Rozprawy doktorskie na temat „Categorization dynamics”

Cette thèse porte sur la question de la sémantique causative. Elle propose une typologie sémantique pour les verbes causatifs analytiques CAUSE, MAKE, HAVE, GET et LET, fondée sur le modèle de la dynamique des forces. Le premier chapitre est une étude épistémologique de l’essence de la notion de causation et de son expression linguistique en anglais. Le second chapitre situe le cadre théorique de notre analyse, à savoir la sémantique cognitive. Le troisième chapitre est un tour d’horizon des idées les plus récurrentes dans la littérature au sujet de la sémantique des verbes causatifs analytiques. Dans le quatrième chapitre, nous proposons une étude de corpus portant sur les propriétés lexico-sémantiques des verbes CAUSE, MAKE, HAVE, GET et LET. Sur la base des donnés empiriques de notre étude de corpus, nous présentons, dans le dernier chapitre, une nouvelle typologie sémantique pour les verbes causatifs analytiques anglais
This thesis deals with the semantics of causative constructions. It develops a semantic typology for English periphrastic causative verbs CAUSE, MAKE, HAVE, GET, and LET, based on the force-dynamics model. The first chapter aims to capture the essence of the notion of causation from an epistemic as well as a linguistic viewpoints. The second chapter sets the theoretical framework, which is cognitive semantics. The third chapter offers a discussion of some of the most commonly shared hypotheses about the semantics of English periphrastic causative verbs in literature. In the fourth chapter, we propose a corpus study of the lexico-semantic features of the verbs CAUSE, MAKE, HAVE, GET and LET. The last chapter presents a newly semantic typology for English periphrastic causative verbs, drawn upon the data we collected from our corpus study

Ce travail doctoral s'intéresse aux effets des processus de catégorisations sociales à l'œuvre dans un marketing politique multiculturel. Cette recherche s'intéresse particulièrement au ciblage des électeurs binationaux Français et issus de l'un des pays du Maghreb et la manière avec laquelle il affecte les dynamiques identitaires des électeurs binationaux et leurs conceptions de la citoyenneté. Nous distinguons quatre dynamiques catégorielles : de repli identitaire, de probation, adaptative et d'ancrage. Les individus vont intégrer, selon les contextes et les discours politiques auxquels ils sont confrontés, l'une de ces quatre dynamiques. Plusieurs rapports à la citoyenneté ont été identifiés, démontrant tous un besoin incessant de citoyenneté des individus. Les pratiques de marketing politique jouent dans ce cadre un rôle important dans la fabrique de catégories, faisant ainsi l'objet de plusieurs critiques. Ce travail de recherche contribue aux théories de l'identité sociale, au marketing politique et aux études marketing sur l'ethnicité
This doctoral work focuses on the effects of social categorization processes in multicultural political marketing. This research specifically examines the targeting of French dual citizens of North African descent and how it affects the identity dynamics of these dual citizens and their perceptions of citizenship. Four categorical dynamics are distinguished: identity withdrawal, probationary, adaptive, and anchoring. Individuals will adopt one of these four dynamics depending on the contexts and political discourses they encounter. Several forms of citizenship relationships have been identified, all demonstrating an ongoing need for citizenship among individuals. Political marketing practices play a significant role in shaping categories within this framework, thus subject to various criticisms. This research contributes to social identity theories, political marketing, and marketing studies on ethnicity

Ontologies provide an organization of language, in the form of a network or graph, which is amenable to computational analysis while remaining human-readable. Although they are used in a variety of disciplines, ontologies in the biomedical field, such as Gene Ontology, are of interest for their role in organizing terminology used to describe—among other concepts—the functions, locations, and processes of genes and gene-products. Due to the consistency and level of automation that ontologies provide for such annotations, methods for finding enriched biological terminology from a set of differentially identified genes in a tissue or cell sample have been developed to aid in the elucidation of disease pathology and unknown biochemical pathways. However, despite their immense utility, biomedical ontologies have significant limitations and caveats. One major issue is that gene annotation enrichment analyses often result in many redundant, individually enriched ontological terms that are highly specific and weakly justified by statistical significance. These large sets of weakly enriched terms are difficult to interpret without manually sorting into appropriate functional or descriptive categories. Also, relationships that organize the terminology within these ontologies do not contain descriptions of semantic scoping or scaling among terms. Therefore, there exists some ambiguity, which complicates the automation of categorizing terms to improve interpretability. We emphasize that existing methods enable the danger of producing incorrect mappings to categories as a result of these ambiguities, unless simplified and incomplete versions of these ontologies are used which omit problematic relations. Such ambiguities could have a significant impact on term categorization, as we have calculated upper boundary estimates of potential false categorizations as high as 121,579 for the misinterpretation of a single scoping relation, has_part, which accounts for approximately 18% of the total possible mappings between terms in the Gene Ontology. However, the omission of problematic relationships results in a significant loss of retrievable information. In the Gene Ontology, this accounts for a 6% reduction for the omission of a single relation. However, this percentage should increase drastically when considering all relations in an ontology. To address these issues, we have developed methods which categorize individual ontology terms into broad, biologically-related concepts to improve the interpretability and statistical significance of gene-annotation enrichment studies, meanwhile addressing the lack of semantic scoping and scaling descriptions among ontological relationships so that annotation enrichment analyses can be performed across a more complete representation of the ontological graph. We show that, when compared to similar term categorization methods, our method produces categorizations that match hand-curated ones with similar or better accuracy, while not requiring the user to compile lists of individual ontology term IDs. Furthermore, our handling of problematic relations produces a more complete representation of ontological information from a scoping perspective, and we demonstrate instances where medically-relevant terms--and by extension putative gene targets--are identified in our annotation enrichment results that would be otherwise missed when using traditional methods. Additionally, we observed a marginal, yet consistent improvement of statistical power in enrichment results when our methods were used, compared to traditional enrichment analyses that utilize ontological ancestors. Finally, using scalable and reproducible data workflow pipelines, we have applied our methods to several genomic, transcriptomic, and proteomic collaborative projects.

A document organization is a collection of documents composed of labeled clusters that contain similar documents. In any information system, a collection of documents always changes as time goes, since users access the collection to delete, add, and update documents. Dynamic Document Organization is a document organization that adapts automatically to such variable document collections. DDO poses two challenges, because of the decentralized mode of access. First, some clusters may have many documents, while others may have very few. Second, documents belonging to new topics may be added to the information system very often. Considering these two points, we need to reorganize the collection of documents, even if it was organized previously. Both text categorization and text clustering are limited when implementing DDO (Dynamic Document Organization) individually. Text categorization requires the manual preliminary tasks of the predefinition of a classification system and the preparation of sample labeled documents. Text clustering generates only unnamed clusters alone; each cluster should be labeled, manually by scanning contained documents. Therefore, this dissertation proposes approaches to the implementation of DDO that combined text clustering, cluster identification, and text categorization.

Ce travail de recherche a pour but d'investiguer du point de vue psychosocial l'identité " raciale " de la population des Noirs en France en particulier dans leur rapport avec les Blancs afin de mettre au jour les dynamiques identitaires qui existent entre les groupes.Le cadre théorique relève de la catégorisation sociale et plus particulièrement de l'un de ses développements, à savoir le concept des partitions sociales. Cette conception postule l'existence de plusieurs types de relations intercatégorielles (les partitions) qui permettent d'appréhender la nature du positionnement identitaire des individus dans le cadre d'une relation intergroupe (en fonction de l'intérêt identitaire de chacun) en y ajoutant un aspect dynamique à travers la notion de co-construction. La méthodologie utilisée relève principalement des représentations mutuelles, c'est-à-dire des représentations intercatégorielles (la représentation que l'on a de sa catégorie d'appartenance et celle que l'on a de l'exogroupe considéré) qui nous permettent de dégager les dynamiques identitaires existant entre des groupes socio-culturellement différents. Dans cette perspective, une série de d'études a été mise en œuvre. La première a mis en évidence l'existence de stratégies identitaires différentes de la part des Noirs antillais selon qu'ils vivent en Métropole ou aux Antilles. Les Antillais de Métropole s'insèrent dans une partition communautaire alors que ceux des Antilles semblent adopter une stratégie d'évitement en se recatégorisant à un niveau supraordonné. La deuxième étude a montré qu'en fonction de leur interlocuteur (opérationnalisé par les mots inducteurs " Black ", " Blanc " et " Beur "), les Français d'origine Afro-antillaise, Européenne ou Maghrébine adoptaient des positionnements identitaires différents. Enfin, une troisième étude a permis de dégager l'existence d'une discrimination implicite vis-à-vis des Noirs en France s'appuyant sur un stéréotype structuré de la même façon qu'aux États-Unis mais ne s'exprimant pas explicitement. De façon générale, ces études semblent montrer que l'appréhension des relations interethniques par le biais des partitions (et non plus en simples termes d'endofavoritisme ou d'exodéfavoritisme) permet de mieux comprendre les relations existant entre différentes communautés dans un contexte intra-national.

Queue buildup is a problem that every fast food restaurant would like to reduce, as its effect on customer satisfaction and restaurant revenue is obviously negative. This thesis looks at how a dynamic menu management system would affect fast food restaurants, specifically how the promotion of fast products, in certain situations, would affect total revenue and queue lengths. A neural network was used to categorize products depending on their preparation time and to predict queue waiting times. A simulation approach was used to find out how the menu management system would affect the restaurant. The results are that in situations where long queues are formed, the promotion of fast products increases revenue by 7.8% and decreases queue lengths by 13.3%. In situations where little to no queues are formed, the same system still decreases queue lengths by 4.7%, but also decreases revenue by 0.2%

Cette thèse aborde la question du Changement Catégoriel dans la Perception et propose une approche Dynamique qui s'appuie sur la Théorie des Catastrophes et le modèle de la FRONCE. Dans nos travaux, nous assimilons les catégories perceptives à des attracteurs d'un système dynamique non linéaire. Si notre postulat est juste, alors, un ensemble de phénomènes non linéaires théoriquement associés doivent pouvoir être observé conjointement dans un changement catégoriel qui pourrait alors être considéré comme une transition de phase. Nous avons porté notre attention sur le phénomène d'Hystérèse, déjà abondamment exploré en perception, ainsi que sur les phénomènes de divergence et de ralentissement critique. Au travers d'une série de sept expérimentations, nous confrontons l'ensemble de ces hypothèses. Nous avons confirmé la présence de ces trois phénomènes lors d'un changement catégoriel. Les résultats montrent que l'hystérèse, la divergence et le ralentissement critique sont des phénomènes observés conjointement lors d'un changement de catégorie, mais qu'ils sont aussi quantitativement reliés, comme permettait de le prédire le modèle de la fronce issue de la théorie des catastrophes. D'une manière générale, nous avons conclu que les catégories perceptives étaient fortement similaires à des d'attracteurs et que l'étude et la modélisation des changements dans la perception devraient présenter un aspect nécessairement dynamique.

no
In three experiments we investigated the dynamics of impression formation when perceivers encounter unsurprising (e.g. male mechanic) versus surprising (e.g. female mechanic) social category conjunctions. In Experiment 1, participants took longer to form an impression of targets described using a surprising versus an unsurprising conjunction of categorizations. In Experiment 2, we investigated the stages during which impressions of category conjunctions are formed. While unsurprising category combinations were characterized with reference to ‘constituent’ stereotypic traits, surprising combinations were characterized initially by stereotypic traits but later by ‘emergent’ impressions. In Experiment 3, we investigated motivational states that drive the dynamics of category conjunction. We found that higher Personal Need for Structure (PNS) predicted the use of more emergent and fewer constituent attributes in the impressions formed of surprising combinations. Across all three experiments, more ‘causal attributes’ were used in descriptions of the surprising combination. We discuss the implications of these findings for developing a model of the dynamics and composition of social category conjunctions.

碩士
國立中正大學
電機工程研究所
92
Adaptive Resonance Theory (ART) is a kind of unsupervised learning network pattern with advantages of stability and plasticity, which mean that network can keep old data properly and learn new ones quickly in the mean time. The objective of this thesis is to implement an ART-based chip by Flexible Adaptable-Size Topology (FAST). The chip designed in Verilog-HDL with XCV800 FPGA plays a role as a categorization controller. The main architecture of which includes three parts. They are "Similarity", "Learning" and "Pruning", respectively. The architecture updates location and size of receptive field dynamically, makes categorization boundaries variable, and prunes neurons with overlapped receptive field to save space. Finally, the categorized results computed by FPGA are sent to PC-simulated AHC which controls a PC-simulated plant, cart-pole system, and tries to make it balanced.

碩士
國立中正大學
心理學研究所
99
For the needs of management, Chinese supervisors tend to divided subordinates into in-group member (zi ji ren) and out-group member (wai ren), and to give different types of treatment. Although there are many research exist, but the distinction criteria between in-group member and out-group member, or how Chinese supervisors to show different treatment, even more, the dynamic of in-group and out-group member, still lack of study to clarify these issues at present. To fill these gaps, in this study, we used in-depth interview to collected data from organizations in Taiwan. The data collected from 7 work group samples which consist of 7 supervisors and 11 subordinates shows that: (1) Supervisors will treat in-group members with “rewarding promotion”, “highly hopes”, “caring concern”, “friendly relationship”, “trustful confident”, and “empowering support”, but treat out-group members with “rigorous” and “neglected segregation”. (2) “Pre-existing relationship”, “personal relationship”, “congenial”, “aggressive active”, “cautious and conscientious”, “unconditionally obedient”, “try to figure out the supervisor’s meaning”, and “highly ability” are in-group member’s characteristics; but “a mutually respectful relationship”, “superficial and perfunctory”, and “get one’s own way” may be out-group member’s characteristics. (3) In the course of the out-group member to the in-group member, employees will “coordinate with supervisor’s requirements”, “working hard”, or “energetic behavior in crisis” and in-group member may play an important role in this course. (4) But in the course of the in-group member to the out-group member, it will start from supervisor, and because of in-group members “unable to meet the expectations of supervisor”, “supervisor and subordinate go their own way”, or “to go against supversior’s trust of subordinate” results in end up with alienation. Finally, based on these results, contributions and limitations are discussed, and suggestions are provided for future studies and managerial practices in Chinese organizations.

Radio frequency identification (RFID) is a technology used in many applications for the identification of objects. This thesis presents a concept of how libraries could use RFID technology to locate physical items within the library. The ability to locate items within the library changes the way users interact with physical material, creates new ways of user collaboration, and influences the ability to browse the shelves for physical items. Several implementation scenarios are presented in detail and implications on collaboration and browsing are analyzed.

Concepts are widely agreed to be the basic constituents of thought. Amongst philosophers and psychologists, however, the question of how concepts are structured has been a longstanding problem and a locus of disagreement. I draw on recent work describing how representational content is ascribed to populations of neurons to develop a novel solution to this problem. Because disputes over the structure of concepts often reflect divergent explanatory goals, I begin by arguing for a set of six criteria that a good theory ought to accommodate. These criteria address philosophical concerns related to content, reference, scope, publicity, and compositionality, and psychological concerns related to categorization phenomena and neural plausibility. Next, I evaluate a number of existing theoretical approaches in relation to these six criteria. I consider classical views that identify concepts with definitions, similarity-based views that identify concepts with prototypes or exemplars, theory-based views that identify concepts with explanatory schemas, and atomistic views that identify concepts with unstructured mental symbols that enter into law-like relations with their referents. I conclude that none of these accounts can satisfactorily accommodate all of the criteria. I then describe the theory of representational content that I employ to motivate a novel account of concept structure. I briefly defend this theory against competitors, and I describe how it can be scaled from the level of basic perceptual representations to the level of highly complex conceptual representations. On the basis of this description, I contend that concepts are structured dynamically through sets of transformations of single source representation, and that the content of a given concept specifies the set of potential transformations it can enter into. I conclude by demonstrating that the ability of this account to meet all of the criteria introduced beforehand. I consider objections to my views throughout.

The purpose of this thesis was to apply mainstream theoretical social psychology to group dynamics in isolated confined extreme (ICE) environments, with a particular focus on the context of long duration spaceflight. This was done by providing a thorough review of both the ICE environment psychology literature and the social psychological literature. From this it emerged that significant gaps existed within the ICE literature around the understanding of group processes. A review of relevant social psychology literature suggested that the social identity perspective may have much to offer in advancing the understanding of group processes in these environments. It was argued that processes related to social identification plays a key role in the team functionality. Of particular note was how social ‘norms’ could influence social identification processes and group functioning. Crew heterogeneity had been cited as a potential concern in much ICE literature, so group norms that related to diversity were investigated. Three theoretical chapters provided a review of space and ICE environment psychology literature, social psychology literature and diversity literature. Five experiments are reported in this thesis. The first, an exploratory field study which study took place during an expedition into the Australian Outback, made general predictions about how social identification might be constrained by the extreme environment context. The study demonstrated that identity processes did indeed influence group functionality, but also that the presence of pro-diversity group norms seemed to have a positive influence. In order to explore this issue in more detail, four studies were conducted. These included a field study at the Mars Desert Research station in Utah, two laboratory studies and a final field study that took place during a winter-over period at the Concordia Research station in Antarctica. Collectively these studies demonstrated that ingroup identification played an important role in driving group functionality in extreme environments. In addition they demonstrated that the presence of a pro-diversity group norm had the potential to increase group identification while also reducing the likelihood of negative outcomes (crew alienation and sub-group conflict). Multilevel Linear Modelling and Social Network Analysis were used in two of the studies and if adopted have the potential to significantly improve the validity of future extreme environment research. By using these techniques in this thesis, the certainty surrounding the research findings was greatly enhanced. Taken together the work of this thesis demonstrated three key points: that mainstream social psychology theory and research has much to offer extreme environment research.; social identification processes play a key role in the group dynamics of people in extreme environments; and that a pro-diversity norm, under certain conditions, can play an important role in maximising group functioning. The contribution of this thesis, then, has implications for the ICE literature as well as the social/organisational literature on group diversity and finally the kind of group culture that should be encouraged in extreme environments.