com posições como estímulo. Resultados positivos nos testes foram encontrados quando as relações testadas eram topograficamente semelhantes às relações treinadas. No Experimento II, treinos de discriminações simples com três conjuntos de estímulos visuais (cores e formas arbitrárias) mostraram diferenças na "discriminabilidade" dos estímulos. No Experimento III os sujeitos foram submetidos a treino de pareamento por identidade e testes de reflexividade com dois dos três conjuntos de estímulos facilmente discriminados no Experimento II e mais três novos conjuntos de estímulos. Os resultados dos testes foram positivos quando os estímulos já tinham sido apresentados antes em treinos de discriminações simples simultâneas, que possivelmente funcionaram como um treino de "matching-to-sample" com atraso. No Experimento IV os sujeitos foram submetidos a treinos de relações condicionais arbitrárias e a testes de simetria, cujos resultados foram negativos. Escolhas corretas foram reforçadas em todas as tentativas, em todos os experimentos. Todas as tentativas de "matching-to-sample" eram com atraso zero. A interpretação dos dados baseia-se na suposição de que relações de equivalência são relações ambientais. O pré-requisito para a emergência das propriedades definidoras é a aquisição de controle por essas propriedades enquanto relações ambientais, o que demanda extenso treino. / Stimulus equivalence, a largely demonstrated behavioral phenomenon in humans has been hardly found in non-humans likely due to procedural difficulties. Unprogrammed control by stimulus position has been pointed as interfering with matching-to-sample performances, which leads to negative results in tests of equivalence defining properties (reflexivity, symmetry and transitivity). Two lines of research were pursued that investigate: 1) Location as stimulus in conditional discriminations and tests of emergent properties and 2) Training and tests of conditional relations with no relation between location and function of stimuli. Four experiments with 1 female and 2 male Cebus apella were conducted. In Experiment I subjects were submitted to training of conditional discriminations and tests of emergent properties with location as stimulus, with positive results when the tested relations were similar to those trained. In Experiment II, training of simultaneous simple discriminations with three sets of visual stimuli (colors and arbitrary forms) showed differences in discriminability of the stimuli. In Experiment III, with two of the easily discriminated stimulus sets and three new stimulus sets, subjects received training of identity matching-to-sample and reflexivity tests, with positive results when the stimuli had been presented before in simultaneous simple discriminations, that possibly functioned as delayed identity matching-to-sample. In Experiment IV, subjects were submitted to arbitrary matching-to-sample training, and to symmetry tests with negative results. Correct choices in all trials were reinforced. All matching-to-sample trials were zero-delay. Interpretation of data relay on the assumption that equivalence relations are environmental relations. The prerequisite for defining properties emergence may be the acquisition of control by these defining properties as environmental relations, which demand extended training.

Activity retrieval is a growing field in electrical engineering that specializes in the search and retrieval of relevant activities and events in video corpora. With the affordability and popularity of cameras for government, personal and retail use, the quantity of available video data is rapidly outscaling our ability to reason over it. Towards the end of empowering users to navigate and interact with the contents of these video corpora, we propose a framework for exploratory search that emphasizes activity structure and search space reduction over complex feature representations. Exploratory search is a user driven process wherein a person provides a system with a query describing the activity, event, or object he is interested in finding. Typically, this description takes the implicit form of one or more exemplar videos, but it can also involve an explicit description. The system returns candidate matches, followed by query refinement and iteration. System performance is judged by the run-time of the system and the precision/recall curve of of the query matches returned. Scaling is one of the primary challenges in video search. From vast web-video archives like youtube (1 billion videos and counting) to the 30 million active surveillance cameras shooting an estimated 4 billion hours of footage every week in the United States, trying to find a set of matches can be like looking for a needle in a haystack. Our goal is to create an efficient archival representation of video corpora that can be calculated in real-time as video streams in, and then enables a user to quickly get a set of results that match. First, we design a system for rapidly identifying simple queries in large-scale video corpora. Instead of focusing on feature design, our system focuses on the spatiotemporal relationships between those features as a means of disambiguating an activity of interest from background. We define a semantic feature vocabulary of concepts that are both readily extracted from video and easily understood by an operator. As data streams in, features are hashed to an inverted index and retrieved in constant time after the system is presented with a user's query. We take a zero-shot approach to exploratory search: the user manually assembles vocabulary elements like color, speed, size and type into a graph. Given that information, we perform an initial downsampling of the archived data, and design a novel dynamic programming approach based on genome-sequencing to search for similar patterns. Experimental results indicate that this approach outperforms other methods for detecting activities in surveillance video datasets. Second, we address the problem of representing complex activities that take place over long spans of space and time. Subgraph and graph matching methods have seen limited use in exploratory search because both problems are provably NP-hard. In this work, we render these problems computationally tractable by identifying the maximally discriminative spanning tree (MDST), and using dynamic programming to optimally reduce the archive data based on a custom algorithm for tree-matching in attributed relational graphs. We demonstrate the efficacy of this approach on popular surveillance video datasets in several modalities. Finally, we design an approach for successive search space reduction in subgraph matching problems. Given a query graph and archival data, our algorithm iteratively selects spanning trees from the query graph that optimize the expected search space reduction at each step until the archive converges. We use this approach to efficiently reason over video surveillance datasets, simulated data, as well as large graphs of protein data.

Les technologies de l'informatique et de la robotique sont en perpétuelle évolution. S'appuyant sur cette évolution technologique, les systèmes d’aide à l’opérateur restent un domaine de recherche d’actualité. Le principal défi des systèmes de la future génération est d'être "intelligents", sensibles au contexte dans un environnement complexe et imprévisible. Cette thèse entre dans ce cadre et traite de la compréhension dynamique du contexte par un robot évoluant dans un tel environnement. En particulier, elle s'intéresse à la question suivante: comment rendre un robot capable de réagir face aux situations de blocage, imprévues dans son plan d’action initial, pour accomplir l’objectif fixé par l’opérateur ? Dans la littérature, ce problème a été soulevé et résolu en partie en programmant dans le système robotique, certaines des fonctions rendant le robot plus autonome. Cependant, l'intégration de ces fonctions dans un même cadre est manquante et plusieurs recherches dans ce sens sont en cours. Dans nos travaux nous proposons un système supportant une approche complète et générique, qui assure à un robot la capacité d’être conscient de la situation de blocage dans laquelle il se trouve et de comprendre et faire face aux situations de blocage rencontrées. Notre approche, nommée Robot Situation AWareness (RSAW) est inspirée de la notion de Situation Awareness (SA) qui a fait ses preuves dans de nombreux domaines notamment dans l’aviation. Nos principales contributions dans RSAW portent sur la conception d’un cadre sémantique intégrant la capacité de compréhension, fondé sur une représentation des connaissances générique, donnant la possibilité d’appliquer des techniques de raisonnement empruntées aux sciences cognitives. L’intégration de RSAW dans un système robotique a également été étudiée, conçue et mise en œuvre dans un système à couches. Ce système d'expérimentation est le robot SAM (Smart Autonomous Majordomo) doté du système AVISO et développé par le CEA-LIST. Les résultats des expérimentations élaborées dans le cadre des travaux menés dans cette thèse sont concluants et prometteurs / Computer technology and robotics are in perpetual evolution. Based on this technological evolution, the operator support systems remain a topical domain of research. The main challenge for the next generation of systems is to be "intelligent", aware of the context in a complex and unpredictable environment. This thesis is into this framework and addresses the dynamic understanding of the context by a robot evolving in such an environment. In particular, the work is interested in the question: How to make a robot able to react to blocked situations unplanned in its initial action plan to achieve the goal set by the operator?In the literature, this issue was raised and resolved in part by programming in robotic system, some of the features making a robot more autonomous. However, the integration of these functions in one framework is missing and more research in this direction is underway. In our work we propose a system supporting a complete and generic approach that ensures a robot the ability to be aware of the blocking situation in which it is found, to understand and deal with deadlock situations encountered. Our approach, called Robot Situation Awareness (RSAW) is inspired by the notion of Situation Awareness (SA), which has been proven in many areas especially in aviation. Our main contributions in RSAW involve the design of a semantic framework integrating the understanding capacity, based on a generic representation of knowledge and giving the possibility to apply reasoning techniques borrowed from cognitive science. Integrating RSAW in a robotic system has also been studied, designed and implemented in a layer system. This experimental system is the robot SAM (Smart Autonomous Majordomo) with the AVISO system developed by CEA-LIST. The conducted experiments allowed testing of the deductive reasoning in resolving a blocked situation and confirmed the need to resort to analogical reasoning. Another wave of experimentation has taken place to prove the effectiveness of our choices. The results of experiments developed as part of the work in this thesis are successful and promising

La reconnaissance d'iris est un des systèmes biométriques les plus fiables et les plus précis. Cependant sa robustesse aux dégradations des images d'entrées est limitée. Généralement les systèmes basés sur l'iris peuvent être décomposés en quatre étapes : segmentation, normalisation, extraction de caractéristiques et comparaison. Des dégradations de la qualité des images d'entrées peuvent avoir des répercussions sur chacune de ses étapes. Elles compliquent notamment la segmentation, ce qui peut engendrer des images normalisées contenant des distorsions ou des artefacts non détectés. De plus, la quantité d'information disponible pour la comparaison peut être réduite. Dans cette thèse, nous proposons des solutions pour améliorer la robustesse des étapes d'extraction de caractéristiques et de comparaison à la dégradation des images d'entrées. Nous travaillons avec deux algorithmes pour ces deux étapes, basés sur les convolutions avec des filtres de Gabor 2D, mais des processus de comparaison différents. L'objectif de la première partie de notre travail est de contrôler la qualité et la quantité d'information sélectionnée pour la comparaison dans les images d'iris normalisées. Dans ce but nous avons défini des mesures de qualité locale et globale qui mesurent la quantité d'occlusions et la richesse de la texture dans les images d'iris. Nous utilisons ces mesures pour déterminer la position et le nombre de régions à exploiter pour l'extraction. Dans une seconde partie de ce travail, nous étudions le lien entre la qualité des images et les performances de reconnaissance des deux algorithmes de reconnaissance décrits ci-dessus. Nous montrons que le second est plus robuste aux images dégradées contenant des artefacts, des distorsions ou une texture pauvre. Enfin, nous proposons un système complet pour la reconnaissance d'iris, qui combine l'utilisation de nos mesures de qualités locale et globale pour optimiser les performances des algorithmes d'extraction de caractéristiques et de comparaison / Iris recognition has become one of the most reliable and accurate biometric systems available. However its robustness to degradations of the input images is limited. Generally iris based systems can be cut into four steps : segmentation, normalization, feature extraction and matching. Degradations of the input image quality can have repercussions on all of these steps. For instance, they make the segmentation more difficult which can result in normalized iris images that contain distortion or undetected artefacts. Moreover the amount of information available for matching can be reduced. In this thesis we propose methods to improve the robustness of the feature extraction and matching steps to degraded input images. We work with two algorithms for these two steps. They are both based on convolution with 2D Gabor filters but use different techniques for matching. The first part of our work is aimed at controlling the quality and quantity of information selected in the normalized iris images for matching. To this end we defined local and global quality metrics that measure the amount of occlusion and the richness of texture in iris images. We use these measures to determine the position and the number of regions to exploit for feature extraction and matching. In the second part, we study the link between image quality and the performance of the two recognition algoritms just described. We show that the second one is more robust to degraded images that contain artefacts, distortion or a poor iris texture. Finally, we propose a complete system for iris recognition that combines the use of our local and global quality metrics to optimize recognition performance

The modern engineering design process includes computer software packages that require approximations to be made when representing geometries. These approximations lead to inherent discrepancies between the design geometry of a part or assembly and the corresponding manufactured geometry. Further approximations are made during the analysis portion of the design process. Manufacturing defects can also occur, which increase the discrepancies between the design and manufactured geometry. These approximations combined with manufacturing defects lead to discrepancies which, for high precision parts, such as jet engine compressor blades, can affect the modal analysis results. In order to account for the manufacturing defects during analysis, mesh morphing is used to morph a structural finite element analysis mesh to match the geometry of compressor blades with simulated manufacturing defects. The mesh morphing process is improved by providing a novel method to convert heteromorphic shape matching within Sculptor to homeomorphic shape matching. This novel method is automated using Java and the NX API. The heteromorphic to homeomorphic conversion method is determined to be valid due to its post-mesh morphing maximum deviations being on the same order as the post-mesh morphing maximum deviations of the ideal homeomorphic case. The usefulness of the automated heteromorphic to homeomorphic conversion method is demonstrated by simulating manufacturing defects on the pressure surface of a compressor blade model, morphing a structural finite element analysis mesh to match the geometry of compressor blades with simulated manufacturing defects, performing a modal analysis, and making observations on the effect of the simulated manufacturing defects on the modal characteristics of the compressor blade.

This manuscript includes two studies. The research design for study 1 was a single-subject reversal design, while study 2 was a case study with 5 experimental conditions. These studies investigated the effects of a Tier 3 peer-matching self-management intervention on two elementary school students who had previously been less responsive to Tier 1 and Tier 2 interventions. The Tier 3 self-management intervention, which was implemented in the classroom, included daily electronic communication between teachers and the two students' parents. Results indicated that this intervention effectively reduced disruptive behaviors and increased total engagement when implemented with integrity; without integrity, results were variable.

Cystic fibrosis (CF) is the most common life-restricting, genetically inherited disease among Caucasians affecting approximately 30,000 people in the United States. Lung disease is the major cause of morbidity and mortality in CF. A number of oral, inhaled, and intravenous therapies are available to combat CF lung disease. Of these, this research project focused on inhaled dornase alfa, oral azithromycin, inhaled tobramycin, and inhaled aztreonam. Data to address three research aims were requested and obtained from the Cystic Fibrosis Foundation Patient Registry (CFFPR). The first aim examined the use of inhaled dornase alfa in younger children with CF. With no clinical efficacy data of dornase alfa in children ≤ 6 years of age, the study utilized subsequent forced expiratory volume in 1 second (FEV₁) measured between 6 - 7 years of age, to assess the effectiveness of long-term dornase alfa use ≤ 6 years of age. Propensity score methods were used to reduce the likelihood of treatment indication bias. The results suggested that receiving treatment with dornase alfa before 6 years of age did not improve FEV₁ between 6 - 7 years. Unmeasured covariates leading to treatment indication bias were likely one of the key explanations for these results. Additionally, lack of a more sensitive outcome than FEV₁ to assess lung function in young patients with early lung damage was thought to be another reason for the failure to reject the null hypothesis. The second aim assessed the long-term clinical effectiveness of chronic azithromycin use on the rate of FEV₁ decline in CF patients between 6 - 20 years of age. This study was novel in that the rate of FEV₁ decline, rather than change in FEV₁ from baseline, was the primary outcome, which was characterized using propensity score matching followed by a linear mixed model analysis. The results of the analysis suggested that the rate of FEV₁ decline was slower in patients who did not receive chronic treatment with azithromycin. Treatment indication bias was thought to play an important role in the direction of the association between treatment and outcome. Associations between FEV₁ % predicted and many of the other study variables included in the analysis were consistent with previous studies. The final aim compared the clinical effectiveness of a combination of inhaled tobramycin and aztreonam with inhaled tobramycin alone on the rate of FEV₁ decline in CF patients between 6 - 20 years of age. This aim was novel in that the effect of this combination treatment on rate of decline in FEV₁ has never been assessed. A linear mixed model analysis was used after matching patients in the two treatment groups on their propensity scores. Once again, the results were contrary to the alternative hypothesis with the combination group having a steeper rate of FEV₁ decline than the group that was treated with tobramycin alone. An important reason for this result was thought to be unresolved treatment indication bias that could not be eliminated even with the use of the propensity score methods used to test the associated hypothesis. The use of validated methods of analysis, i.e., propensity scores, to counter treatment indication bias using the largest available observational dataset for CF, was one of the key strengths of this study. Moreover, this study highlighted important weaknesses in the CFFPR with regards to lack of data on patient and physician-level variables - an area of active interest for the Cystic Fibrosis Foundation.

Auditory recognition memory in non-human primates is not well understood. Monkeys have difficulty acquiring auditory memory tasks, and limited capability maintaining auditory information over memory delays, relative to studies of visual memory. Neural substrates of auditory discrimination and recognition memory depend on superior temporal gyrus (STG), instead of rhinal cortex necessary for visual memory (Fritz et al., 2005). The current project assessed behavioral and neural correlates of auditory processing and memory function in monkeys, particularly focusing on the dorsal temporal pole (dTP), the rostral portion of STG. Chapter 2 examined recognition memory of monkeys under influences of various sound types. In a delayed matching-to-sample (DMTS) task, rhesus monkeys were trained to determine if two sounds, separated by a 5-second delay, were same (match trials) or different (nonmatch trials). Results demonstrated monkey vocalizations served as better cues than other sound types for auditory memory performance. Memory improvements may be due to familiarity and biological significance of con-specific sounds, analogous to using facial stimuli during visual tasks. Chapter 3 examined neuronal activity of dTP, when two monkeys performed an auditory DTMS task and listened to sound stimuli. Population encoding of sample stimuli in dTP was closely associated with memory accuracy. Moreover, a suppression effect on identical sounds was present, similar to processing in the ventral visual processing stream, inferior temporal cortex (ITC) and ventral temporal pole (vTP). Delay-related activity of dTP was weak, limited and short-lived, in contrast to visual studies reporting sustained activity over memory delays in ITC, vTP and prefrontal cortex. The findings provide preliminary evidence on why monkeys show limited memory capability, compared to visual memory, for auditory information. Neurons of dTP were sound-selective, and mainly evoked by one to four discrete stimuli only. Sound types and simple acoustic properties of sound stimuli cannot completely account for response profiles of dTP neurons. The findings suggest dTP is a higher order auditory area, and receives information from various auditory areas along STG. Dorsal temporal pole fits into proposals of neural networks for auditory processing, in which a hierarchical organization of information flow exists within the primate auditory nervous system.

The role of digital humans in product design and assessment is ever increasing. Accurate digital human models are used to provide feedback on virtual prototypes of products, thus reducing costs and shortening the design cycle. An essential part of product assessment in the virtual world is the ability of the human model to interact correctly and naturally with the product model. This involves reaching, grasping and manipulation. This work addresses the difficult problem of grasp planning for digital humans. We develop a semi-interactive system for synthesizing grasps based on the object's shape, and implement this system for SantosTM, the digital human developed at the Virtual Soldier Research Program at the University of Iowa. The system is composed of three main parts: First, a shape matching module that creates an initial power grasp for the object based on a database of pre-calculated grasps. Second, an optimization based module provides control of the fingertip locations. This can be used to synthesize precision grasps under the user's guidance. Finally, a grasp quality module provides feedback about the grasp's mechanical stability. The novelty of our approach lies in the fact that it takes into consideration the upper body posture when planning the grasp, so the whole arm and the torso are involved in the grasp.

Information contained in chemical compounds, fingerprint databases, social networks, and interactions between websites all have one thing in common: they can be represented as graphs. The need to analyze, compare, and classify graph datasets has become more evident over the last decade. The graph isomorphism problem is known to belong to the NP class, and the subgraph isomorphism problem is known to be an NP-complete problem. Several error-tolerant graph matching techniques have been developed during the last two decades in order to overcome the computational complexity associated with these problems. Some of these techniques rely upon similarity measures based on the topology of the graphs. Random walks and edit distance kernels are examples of such methods. In conjunction with learning algorithms like back-propagation neural networks, k-nearest neighbor, and support vector machines (SVM), these methods provide a way of classifying graphs based on a training set of labeled instances. This thesis presents a novel approach to error-tolerant graph matching based on current flow analysis. Analysis of current flow in electrical networks is a technique that uses the voltages and currents obtained through nodal analysis of a graph representing an electrical circuit. Current flow analysis in electrical networks shares some interesting connections with the number of random walks along the graph. We propose an algorithm to calculate a similarity measure between two graphs based on the current flows along geodesics of the same degree. This similarity measure can be applied over large graph datasets, allowing these datasets to be compared in a reasonable amount of time. This thesis investigates the classification potential of several data mining algorithms based on the information extracted from a graph dataset and represented as current flow vectors. We describe our operational prototype and evaluate its effectiveness on the NCI-HIV dataset.