Dissertations / Theses on the topic 'Probabilistic representation'
To see the other types of publications on this topic, follow the link: Probabilistic representation.
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'Probabilistic representation.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
1
Helmkay, Owen. "Information representation, problem format, and mental algorithms in probabilistic reasoning." Thesis, National Library of Canada = Bibliothèque nationale du Canada, 2001. http://www.collectionscanada.ca/obj/s4/f2/dsk3/ftp04/NQ66153.pdf.
Full text
2
Tarrago, Pierre. "Non-commutative generalization of some probabilistic results from representation theory." Thesis, Paris Est, 2015. http://www.theses.fr/2015PESC1123/document.
Full textAbstract:
Le sujet de cette thèse est la généralisation non-commutative de résultats probabilistes venant de la théorie des représentations. Les résultats obtenus se divisent en trois parties distinctes. Dans la première partie de la thèse, le concept de groupe quantique easy est étendu au cas unitaire. Tout d'abord, nous donnons une classification de l'ensemble des groupes quantiques easy unitaires dans le cas libre et classique. Nous étendons ensuite les résultats probabilistes de au cas unitaire. La deuxième partie de la thèse est consacrée à une étude du produit en couronne libre. Dans un premier temps, nous décrivons les entrelaceurs des représentations dans le cas particulier d'un produit en couronne libre avec le groupe symétrique libre: cette description permet également d'obtenir plusieurs résultats probabilistes. Dans un deuxième temps, nous établissons un lien entre le produit en couronne libre et les algèbres planaires: ce lien mène à une preuve d'une conjecture de Banica et Bichon. Dans la troisième partie de la thèse, nous étudions un analoque du graphe de Young qui encode la structure multiplicative des fonctions fondamentales quasi-symétriques. La frontière minimale de ce graphe a déjà été décrite par Gnedin et Olshanski. Nous prouvons que la frontière minimale coïncide avec la frontière de Martin. Au cours de cette preuve, nous montrons plusieurs résultats combinatoires asymptotiques concernant les diagrammes de Young en rubanThe subject of this thesis is the non-commutative generalization of some probabilistic results that occur in representation theory. The results of the thesis are divided into three different parts. In the first part of the thesis, we classify all unitary easy quantum groups whose intertwiner spaces are described by non-crossing partitions, and develop the Weingarten calculus on these quantum groups. As an application of the previous work, we recover the results of Diaconis and Shahshahani on the unitary group and extend those results to the free unitary group. In the second part of the thesis, we study the free wreath product. First, we study the free wreath product with the free symmetric group by giving a description of the intertwiner spaces: several probabilistic results are deduced from this description. Then, we relate the intertwiner spaces of a free wreath product with the free product of planar algebras, an object which has been defined by Bisch and Jones. This relation allows us to prove the conjecture of Banica and Bichon. In the last part of the thesis, we prove that the minimal and the Martin boundaries of a graph introduced by Gnedin and Olshanski are the same. In order to prove this, we give some precise estimates on the uniform standard filling of a large ribbon Young diagram. This yields several asymptotic results on the filling of large ribbon Young diagrams
3
Shen, Amelia H. (Amelia Huimin). "Probabilistic representation and manipulation of Boolean functions using free Boolean diagrams." Thesis, Massachusetts Institute of Technology, 1994. http://hdl.handle.net/1721.1/34087.
Full textAbstract:
Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1994.Includes bibliographical references (p. 145-149).
by Amelia Huimin Shen.
Ph.D.
4
Lloyd, James Robert. "Representation, learning, description and criticism of probabilistic models with applications to networks, functions and relational data." Thesis, University of Cambridge, 2015. https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.709264.
Full text
5
Vasudevan, Shrihari. "Spatial cognition for mobile robots : a hierarchical probabilistic concept-oriented representation of space." Zürich : ETH, 2008. http://e-collection.ethbib.ethz.ch/show?type=diss&nr=17612.
Full text
6
Lavis, Benjamin Mark Mechanical & Manufacturing Engineering Faculty of Engineering UNSW. "Spatially reconfigurable and non-parametric representation of dynamic bayesian beliefs." Publisher:University of New South Wales. Mechanical & Manufacturing Engineering, 2008. http://handle.unsw.edu.au/1959.4/41468.
Full textAbstract:
This thesis presents a means for representing and computing beliefs in the form of arbitrary probability density functions with a guarantee for the ongoing validity of such beliefs over indefinte time frames. The foremost aspect of this proposal is the introduction of a general, theoretical, solution to the guaranteed state estimation problem from within the recursive Bayesian estimation framework. The solution presented here determines the minimum space required, at each stage of the estimation process, to represent the belief with limited, or no, loss of information. Beyond this purely theoretical aspect, a number of numerical techniques, capable of determining the required space and performing the appropriate spatial reconfiguration, whilst also computing and representing the belief functions, are developed. This includes a new, hybrid particle-element approach to recursive Bayesian estimation. The advantage of spatial reconfiguration as presented here is that it ensures that the belief functions consider all plausible states of the target system, without altering the recursive Bayesian estimation equations used to form those beliefs. Furthermore, spatial reconfiguration as proposed in this dissertation enhances the estimation process since it allows computational resources to be concentrated on only those states considered plausible. Autonomous maritime search and rescue is used as a focus application throughout this dissertation since the searching-and-tracking requirements of the problem involve uncertainty, the use of arbitrary belief functions and dynamic target systems. Nevertheless, the theoretical development in this dissertation has been kept general and independent of an application, and as such the theory and techniques presented here may be applied to any problem involving dynamic Bayesian beliefs. A number of numerical experiments and simulations show the efficacy of the proposed spatially reconfigurable representations, not only in ensuring the validity of the belief functions over indefinite time frames, but also in reducing computation time and improving the accuracy of function approximation. Improvements of an order of magnitude were achieved when compared with traditional, spatially static representations.
7
Geilke, Michael [Verfasser]. "Online density estimates : a probabilistic condensed representation of data for knowledge discovery / Michael Geilke." Mainz : Universitätsbibliothek Mainz, 2017. http://d-nb.info/1147611165/34.
Full text
8
Zanitti, Gaston Ezequiel. "Development of a probabilistic domain-specific language for brain connectivity including heterogeneous knowledge representation." Electronic Thesis or Diss., université Paris-Saclay, 2023. http://www.theses.fr/2023UPASG022.
Full textAbstract:
Grâce aux récents progrès technologiques, le chercheur en neurosciences dispose d'une quantité croissante de jeux de données pour étudier le cerveau. La multiplicité des travaux dédiés a également produit des ontologies encodant des connaissances à la pointe concernant les différentes aires, les schémas d'activation, les mots-clés associés aux études, etc. Il existe d'autre part une incertitude inhérente aux images cérébrales, du fait de la mise en correspondance entre voxels - ou pixels 3D - et points réels sur le cerveau de différents sujets. Malheureusement, à ce jour, aucun cadre unifié ne permet l'accès à cette mine de données hétérogènes avec l'incertitude associée, obligeant le chercheur à se tourner vers des outils ad hoc. Dans cette étude, nous présentons NeuroLang, un langage probabiliste basé sur de la logique de premier ordre, comprenant des règles existentielles, de l'incertitude probabiliste, l'intégration d'ontologies reposant sur l'hypothèse du monde ouvert, ainsi que des mécanismes garantissant une réponse aux requêtes résolvables, même sur de très grandes bases de données. Nous soutenons que NeuroLang, par l'expressivité de son langage de requête, contribuera à grandement améliorer la recherche en neurosciences, en donnant notamment la possibilité d'intégrer de manière transparente des données hétérogènes, telles que des ontologies avec des atlas cérébraux probabilistes. Dans ce cas-ci, des domaines cognitifs - à la granularité fine - et des régions cérébrales seront associés via un ensemble de critères formels, favorisant ainsi la communication et la reproductibilité des résultats d'études sur les fonctions cérébrales. Aussi croyons-nous que NeuroLang est à même de se positionner en tête sur ces approches numériques qui visent à formaliser la recherche neuroscientifique à grande échelle via la programmation probabiliste et logique du premier ordreResearchers in neuroscience have a growing number of datasets available to study the brain, which is made possible by recent technological advances. Given the extent to which the brain has been studied, there is also available ontological knowledge encoding the current state of the art regarding its different areas, activation patterns, keywords associated with studies, etc. Furthermore, there is inherent uncertainty associated with brain scans arising from the mapping between voxels -3D pixels- and actual points in different individual brains. Unfortunately, there is currently no unifying framework for accessing such collections of rich heterogeneous data under uncertainty, making it necessary for researchers to rely on ad hoc tools. In this work we introduce NeuroLang, a probabilistic language based on first-order logic with existential rules, probabilistic uncertainty, ontologies integration under the open world assumption, and built-in mechanisms to guarantee tractable query answering over very large datasets. We propose that NeuroLang provides a substantial improvement to cognitive neuroscience research through the expressive power of its query language. We can leverage the ability of NeuroLang to seamlessly integrate useful heterogeneous data, such as ontologies and probabilistic brain atlases, to map fine-grained cognitive domains to brain regions through a set of formal criteria, promoting shareable and highly reproducible research on the domains of brain function. We believe that NeuroLang is well suited for leading computational approaches to formalize large-scale neuroscience research through probabilistic first-order logic programming
9
Tarrago, Pierre [Verfasser], and Roland [Akademischer Betreuer] Speicher. "Non-commutative generalization of some probabilistic results from representation theory / Pierre Tarrago. Betreuer: Roland Speicher." Saarbrücken : Saarländische Universitäts- und Landesbibliothek, 2015. http://d-nb.info/1079840249/34.
Full text
10
Nayak, Sunita. "Representation and learning for sign language recognition." [Tampa, Fla] : University of South Florida, 2008. http://purl.fcla.edu/usf/dc/et/SFE0002362.
Full text
11
El-Shaer, Mennat Allah. "An Experimental Evaluation of Probabilistic Deep Networks for Real-time Traffic Scene Representation using Graphical Processing Units." The Ohio State University, 2019. http://rave.ohiolink.edu/etdc/view?acc_num=osu1546539166677894.
Full text
12
Paraschos, Alexandros [Verfasser], Jan [Akademischer Betreuer] Peters, Gerhard [Akademischer Betreuer] Neumann, and Sylvain [Akademischer Betreuer] Calinon. "Robot Skill Representation, Learning and Control with Probabilistic Movement Primitives / Alexandros Paraschos ; Jan Peters, Gerhard Neumann, Sylvain Calinon." Darmstadt : Universitäts- und Landesbibliothek Darmstadt, 2017. http://d-nb.info/1147968381/34.
Full text
13
Ogul, Hasan. "Computational Representation Of Protein Sequences For Homology Detection And Classification." Phd thesis, METU, 2006. http://etd.lib.metu.edu.tr/upload/12606997/index.pdf.
Full textAbstract:
Machine learning techniques have been widely used for classification problems in computational biology. They require that the input must be a collection of fixedlength feature vectors. Since proteins are of varying lengths, there is a need for a
means of representing protein sequences by a fixed-number of features. This thesis
introduces three novel methods for this purpose: n-peptide compositions with
reduced alphabets, pairwise similarity scores by maximal unique matches, and
pairwise similarity scores by probabilistic suffix trees.
New sequence representations described in the thesis are applied on three
challenging problems of computational biology: remote homology detection,
subcellular localization prediction, and solvent accessibility prediction, with some
problem-specific modifications. Rigorous experiments are conducted on common
benchmarking datasets, and a comparative analysis is performed between the new
methods and the existing ones for each problem.
On remote homology detection tests, all three methods achieve competitive
accuracies with the state-of-the-art methods, while being much more efficient. A
combination of new representations are used to devise a hybrid system, called
PredLOC, for predicting subcellular localization of proteins and it is tested on two
distinct eukaryotic datasets. To the best of author&rsquos knowledge, the accuracy achieved by PredLOC is the highest one ever reported on those datasets. The maximal unique match method is resulted with only a slight improvement in solvent accessibility predictions.
14
Ramos, Fabio Tozeto. "Recognising, Representing and Mapping Natural Features in Unstructured Environments." Australian Centre for Field Robotics, Department of Aerospace, Mechanical and Mechatronic Engineering, 2008. http://hdl.handle.net/2123/2322.
Full textAbstract:
Doctor of Philosophy (PhD)This thesis addresses the problem of building statistical models for multi-sensor perception in unstructured outdoor environments. The perception problem is divided into three distinct tasks: recognition, representation and association. Recognition is cast as a statistical classification problem where inputs are images or a combination of images and ranging information. Given the complexity and variability of natural environments, this thesis investigates the use of Bayesian statistics and supervised dimensionality reduction to incorporate prior information and fuse sensory data. A compact probabilistic representation of natural objects is essential for many problems in field robotics. This thesis presents techniques for combining non-linear dimensionality reduction with parametric learning through Expectation Maximisation to build general representations of natural features. Once created these models need to be rapidly processed to account for incoming information. To this end, techniques for efficient probabilistic inference are proposed. The robustness of localisation and mapping algorithms is directly related to reliable data association. Conventional algorithms employ only geometric information which can become inconsistent for large trajectories. A new data association algorithm incorporating visual and geometric information is proposed to improve the reliability of this task. The method uses a compact probabilistic representation of objects to fuse visual and geometric information for the association decision. The main contributions of this thesis are: 1) a stochastic representation of objects through non-linear dimensionality reduction; 2) a landmark recognition system using a visual and ranging sensors; 3) a data association algorithm combining appearance and position properties; 4) a real-time algorithm for detection and segmentation of natural objects from few training images and 5) a real-time place recognition system combining dimensionality reduction and Bayesian learning. The theoretical contributions of this thesis are demonstrated with a series of experiments in unstructured environments. In particular, the combination of recognition, representation and association algorithms is applied to the Simultaneous Localisation and Mapping problem (SLAM) to close large loops in outdoor trajectories, proving the benefits of the proposed methodology.
15
Ramos, Fabio Tozeto. "Recognising, Representing and Mapping Natural Features in Unstructured Environments." Thesis, The University of Sydney, 2007. http://hdl.handle.net/2123/2322.
Full textAbstract:
This thesis addresses the problem of building statistical models for multi-sensor perception in unstructured outdoor environments. The perception problem is divided into three distinct tasks: recognition, representation and association. Recognition is cast as a statistical classification problem where inputs are images or a combination of images and ranging information. Given the complexity and variability of natural environments, this thesis investigates the use of Bayesian statistics and supervised dimensionality reduction to incorporate prior information and fuse sensory data. A compact probabilistic representation of natural objects is essential for many problems in field robotics. This thesis presents techniques for combining non-linear dimensionality reduction with parametric learning through Expectation Maximisation to build general representations of natural features. Once created these models need to be rapidly processed to account for incoming information. To this end, techniques for efficient probabilistic inference are proposed. The robustness of localisation and mapping algorithms is directly related to reliable data association. Conventional algorithms employ only geometric information which can become inconsistent for large trajectories. A new data association algorithm incorporating visual and geometric information is proposed to improve the reliability of this task. The method uses a compact probabilistic representation of objects to fuse visual and geometric information for the association decision. The main contributions of this thesis are: 1) a stochastic representation of objects through non-linear dimensionality reduction; 2) a landmark recognition system using a visual and ranging sensors; 3) a data association algorithm combining appearance and position properties; 4) a real-time algorithm for detection and segmentation of natural objects from few training images and 5) a real-time place recognition system combining dimensionality reduction and Bayesian learning. The theoretical contributions of this thesis are demonstrated with a series of experiments in unstructured environments. In particular, the combination of recognition, representation and association algorithms is applied to the Simultaneous Localisation and Mapping problem (SLAM) to close large loops in outdoor trajectories, proving the benefits of the proposed methodology.
16
Stenson, Matthew P. "Analysis of higher order terms in the Gram-Charlier type a representation of equivalent load used in probabilistic simulation of electric power systems." Ohio University / OhioLINK, 1987. http://rave.ohiolink.edu/etdc/view?acc_num=ohiou1183062589.
Full text
17
GARBARINO, DAVIDE. "Acknowledging the structured nature of real-world data with graphs embeddings and probabilistic inference methods." Doctoral thesis, Università degli studi di Genova, 2022. http://hdl.handle.net/11567/1092453.
Full textAbstract:
In the artificial intelligence community there is a growing consensus that real
world data is naturally represented as graphs because they can easily incorporate complexity at several levels, e.g. hierarchies or time dependencies. In this context, this thesis studies two main branches for structured data.
In the first part we explore how state-of-the-art machine learning methods can be extended to graph modeled data provided that one is able to represent graphs in vector spaces. Such extensions can be applied to analyze several kinds of real-world data and tackle different problems. Here we study the
following problems: a) understand the relational nature and evolution of websites which belong to different categories (e-commerce, academic (p.a.) and encyclopedic (forum)); b) model tennis players scores based on different game surfaces and tournaments in order to predict matches results; c) analyze preter-
m-infants motion patterns able to characterize possible neuro degenerative disorders and d) build an academic collaboration recommender system able to model academic groups and individual research interest while suggesting possible researchers to connect with, topics of interest and representative publications to external users.
In the second part we focus on graphs inference methods from data which present two main challenges: missing data and non-stationary time dependency. In particular, we study the problem of inferring Gaussian Graphical
Models in the following settings: a) inference of Gaussian Graphical Models when data are missing or latent in the context of multiclass or temporal network inference and b) inference of time-varying Gaussian Graphical Models
when data is multivariate and non-stationary. Such methods have a natural application in the composition of an optimized stock markets portfolio. Overall this work sheds light on how to acknowledge the intrinsic structure
of data with the aim of building statistical models that are able to capture the actual complexity of the real world.
18
Schustek, Philipp. "Probabilistic models for human judgments about uncertainty in intuitive inference tasks." Doctoral thesis, Universitat Pompeu Fabra, 2018. http://hdl.handle.net/10803/586057.
Full textAbstract:
Updating beliefs to maintain coherence with observational evidence is a cornerstone of rationality. This entails the compliance with probabilistic principles which acknowledge that real-world observations are consistent with several possible interpretations. This work presents two novel experimental paradigms and computational analyses of how human participants quantify uncertainty in perceptual inference tasks. Their behavioral responses feature non-trivial patterns of probabilistic inference such as reliability-based belief updating over hierarchical state representations of the environment. Despite characteristic generalization biases, behavior cannot be explained well by alternative heuristic accounts. These results suggest that uncertainty is an integral part of our inferences and that we indeed have the potential to resort to rational inference mechanisms that adhere to probabilistic principles. Furthermore, they appear consistent with ubiquitous representations of uncertainty posited by framework theories such as Bayesian hierarchical modeling and predictive coding.Un pilar fundamental de la racionalidad es actualizar las creencias con la finalidad de mantener la coherencia con la evidencia observacional. Esto implica cumplir con principios probabilísticos, los cuales reconocen que las observaciones del mundo real son consistentes con varias interpretaciones posibles. Este estudio presenta dos novedosas pruebas experimentales, así como análisis computacionales, de cómo participantes humanos cuantifican la incertidumbre en tareas de inferencia perceptiva. Sus respuestas conductuales muestran patrones no triviales de inferencia probabilística, tales como la actualización de creencias basadas en la confiabilidad sobre las representaciones jerárquicas del estado del entorno. A pesar de los sesgos característicos de generalización, el comportamiento no puede ser correctamente explicado con descripciones heurísticas alternativas. Estos resultados sugieren que la incertidumbre es una parte integral de nuestras inferencias y que efectivamente tenemos el potencial para recurrir a mecanismos de inferencia racional, los cuales adhieren a principios probabilísticos. Además, dichos resultados son compatibles con la idea de que representaciones de incertidumbre internas son ubicuas, lo cual presuponen teorías generales como Bayesian hierarchical modeling y predictive coding.
19
Lee, Wooyoung. "Learning Statistical Features of Scene Images." Research Showcase @ CMU, 2014. http://repository.cmu.edu/dissertations/540.
Full textAbstract:
Scene perception is a fundamental aspect of vision. Humans are capable of analyzing behaviorally-relevant scene properties such as spatial layouts or scene categories very quickly, even from low resolution versions of scenes. Although humans perform these tasks effortlessly, they are very challenging for machines. Developing methods that well capture the properties of the representation used by the visual system will be useful for building computational models that are more consistent with perception. While it is common to use hand-engineered features that extract information from predefined dimensions, they require careful tuning of parameters and do not generalize well to other tasks or larger datasets. This thesis is driven by the hypothesis that the perceptual representations are adapted to the statistical properties of natural visual scenes. For developing statistical features for global-scale structures (low spatial frequency information that encompasses entire scenes), I propose to train hierarchical probabilistic models on whole scene images. I first investigate statistical clusters of scene images by training a mixture model under the assumption that each image can be decoded by sparse and independent coefficients. Each cluster discovered by the unsupervised classifier is consistent with the high-level semantic categories (such as indoor, outdoor-natural and outdoor-manmade) as well as perceptual layout properties (mean depth, openness and perspective). To address the limitation of mixture models in their assumptions of a discrete number of underlying clusters, I further investigate a continuous representation for the distributions of whole scenes. The model parameters optimized for natural visual scenes reveal a compact representation that encodes their global-scale structures. I develop a probabilistic similarity measure based on the model and demonstrate its consistency with the perceptual similarities. Lastly, to learn the representations that better encode the manifold structures in general high-dimensional image space, I develop the image normalization process to find a set of canonical images that anchors the probabilistic distributions around the real data manifolds. The canonical images are employed as the centers of the conditional multivariate Gaussian distributions. This approach allows to learn more detailed structures of the local manifolds resulting in improved representation of the high level properties of scene images.
20
Chrastansky, Alena [Verfasser], and Hans Von [Akademischer Betreuer] Storch. "Multi-decadal reconstruction and probabilistic representation of weather-related variability in North Sea coast chronic oil pollution / Alena Chrastansky. Betreuer: Hans von Storch." Hamburg : Staats- und Universitätsbibliothek Hamburg, 2011. http://d-nb.info/102042236X/34.
Full text
21
Nyga, Daniel [Verfasser], Michael [Akademischer Betreuer] [Gutachter] Beetz, and Anthony G. [Gutachter] Cohn. "Interpretation of Natural-language Robot Instructions: Probabilistic Knowledge Representation, Learning, and Reasoning / Daniel Nyga ; Gutachter: Michael Beetz, Anthony G. Cohn ; Betreuer: Michael Beetz." Bremen : Staats- und Universitätsbibliothek Bremen, 2017. http://d-nb.info/1132756944/34.
Full text
22
Yan, Chang. "Neural Representation of Working Memory Contents at Different Levels of Abstraction." Doctoral thesis, Humboldt-Universität zu Berlin, 2020. http://dx.doi.org/10.18452/22232.
Full textAbstract:
Die Erforschung der neuronaler Grundlagen des Arbeitsgedächtnisses (WM) fand breite Aufmerksamkeit, konzentrierte sich aber auf die Speicherung sensorischer Inhalte. Beweise für die kurzfristige Aufrechterhaltung abstrakter, verbaler oder kategorischer Informationen sind selten. Ziel dieser Arbeit ist die Untersuchung der neuronalen Repräsentation von WM-Inhalten auf verschiedenen Abstraktionsebenen. Ich stelle hier drei empirische Studien vor, in denen fMRT, multivariate Musteranalyse oder probabilistische Modelle als Hauptmethoden eingesetzt wurden. Die erste Studie identifizierte kortikale Regionen, die den WM-Inhalt eines Skripts behielten. Chinesische Muttersprachler wurden gebeten, sich bekannte chinesische Zeichen zu merken, was die verbale Kodierung stark fördern. Die Ergebnisse zeigten links lateralisierte sprachbezogene Hirnareale als Kandidatenspeicher für verbale Inhalte. Die zweite und dritte Studie zielten darauf ab, die Hypothese zu testen, dass Farbe als eine Kombination aus einer visuellen Repräsentation und einer kategorischen Repräsentation gespeichert wird. Die zweite Studie verwendete ein sensorisches Kodierungsmodell und ein empirisch basiertes kategorisches Kodierungsmodell, um jeweils zwei Quellen neuronaler Repräsentationen zu charakterisieren. Farbinformationen wurden in drei farbbezogenen ROIs dekodiert: V1, V4, VO1, und insbesondere wurde eine Erhöhung der kategorischen Repräsentation in vorderen kortikalen Arealen beobachtet. In der dritten Studie wurde die verzögerte Verhaltensreaktion untersucht, die ein systematisches Bias-Muster zeigte; es wurde ein probabilistisches Dual-Content-Modell implementiert, das ein mit den experimentellen Ergebnissen hoch korreliertes Antwortmuster erzeugte; dies bestätigte die Hypothese der mnemonischen Dual-Content Repräsentation. Diese Studien zusammen schlagen eine Arbeitsteilung entlang der rostro-kaudalen Achse des Gehirns, die auf der Abstraktionsebene der gespeicherten Inhalte basiert.Research on the neural basis of working memory (WM) has received broad attention but has focused on storage of sensory content. Evidence on short-term maintenance of abstract verbal or categorical information is scarce. This thesis aims to investigate neural representation of WM content at different levels of abstraction. I present here three empirical studies that employed fMRI, multivariate pattern analysis or probabilistic modeling as major methods. The first study identified cortical regions that retained WM content of a script. Native Chinese speakers were asked to memorize well-known Chinese characters which strongly facilitated verbal coding. Results indicated left lateralized language-related brain areas as candidate stores for verbal content. The second and the third studies aimed to test the hypothesis that color is memorized as a combination of the low-level visual representation and the abstract categorical representation. The second study utilized a conventional sensory encoding model and a novel empirical-based categorical encoding model to characterize two sources of neural representations. Color information was decoded in three color-related ROIs: V1, V4, VO1, and notably, an elevation in categorical representation was observed in more anterior cortices. In the third study, the delayed behavioral response was examined, which exhibited a systematic bias pattern; a probabilistic dual-content model was implemented, which produced response patterns highly correlated with experimental results; this confirmed the hypothesis of dual-content mnemonic representations. These studies together suggest a division of labor along the rostral-caudal axis of the brain, based on the abstraction level of memorized contents.
23
Silvestre, André Meyer. "Raciocínio probabilístico aplicado ao diagnóstico de insuficiência cardíaca congestiva (ICC)." reponame:Biblioteca Digital de Teses e Dissertações da UFRGS, 2003. http://hdl.handle.net/10183/12679.
Full textAbstract:
As Redes Bayesianas constituem um modelo computacional adequado para a realização de inferências probabilísticas em domínios que envolvem a incerteza. O raciocínio diagnóstico médico pode ser caracterizado como um ato de inferência probabilística em um domínio incerto, onde a elaboração de hipóteses diagnósticas é representada pela estratificação de doenças em função das probabilidades a elas associadas. A presente dissertação faz uma pesquisa sobre a metodologia para construção/validação de redes bayesianas voltadas à área médica, e utiliza estes conhecimentos para o desenvolvimento de uma rede probabilística para o auxílio diagnóstico da Insuficiência Cardíaca (IC). Esta rede bayesiana, implementada como parte do sistema SEAMED/AMPLIA, teria o papel de alerta para o diagnóstico e tratamento precoce da IC, o que proporcionaria uma maior agilidade e eficiência no atendimento de pacientes portadores desta patologia.Bayesian networks (BN) constitute an adequate computational model to make probabilistic inference in domains that involve uncertainty. Medical diagnostic reasoning may be characterized as an act of probabilistic inference in an uncertain domain, where diagnostic hypotheses elaboration is represented by the stratification of diseases according to the related probabilities. The present dissertation researches the methodology used in the construction/validation of Bayesian Networks related to the medical field, and makes use of this knowledge for the development of a probabilistic network to aid in the diagnosis of Heart Failure (HF). This BN, implemented as part of the SEAMED/AMPLIA System, would engage in the role of alerting for early diagnosis and treatment of HF, which could provide faster and more efficient healthcare of patients carrying this pathology.
24
Jain, Dominik [Verfasser], Michael [Akademischer Betreuer] Beetz, and Marc [Akademischer Betreuer] Toussaint. "Probabilistic Cognition for Technical Systems : Statistical Relational Models for High-Level Knowledge Representation, Learning and Reasoning / Dominik Jain. Gutachter: Michael Beetz ; Marc Toussaint. Betreuer: Michael Beetz." München : Universitätsbibliothek der TU München, 2012. http://d-nb.info/1031076190/34.
Full text
25
Shan, Yin Information Technology & Electrical Engineering Australian Defence Force Academy UNSW. "Program distribution estimation with grammar models." Awarded by:University of New South Wales - Australian Defence Force Academy. School of Information Technology and Electrical Engineering, 2005. http://handle.unsw.edu.au/1959.4/38737.
Full textAbstract:
This thesis studies grammar-based approaches in the application of Estimation of Distribution Algorithms (EDA) to the tree representation widely used in Genetic Programming (GP). Although EDA is becoming one of the most active fields in Evolutionary computation (EC), the solution representation in most EDA is a Genetic Algorithms (GA) style linear representation. The more complex tree representations, resembling GP, have received only limited exploration. This is unfortunate, because tree representations provide a natural and expressive way of representing solutions for many problems. This thesis aims to help fill this gap, exploring grammar-based approaches to extending EDA to GP-style tree representations. This thesis firstly provides a comprehensive survey of current research on EDA with emphasis on EDA with GP-style tree representation. The thesis attempts to clarify the relationship between EDA with conventional linear representations and those with a GP-style tree representation, and to reveal the unique difficulties which face this research. Secondly, the thesis identifies desirable properties of probabilistic models for EDA with GP-style tree representation, and derives the PRODIGY framework as a consequence. Thirdly, following the PRODIGY framework, three methods are proposed. The first method is Program Evolution with Explicit Learning (PEEL). Its incremental general-to-specific grammar learning method balances the effectiveness and efficiency of the grammar learning. The second method is Grammar Model-based Program Evolution (GMPE). GMPE realises the PRODIGY framework by introducing elegant inference methods from the formal grammar field. GMPE provides good performance on some problems, but also provides a means to better understand some aspects of conventional GP, especially the building block hypothesis. The third method is Swift GMPE (sGMPE), which is an extension of GMPE, aiming at reducing the computational cost. Fourthly, a more accurate Minimum Message Length metric for grammar learning in PRODIGY is derived in this thesis. This metric leads to improved performance in the GMPE system, but may also be useful in grammar learning in general. It is also relevant to the learning of other probabilistic graphical models.
26
Izydorczyk, Lucas. "Probabilistic backward McKean numerical methods for PDEs and one application to energy management." Electronic Thesis or Diss., Institut polytechnique de Paris, 2021. http://www.theses.fr/2021IPPAE008.
Full textAbstract:
Cette thèse s'intéresse aux équations différentielles stochastiques de type McKean(EDS) et à leur utilisation pour représenter des équations aux dérivées partielles (EDP) non linéaires. Ces équations ne dépendent pas seulement du temps et de la position d'une certaine particule mais également de sa loi. En particulier nous traitons le cas inhabituel de la représentation d'EDP de type Fokker-Planck avec condition terminale fixée. Nous discutons existence et unicité pour ces EDP et de leur représentation sous la forme d'une EDS de type McKean, dont l'unique solutioncorrespond à la dynamique du retourné dans le temps d'un processus de diffusion.Nous introduisons la notion de représentation complètement non-linéaire d'une EDP semilinéaire. Celle-ci consiste dans le couplage d'une EDS rétrograde et d'un processus solution d'une EDS évoluant de manière rétrograde dans le temps. Nous discutons également une application à la représentation d'une équation d'Hamilton-Jacobi-Bellman (HJB) en contrôle stochastique. Sur cette base, nous proposonsun algorithme de Monte-Carlo pour résoudre des problèmes de contrôle. Celui ciest avantageux en termes d'efficience calculatoire et de mémoire, en comparaisonavec les approches traditionnelles progressive rétrograde. Nous appliquons cette méthode dans le contexte de la gestion de la demande dans les réseaux électriques. Pour finir, nous faisons le point sur l'utilisation d'EDS de type McKean généralisées pour représenter des EDP non-linéaires et non-conservatives plus générales que Fokker-PlanckThis thesis concerns McKean Stochastic Differential Equations (SDEs) to representpossibly non-linear Partial Differential Equations (PDEs). Those depend not onlyon the time and position of a given particle, but also on its probability law. In particular, we treat the unusual case of Fokker-Planck type PDEs with prescribed final data. We discuss existence and uniqueness for those equations and provide a probabilistic representation in the form of McKean type equation, whose unique solution corresponds to the time-reversal dynamics of a diffusion process.We introduce the notion of fully backward representation of a semilinear PDE: thatconsists in fact in the coupling of a classical Backward SDE with an underlying processevolving backwardly in time. We also discuss an application to the representationof Hamilton-Jacobi-Bellman Equation (HJB) in stochastic control. Based on this, we propose a Monte-Carlo algorithm to solve some control problems which has advantages in terms of computational efficiency and memory whencompared to traditional forward-backward approaches. We apply this method in the context of demand side management problems occurring in power systems. Finally, we survey the use of generalized McKean SDEs to represent non-linear and non-conservative extensions of Fokker-Planck type PDEs
27
Lees, Benjamin T. "Quantum spin systems, probabilistic representations and phase transitions." Thesis, University of Warwick, 2016. http://wrap.warwick.ac.uk/82123/.
Full textAbstract:
This thesis investigates properties of classical and quantum spin systems on lattices. These models have been widely studied due to their relevance to condensed matter physics. We identify the ground states of an antiferromagnetic RP2 model, these ground states are very di�erent from the ferromagnetic model and there was some disagreement over their structure, we settle this disagreement. Correlation inequalities are proved for the spin- 1/2 XY model and the ground state of the spin-1 XY model. This provides fresh results in a topic that had been stagnant and allows the proof of some new results, for example existence of some correlation functions in the thermodynamic limit. The occurrence of nematic order at low temperature in a quantum nematic model is proved using the method of reflection positivity and infrared bounds. Previous results on this nematic order were achieved indirectly via a probabilistic representation. This result is maintained in the presence of a small antiferromagnetic interaction, this case was not previously covered. Probabilistic representations for quantum spin systems are introduced and some consequences are presented. In particular, N´eel order is proved in a bilinear-biquadratic spin-1 system at low temperature. This result extends the famous result of Dyson, Lieb and Simon [35]. Dilute spin systems are introduced and the occurrence of a phase transition at low temperature characterised by preferential occupation of the even or odd sublattice of a cubic box is proved. This result is the first of its type for such a mixed classical and quantum system. A probabilistic representation of the spin-1 Bose-Hubbard model is also presented and some consequences are proved.
28
Dondrup, Christian. "Human-robot spatial interaction using probabilistic qualitative representations." Thesis, University of Lincoln, 2016. http://eprints.lincoln.ac.uk/28665/.
Full textAbstract:
Current human-aware navigation approaches use a predominantly metric representation of the interaction which makes them susceptible to changes in the environment. In order to accomplish reliable navigation in ever-changing human populated environments, the presented work aims to abstract from the underlying metric representation by using Qualitative Spatial Relations (QSR), namely the Qualitative Trajectory Calculus (QTC), for Human-Robot Spatial Interaction (HRSI). So far, this form of representing HRSI has been used to analyse different types of interactions online. This work extends this representation to be able to classify the interaction type online using incrementally updated QTC state chains, create a belief about the state of the world, and transform this high-level descriptor into low-level movement commands. By using QSRs the system becomes invariant to change in the environment, which is essential for any form of long-term deployment of a robot, but most importantly also allows the transfer of knowledge between similar encounters in different environments to facilitate interaction learning. To create a robust qualitative representation of the interaction, the essence of the movement of the human in relation to the robot and vice-versa is encoded in two new variants of QTC especially designed for HRSI and evaluated in several user studies. To enable interaction learning and facilitate reasoning, they are employed in a probabilistic framework using Hidden Markov Models (HMMs) for online classiffication and evaluation of their appropriateness for the task of human-aware navigation. In order to create a system for an autonomous robot, a perception pipeline for the detection and tracking of humans in the vicinity of the robot is described which serves as an enabling technology to create incrementally updated QTC state chains in real-time using the robot's sensors. Using this framework, the abstraction and generalisability of the QTC based framework is tested by using data from a different study for the classiffication of automatically generated state chains which shows the benefits of using such a highlevel description language. The detriment of using qualitative states to encode interaction is the severe loss of information that would be necessary to generate behaviour from it. To overcome this issue, so-called Velocity Costmaps are introduced which restrict the sampling space of a reactive local planner to only allow the generation of trajectories that correspond to the desired QTC state. This results in a exible and agile behaviour I generation that is able to produce inherently safe paths. In order to classify the current interaction type online and predict the current state for action selection, the HMMs are evolved into a particle filter especially designed to work with QSRs of any kind. This online belief generation is the basis for a exible action selection process that is based on data acquired using Learning from Demonstration (LfD) to encode human judgement into the used model. Thereby, the generated behaviour is not only sociable but also legible and ensures a high experienced comfort as shown in the experiments conducted. LfD itself is a rather underused approach when it comes to human-aware navigation but is facilitated by the qualitative model and allows exploitation of expert knowledge for model generation. Hence, the presented work bridges the gap between the speed and exibility of a sampling based reactive approach by using the particle filter and fast action selection, and the legibility of deliberative planners by using high-level information based on expert knowledge about the unfolding of an interaction.
29
Stuhlmüller, Andreas. "Modeling cognition with probabilistic programs : representations and algorithms." Thesis, Massachusetts Institute of Technology, 2015. http://hdl.handle.net/1721.1/100860.
Full textAbstract:
Thesis: Ph. D., Massachusetts Institute of Technology, Department of Brain and Cognitive Sciences, 2015.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 167-176).
This thesis develops probabilistic programming as a productive metaphor for understanding cognition, both with respect to mental representations and the manipulation of such representations. In the first half of the thesis, I demonstrate the representational power of probabilistic programs in the domains of concept learning and social reasoning. I provide examples of richly structured concepts, defined in terms of systems of relations, subparts, and recursive embeddings, that are naturally expressed as programs and show initial experimental evidence that they match human generalization patterns. I then proceed to models of reasoning about reasoning, a domain where the expressive power of probabilistic programs is necessary to formalize our intuitive domain understanding due to the fact that, unlike previous formalisms, probabilistic programs allow conditioning to be represented in a model, not just applied to a model. I illustrate this insight with programs that model nested reasoning in game theory, artificial intelligence, and linguistics. In the second half, I develop three inference algorithms with the dual intent of showing how to efficiently compute the marginal distributions defined by probabilistic programs, and providing building blocks for process-level accounts of human cognition. First, I describe a Dynamic Programming algorithm for computing the marginal distribution of discrete probabilistic programs by compiling to systems of equations and show that it can make inference in models of "reasoning about reasoning" tractable by merging and reusing subcomputations. Second, I introduce the setting of amortized inference and show how learning inverse models lets us leverage samples generated by other inference algorithms to compile probabilistic models into fast recognition functions. Third, I develop a generic approach to coarse-to-fine inference in probabilistic programs and provide evidence that it can speed up inference in models with large state spaces that have appropriate hierarchical structure. Finally, I substantiate the claim that probabilistic programming is a productive metaphor by outlining new research questions that have been opened up by this line of investigation.
by Andreas Stuhlmüller.
Ph. D.
30
Moraes, Carlos Afonso Silveira. "Registros de Representação Semiótica: Contribuições para o letramento probabilístico no 9º ano do Ensino Fundamental." Universidade Federal de São Carlos, 2017. https://repositorio.ufscar.br/handle/ufscar/9234.
Full textAbstract:
Submitted by Carlos Moraes (carlos.afonso.moraes@ig.com.br) on 2017-12-13T16:09:18Z
No. of bitstreams: 2
Dissetação final corrigida dezembro.pdf: 3482457 bytes, checksum: ae6936e91b4e36f6045411e79cf8d412 (MD5)
Declaração UFSCAR.pdf: 762831 bytes, checksum: f94579acaea54b2f2ac4763c06c19706 (MD5)Approved for entry into archive by Milena Rubi ( ri.bso@ufscar.br) on 2017-12-13T17:45:58Z (GMT) No. of bitstreams: 2 Dissetação final corrigida dezembro.pdf: 3482457 bytes, checksum: ae6936e91b4e36f6045411e79cf8d412 (MD5) Declaração UFSCAR.pdf: 762831 bytes, checksum: f94579acaea54b2f2ac4763c06c19706 (MD5)
Approved for entry into archive by Milena Rubi ( ri.bso@ufscar.br) on 2017-12-13T17:46:07Z (GMT) No. of bitstreams: 2 Dissetação final corrigida dezembro.pdf: 3482457 bytes, checksum: ae6936e91b4e36f6045411e79cf8d412 (MD5) Declaração UFSCAR.pdf: 762831 bytes, checksum: f94579acaea54b2f2ac4763c06c19706 (MD5)
Made available in DSpace on 2017-12-13T17:46:24Z (GMT). No. of bitstreams: 2 Dissetação final corrigida dezembro.pdf: 3482457 bytes, checksum: ae6936e91b4e36f6045411e79cf8d412 (MD5) Declaração UFSCAR.pdf: 762831 bytes, checksum: f94579acaea54b2f2ac4763c06c19706 (MD5) Previous issue date: 2017-10-31
Não recebi financiamento
This research had the objective of describing and analyzing a teaching-learning concept of Probability in two classes of the ninth elementary school, in a municipal public school in Salto de Pirapora, in the interior of the State of São Paulo. The acquisition of probabilistic language in learning concept of probability was a motivating factor for the research project. The theoretical contributions of this research involved the records of semiotic representation by Raymond Duval and the literary probabilistic in the perspective of Iddo Gal. The guiding question of the research was: "How are records of semiotic representation mobilized and coordinated in tasks involving the context probabilistic? "A field work was elaborated with activities involving classical and frequentist probability, counting and statistics and a didactic sequence using experiments sample space, probability of simple events, events composites, bar graphs, relative frequency, frequency distribution and the tree diagram. As a teacher-researcher, the production of information originated from activities developed by students in the form of written protocols, in addition to audio records of dialogues that occurred in the correction of activities and records in the logbook. The results of the analysis of the empirical material of the research revealed that the students used different registers of semiotic representation in the resolution of tasks. The mobilization and coordination of these registers support the development of students' probabilistic literacy. Like this work was derived from the analysis of a pedagogical practice, it is expected there are contributions to the teaching practice in content involving combinatorial, statistical and probability for elementary school.
Esta pesquisa teve por objetivo descrever e analisar um cenário de ensinoaprendizagem do conceito de Probabilidade em duas classes do nono ano do Ensino Fundamental, em uma escola pública da rede municipal de ensino do município de Salto de Pirapora, interior do Estado de São Paulo. A aquisição da linguagem probabilística na aprendizagem de conceitos relativos à probabilidade foi um elemento motivador para o projeto de pesquisa. Os aportes teóricos dessa pesquisa envolveu os registros de representação semiótica por Raymond Duval e o letramento probabilístico na perspectiva de Iddo Gal. A questão orientadora da investigação foi: “Como os registros de representação semiótica são mobilizados e coordenados em tarefas envolvendo o contexto probabilístico?” Foi elaborado um trabalho de campo com atividades envolvendo a probabilidade clássica e frequentista, processos de contagem e estatística e uma sequência didática que utiliza experimentos aleatórios, espaço amostral, probabilidade de eventos simples, eventos compostos, gráficos de barra, frequência relativa, tabela de distribuição de frequência e o diagrama da árvore. Na condição de professor-pesquisador, a produção de informações foi oriunda de atividades desenvolvidas pelos alunos na forma de protocolos escritos, além de registros em áudio de diálogos ocorridos na correção das atividades e registros elaborados no diário de bordo. Os resultados da análise do material empírico da pesquisa revelaram nessa pesquisa de que os alunos utilizaram diferentes registros de representação semiótica na resolução das tarefas. A mobilização e coordenação desses registros favoreceram o desenvolvimento do letramento probabilístico dos alunos. Como este trabalho foi oriundo da análise de uma prática pedagógica, espera-se que haja contribuições para a prática docente em conteúdos envolvendo combinatória, estatística e probabilidade para o Ensino Fundamental.
31
Golmard, Jean-Louis. "Les reseaux probabilistes : representation, utilisation et acquisition des connaissances." Paris 6, 1992. http://www.theses.fr/1992PA066158.
Full textAbstract:
Cette these est basee sur quatre articles. Ceux-ci sont precedes de trois chapitres. Les reseaux probabilistes sont definis dans le premier chapitre comme des doublets (graphe, distribution de probabilite). Une typologie des reseaux probabilistes est proposee. Le deuxieme chapitre presente les articles, en decrivant pour chacun d'eux son contexte, un resume et des complements. Le troisieme chapitre conclut la presentation du memoire en decrivant brievement nos perspectives de recherches personnelles, ainsi qu'une application medicale. L'article a propose un formalisme de representation des connaissances base sur un modele appele modele auto-logistique. L'avantage principal de ce modele est son faible nombre de parametres. L'article b traite des algorithmes d'inference. Ceux-ci sont divises en deux categories: les methodes exactes et les methodes stochastiques. On montre que les methodes exactes sont toutes basees sur la transformation des graphes en arbres. Les deux derniers articles decrivent des methodes d'acquisition des connaissances, que la structure du graphe soit connue (article c) ou inconnue (article d). Ils supposent une structure simple du graphe, et l'inobservabilite de certaines variables. Dans les deux articles, une experience de simulation suit la description des methodes
32
Gahli, Ahmed. "Novel probabilistic image representations for information-based image description and analysis." Thesis, University of Nottingham, 1999. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.285686.
Full text
33
Gyftodimos, Elias. "A probabilistic graphical model framework for higher-order term-based representations." Thesis, University of Bristol, 2005. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.425088.
Full text
34
Le, cavil Anthony. "Représentation probabiliste de type progressif d'EDP nonlinéaires nonconservatives et algorithmes particulaires." Thesis, Université Paris-Saclay (ComUE), 2016. http://www.theses.fr/2016SACLY023.
Full textAbstract:
Dans cette thèse, nous proposons une approche progressive (forward) pour la représentation probabiliste d'Equations aux Dérivées Partielles (EDP) nonlinéaires et nonconservatives, permettant ainsi de développer un algorithme particulaire afin d'en estimer numériquement les solutions. Les Equations Différentielles Stochastiques Nonlinéaires de type McKean (NLSDE) étudiées dans la littérature constituent une formulation microscopique d'un phénomène modélisé macroscopiquement par une EDP conservative. Une solution d'une telle NLSDE est la donnée d'un couple $(Y,u)$ où $Y$ est une solution d' équation différentielle stochastique (EDS) dont les coefficients dépendent de $u$ et de $t$ telle que $u(t,cdot)$ est la densité de $Y_t$. La principale contribution de cette thèse est de considérer des EDP nonconservatives, c'est-à- dire des EDP conservatives perturbées par un terme nonlinéaire de la forme $Lambda(u,nabla u)u$. Ceci implique qu'un couple $(Y,u)$ sera solution de la représentation probabiliste associée si $Y$ est un encore un processus stochastique et la relation entre $Y$ et la fonction $u$ sera alors plus complexe. Etant donnée la loi de $Y$, l'existence et l'unicité de $u$ sont démontrées par un argument de type point fixe via une formulation originale de type Feynmann-KacThis thesis performs forward probabilistic representations of nonlinear and nonconservative Partial Differential Equations (PDEs), which allowto numerically estimate the corresponding solutions via an interacting particle system algorithm, mixing Monte-Carlo methods and non-parametric density estimates.In the literature, McKean typeNonlinear Stochastic Differential Equations (NLSDEs) constitute the microscopic modelof a class of PDEs which are conservative. The solution of a NLSDEis generally a couple $(Y,u)$ where $Y$ is a stochastic process solving a stochastic differential equation whose coefficients depend on $u$ and at each time $t$, $u(t,cdot)$ is the law density of the random variable $Y_t$.The main idea of this thesis is to consider this time a non-conservative PDE which is the result of a conservative PDE perturbed by a term of the type $Lambda(u, nabla u) u$. In this case, the solution of the corresponding NLSDE is again a couple $(Y,u)$, where again $Y$ is a stochastic processbut where the link between the function $u$ and $Y$ is more complicated and once fixed the law of $Y$, $u$ is determined by a fixed pointargument via an innovating Feynmann-Kac type formula
35
Ule, Tylman. "Treebank refinement optimising representations of syntactic analyses for probabilistic context-free parsing /." [S.l. : s.n.], 2007.
Find full text
36
Le, cavil Anthony. "Représentation probabiliste de type progressif d'EDP nonlinéaires nonconservatives et algorithmes particulaires." Electronic Thesis or Diss., Université Paris-Saclay (ComUE), 2016. http://www.theses.fr/2016SACLY023.
Full textAbstract:
Dans cette thèse, nous proposons une approche progressive (forward) pour la représentation probabiliste d'Equations aux Dérivées Partielles (EDP) nonlinéaires et nonconservatives, permettant ainsi de développer un algorithme particulaire afin d'en estimer numériquement les solutions. Les Equations Différentielles Stochastiques Nonlinéaires de type McKean (NLSDE) étudiées dans la littérature constituent une formulation microscopique d'un phénomène modélisé macroscopiquement par une EDP conservative. Une solution d'une telle NLSDE est la donnée d'un couple (Y,u) où Y est une solution d' équation différentielle stochastique (EDS) dont les coefficients dépendent de u et de t telle que u(t,.) est la densité de Yt. La principale contribution de cette thèse est de considérer des EDP nonconservatives, c'est-à- dire des EDP conservatives perturbées par un terme nonlinéaire de la forme Lambda(u,nabla u)u. Ceci implique qu'un couple (Y,u) sera solution de la représentation probabiliste associée si Y est un encore un processus stochastique et la relation entre Y et la fonction u sera alors plus complexe. Etant donnée la loi de Y, l'existence et l'unicité de u sont démontrées par un argument de type point fixe via une formulation originale de type Feynmann-KacThis thesis performs forward probabilistic representations of nonlinear and nonconservative Partial Differential Equations (PDEs), which allowto numerically estimate the corresponding solutions via an interacting particle system algorithm, mixing Monte-Carlo methods and non-parametric density estimates.In the literature, McKean typeNonlinear Stochastic Differential Equations (NLSDEs) constitute the microscopic modelof a class of PDEs which are conservative. The solution of a NLSDEis generally a couple (Y,u) where Y is a stochastic process solving a stochastic differential equation whose coefficients depend on u and at each time t, u(t,.) is the law density of the random variable Yt.The main idea of this thesis is to consider this time a non-conservative PDE which is the result of a conservative PDE perturbed by a term of the type Lambda(u, nabla u) u. In this case, the solution of the corresponding NLSDE is again a couple (Y,u), where again Y is a stochastic processbut where the link between the function u and Y is more complicated and once fixed the law of Y, u is determined by a fixed pointargument via an innovating Feynmann-Kac type formula
37
Exarchakis, Georgios [Verfasser], Jörg [Akademischer Betreuer] Lücke, and Bruno [Akademischer Betreuer] Olshausen. "Probabilistic models for invariant representations and transformations / Georgios Exarchakis ; Jörg Lücke, Bruno Olshausen." Oldenburg : BIS der Universität Oldenburg, 2016. http://d-nb.info/1141904470/34.
Full text
38
Spiegel, Christoph. "Additive structures and randomness in combinatorics." Doctoral thesis, Universitat Politècnica de Catalunya, 2020. http://hdl.handle.net/10803/669327.
Full textAbstract:
Arithmetic Combinatorics, Combinatorial Number Theory, Structural Additive Theory and Additive Number Theory are just some of the terms used to describe the vast field that sits at the intersection of Number Theory and Combinatorics and which will be the focus of this thesis. Its contents are divided into two main parts, each containing several thematically related results.
The first part deals with the question under what circumstances solutions to arbitrary linear systems of equations usually occur in combinatorial structures..The properties we will be interested in studying in this part relate to the solutions to linear systems of equations.
A first question one might ask concerns the point at which sets of a given size will typically contain a solution. We will establish a threshold and also study the distribution of the number of solutions at that threshold, showing that it converges to a Poisson distribution in certain cases.
Next, Van der Waerden’s Theorem, stating that every finite coloring of the integers contains monochromatic arithmetic progression of arbitrary length, is by some considered to be the first result in Ramsey Theory. Rado generalized van der Waerden’s result by characterizing those linear systems whose solutions satisfy a similar property and Szemerédi strengthened it to a statement concerning density rather than colorings. We will turn our attention towards versions of Rado’s and Szemerédi’s Theorem in random sets, extending previous work of Friedgut, Rödl, Rucin´ski and Schacht in the case of the former and of Conlon, Gowers and Schacht for the latter to include a larger variety of systems and solutions.
Lastly, Chvátal and Erdo¿s suggested studying Maker-Breaker games. These games have deep connections to the theory of random structures and we will build on work of Bednarska and Luczak to establish the threshold for how much a large variety of games need to be biased in favor of the second player. These include games in which the first player wants to occupy a solution to some given linear system, generalizing the van der Waerden games introduced by Beck.
The second part deals with the extremal behavior of sets with interesting additive properties. In particular, we will be interested in bounds or structural descriptions for sets exhibiting some restrictions with regards to either their representation function or their sumset.
First, we will consider Sidon sets, that is sets of integers with pairwise unique differences. We will study a generalization of Sidon sets proposed very recently by Kohayakawa, Lee, Moreira and Rödl, where the pairwise differences are not just distinct, but in fact far apart by a certain measure. We will obtain strong lower bounds for such infinite sets using an approach of Cilleruelo. As a consequence of these bounds, we will also obtain the best current lower bound for Sidon sets in randomly generated infinite sets of integers of high density.
Next, one of the central results at the intersection of Combinatorics and Number Theory is the Freiman–Ruzsa Theorem stating that any finite set of integers of given doubling can be efficiently covered by a generalized arithmetic progression. In the case of particularly small doubling, more precise structural descriptions exist. We will first study results going beyond Freiman’s well-known 3k–4 Theorem in the integers. We will then see an application of these results to sets of small doubling in finite cyclic groups.
Lastly, we will turn our attention towards sets with near-constant representation functions. Erdo¿s and Fuchs established that representation functions of arbitrary sets of integers cannot be too close to being constant. We will first extend the result of Erdo¿s and Fuchs to ordered representation functions. We will then address a related question of Sárközy and Sós regarding weighted representation function.La combinatòria aritmètica, la teoria combinatòria dels nombres, la teoria additiva estructural i la teoria additiva de nombres són alguns dels termes que es fan servir per descriure una branca extensa i activa que es troba en la intersecció de la teoria de nombres i de la combinatòria, i que serà el motiu d'aquesta tesi doctoral. La primera part tracta la qüestió de sota quines circumstàncies es solen produir solucions a sistemes lineals d’equacions arbitràries en estructures additives. Una primera pregunta que s'estudia es refereix al punt en que conjunts d’una mida determinada contindran normalment una solució. Establirem un llindar i estudiarem també la distribució del nombre de solucions en aquest llindar, tot demostrant que en certs casos aquesta distribució convergeix a una distribució de Poisson. El següent tema de la tesis es relaciona amb el teorema de Van der Waerden, que afirma que cada coloració finita dels nombres enters conté una progressió aritmètica monocromàtica de longitud arbitrària. Aquest es considera el primer resultat en la teoria de Ramsey. Rado va generalitzar el resultat de van der Waerden tot caracteritzant en aquells sistemes lineals les solucions de les quals satisfan una propietat similar i Szemerédi la va reforçar amb una versió de densitat del resultat. Centrarem la nostra atenció cap a versions del teorema de Rado i Szemerédi en conjunts aleatoris, ampliant els treballs anteriors de Friedgut, Rödl, Rucinski i Schacht i de Conlon, Gowers i Schacht. Per últim, Chvátal i Erdos van suggerir estudiar estudiar jocs posicionals del tipus Maker-Breaker. Aquests jocs tenen una connexió profunda amb la teoria de les estructures aleatòries i ens basarem en el treball de Bednarska i Luczak per establir el llindar de la quantitat que necessitem per analitzar una gran varietat de jocs en favor del segon jugador. S'inclouen jocs en què el primer jugador vol ocupar una solució d'un sistema lineal d'equacions donat, generalitzant els jocs de van der Waerden introduïts per Beck. La segona part de la tesis tracta sobre el comportament extrem dels conjunts amb propietats additives interessants. Primer, considerarem els conjunts de Sidon, és a dir, conjunts d’enters amb diferències úniques quan es consideren parelles d'elements. Estudiarem una generalització dels conjunts de Sidons proposats recentment per Kohayakawa, Lee, Moreira i Rödl, en que les diferències entre parelles no són només diferents, sinó que, en realitat, estan allunyades una certa proporció en relació a l'element més gran. Obtindrem límits més baixos per a conjunts infinits que els obtinguts pels anteriors autors tot usant una construcció de conjunts de Sidon infinits deguda a Cilleruelo. Com a conseqüència d'aquests límits, obtindrem també el millor límit inferior actual per als conjunts de Sidon en conjunts infinits generats aleatòriament de nombres enters d'alta densitat. A continuació, un dels resultats centrals a la intersecció de la combinatòria i la teoria dels nombres és el teorema de Freiman-Ruzsa, que afirma que el conjunt suma d'un conjunt finit d’enters donats pot ser cobert de manera eficient per una progressió aritmètica generalitzada. En el cas de que el conjunt suma sigui de mida petita, existeixen descripcions estructurals més precises. Primer estudiarem els resultats que van més enllà del conegut teorema de Freiman 3k-4 en els enters. Llavors veurem una aplicació d’aquests resultats a conjunts de dobles petits en grups cíclics finits. Finalment, dirigirem l’atenció cap a conjunts amb funcions de representació gairebé constants. Erdos i Fuchs van establir que les funcions de representació de conjunts arbitraris d’enters no poden estar massa a prop de ser constants. Primer estendrem el resultat d’Erdos i Fuchs a funcions de representació ordenades. A continuació, abordarem una pregunta relacionada de Sárközy i Sós sobre funció de representació ponderada.
39
Hernandez, Gabriel. "A probabilistic-based design approach with game theoretical representations of the enterprise design process." Thesis, Georgia Institute of Technology, 1998. http://hdl.handle.net/1853/33422.
Full text
40
Aliakbari, khoei Mina. "Une approche computationnelle de la dépendance au mouvement du codage de la position dans la système visuel." Thesis, Aix-Marseille, 2014. http://www.theses.fr/2014AIXM4041/document.
Full textAbstract:
Cette thèse est centralisée sur cette question : comment est-ce que le système visuel peut coder efficacement la position des objets en mouvement, en dépit des diverses sources d'incertitude ? Cette étude déploie une hypothèse sur la connaissance a priori de la cohérence temporelle du mouvement (Burgi et al 2000; Yuille and Grzywacz 1989). Nous avons ici étendu le cadre de modélisation précédemment proposé pour expliquer le problème de l'ouverture (Perrinet and Masson, 2012). C'est un cadre d'estimation de mouvement Bayésien mis en oeuvre par un filtrage particulaire, que l'on appelle la prévision basé sur le mouvement (MBP). Sur cette base, nous avons introduit une théorie du codage de position basée sur le mouvement, et étudié comment les mécanismes neuronaux codant la position instantanée de l'objet en mouvement pourraient être affectés par le signal de mouvement le long d'une trajectoire. Les résultats de cette thèse suggèrent que le codage de la position basé sur le mouvement peut constituer un calcul neuronal générique parmi toutes les étapes du système visuel. Cela peut en partie compenser les effets cumulatifs des délais neuronaux dans le codage de la position. En outre, il peut expliquer des changements de position basés sur le mouvement, comme par example, l'Effect de Saut de Flash. Comme un cas particulier, nous avons introduit le modèle de MBP diagonal et avons reproduit la réponse anticipée de populations de neurones dans l'aire cortical V1. Nos résultats indiquent qu'un codage en position efficace et robuste peut être fortement dépendant de l'intégration le long de la trajectoireCoding the position of moving objects is an essential ability of the visual system in fulfilling precise and robust tracking tasks. This thesis is focalized upon this question: How does the visual system efficiently encode the position of moving objects, despite various sources of uncertainty? This study deploys the hypothesis that the visual systems uses prior knowledge on the temporal coherency of motion (Burgi et al 2000; Yuille and Grzywacz 1989). We implemented this prior by extending the modeling framework previously proposed to explain the aperture problem (Perrinet and Masson, 2012), so-called motion-based prediction (MBP). This model is a Bayesian motion estimation framework implemented by particle filtering. Based on that, we have introduced a theory on motion-based position coding, to investigate how neural mechanisms encoding the instantaneous position of moving objects might be affected by motion. Results of this thesis suggest that motion-based position coding might be a generic neural computation among all stages of the visual system. This mechanism might partially compensate the accumulative and restrictive effects of neural delays in position coding. Also it may account for motion-based position shifts as the flash lag effect. As a specific case, results of diagonal MBP model reproduced the anticipatory response of neural populations in the primary visual cortex of macaque monkey. Our results imply that an efficient and robust position coding might be highly dependent on trajectory integration and that it constitutes a key neural signature to study the more general problem of predictive coding in sensory areas
41
Acerbi, Luigi. "Complex internal representations in sensorimotor decision making : a Bayesian investigation." Thesis, University of Edinburgh, 2015. http://hdl.handle.net/1842/16233.
Full textAbstract:
The past twenty years have seen a successful formalization of the idea that perception is a form of probabilistic inference. Bayesian Decision Theory (BDT) provides a neat mathematical framework for describing how an ideal observer and actor should interpret incoming sensory stimuli and act in the face of uncertainty. The predictions of BDT, however, crucially depend on the observer’s internal models, represented in the Bayesian framework by priors, likelihoods, and the loss function. Arguably, only in the simplest scenarios (e.g., with a few Gaussian variables) we can expect a real observer’s internal representations to perfectly match the true statistics of the task at hand, and to conform to exact Bayesian computations, but how humans systematically deviate from BDT in more complex cases is yet to be understood. In this thesis we theoretically and experimentally investigate how people represent and perform probabilistic inference with complex (beyond Gaussian) one-dimensional distributions of stimuli in the context of sensorimotor decision making. The goal is to reconstruct the observers’ internal representations and details of their decision-making process from the behavioural data – by employing Bayesian inference to uncover properties of a system, the ideal observer, that is believed to perform Bayesian inference itself. This “inverse problem” is not unique: in principle, distinct Bayesian observer models can produce very similar behaviours. We circumvented this issue by means of experimental constraints and independent validation of the results. To understand how people represent complex distributions of stimuli in the specific domain of time perception, we conducted a series of psychophysical experiments where participants were asked to reproduce the time interval between a mouse click and a flash, drawn from a session-dependent distribution of intervals. We found that participants could learn smooth approximations of the non-Gaussian experimental distributions, but seemed to have trouble with learning some complex statistical features such as bimodality. To investigate whether this difficulty arose from learning complex distributions or computing with them, we conducted a target estimation experiment in which “priors” where explicitly displayed on screen and therefore did not need to be learnt. Lack of difference in performance between the Gaussian and bimodal conditions in this task suggests that acquiring a bimodal prior, rather than computing with it, is the major difficulty. Model comparison on a large number of Bayesian observer models, representing different assumptions about the noise sources and details of the decision process, revealed a further source of variability in decision making that was modelled as a “stochastic posterior”. Finally, prompted by a secondary finding of the previous experiment, we tested the effect of decision uncertainty on the capacity of the participants to correct for added perturbations in the visual feedback in a centre of mass estimation task. Participants almost completely compensated for the injected error in low uncertainty trials, but only partially so in the high uncertainty ones, even when allowed sufficient time to adjust their response. Surprisingly, though, their overall performance was not significantly affected. This finding is consistent with the behaviour of a Bayesian observer with an additional term in the loss function that represents “effort” – a component of optimal control usually thought to be negligible in sensorimotor estimation tasks. Together, these studies provide new insight into the capacity and limitations people have in learning and performing probabilistic inference with distributions beyond Gaussian. This work also introduces several tools and techniques that can help in the systematic exploration of suboptimal behaviour. Developing a language to describe suboptimality, mismatching representations and approximate inference, as opposed to optimality and exact inference, is a fundamental step to link behavioural studies to actual neural computations.
42
Botha, Jan Abraham. "Probabilistic modelling of morphologically rich languages." Thesis, University of Oxford, 2014. http://ora.ox.ac.uk/objects/uuid:8df7324f-d3b8-47a1-8b0b-3a6feb5f45c7.
Full textAbstract:
This thesis investigates how the sub-structure of words can be accounted for in probabilistic models of language. Such models play an important role in natural language processing tasks such as translation or speech recognition, but often rely on the simplistic assumption that words are opaque symbols. This assumption does not fit morphologically complex language well, where words can have rich internal structure and sub-word elements are shared across distinct word forms. Our approach is to encode basic notions of morphology into the assumptions of three different types of language models, with the intention that leveraging shared sub-word structure can improve model performance and help overcome data sparsity that arises from morphological processes. In the context of n-gram language modelling, we formulate a new Bayesian model that relies on the decomposition of compound words to attain better smoothing, and we develop a new distributed language model that learns vector representations of morphemes and leverages them to link together morphologically related words. In both cases, we show that accounting for word sub-structure improves the models' intrinsic performance and provides benefits when applied to other tasks, including machine translation. We then shift the focus beyond the modelling of word sequences and consider models that automatically learn what the sub-word elements of a given language are, given an unannotated list of words. We formulate a novel model that can learn discontiguous morphemes in addition to the more conventional contiguous morphemes that most previous models are limited to. This approach is demonstrated on Semitic languages, and we find that modelling discontiguous sub-word structures leads to improvements in the task of segmenting words into their contiguous morphemes.
43
Nyman, Peter. "On relations between classical and quantum theories of information and probability." Doctoral thesis, Linnéuniversitetet, Institutionen för datavetenskap, fysik och matematik, DFM, 2011. http://urn.kb.se/resolve?urn=urn:nbn:se:lnu:diva-13830.
Full textAbstract:
In this thesis we study quantum-like representation and simulation of quantum algorithms by using classical computers.The quantum--like representation algorithm (QLRA) was introduced by A. Khrennikov (1997) to solve the ``inverse Born's rule problem'', i.e. to construct a representation of probabilistic data-- measured in any context of science-- and represent this data by a complex or more general probability amplitude which matches a generalization of Born's rule.The outcome from QLRA matches the formula of total probability with an additional trigonometric, hyperbolic or hyper-trigonometric interference term and this is in fact a generalization of the familiar formula of interference of probabilities. We study representation of statistical data (of any origin) by a probability amplitude in a complex algebra and a Clifford algebra (algebra of hyperbolic numbers). The statistical data is collected from measurements of two dichotomous and trichotomous observables respectively. We see that only special statistical data (satisfying a number of nonlinear constraints) have a quantum--like representation. We also study simulations of quantum computers on classical computers.Although it can not be denied that great progress have been made in quantum technologies, it is clear that there is still a huge gap between the creation of experimental quantum computers and realization of a quantum computer that can be used in applications. Therefore the simulation of quantum computations on classical computers became an important part in the attempt to cover this gap between the theoretical mathematical formulation of quantum mechanics and the realization of quantum computers. Of course, it can not be expected that quantum algorithms would help to solve NP problems for polynomial time on classical computers. However, this is not at all the aim of classical simulation. The second part of this thesis is devoted to adaptation of the Mathematica symbolic language to known quantum algorithms and corresponding simulations on classical computers. Concretely we represent Simon's algorithm, Deutsch-Josza algorithm, Shor's algorithm, Grover's algorithm and quantum error-correcting codes in the Mathematica symbolic language. We see that the same framework can be used for all these algorithms. This framework will contain the characteristic property of the symbolic language representation of quantum computing and it will be a straightforward matter to include future algorithms in this framework.
44
Zarzar, Gandler Gabriela. "Evaluation of probabilistic representations for modeling and understanding shape based on synthetic and real sensory data." Thesis, KTH, Skolan för datavetenskap och kommunikation (CSC), 2017. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-215650.
Full textAbstract:
The advancements in robotic perception in the recent years have empowered robots to better execute tasks in various environments. The perception of objects in the robot work space significantly relies on how sensory data is represented. In this context, 3D models of object’s surfaces have been studied as a means to provide useful insights on shape of objects and ultimately enhance robotic perception. This involves several challenges, because sensory data generally presents artifacts, such as noise and incompleteness. To tackle this problem, we employ Gaussian Process Implicit Surface (GPIS), a non-parametric probabilistic reconstruction of object’s surfaces from 3D data points. This thesis investigates different configurations for GPIS, as a means to tackle the extraction of shape information. In our approach we interpret an object’s surface as the level-set of an underlying sparse Gaussian Process (GP) with variational formulation. Results show that the variational formulation for sparse GP enables a reliable approximation to the full GP solution. Experiments are performed on a synthetic and a real sensory data set. We evaluate results by assessing how close the reconstructed surfaces are to the ground-truth correspondences, and how well objects from different categories are clustered based on the obtained representation. Finally we conclude that the proposed solution derives adequate surface representations to reason about object shape and to discriminate objects based on shape information.Framsteg inom robotperception de senaste åren har resulterat i robotar som är bättre på attutföra uppgifter i olika miljöer. Perception av objekt i robotens arbetsmiljö är beroende avhur sensorisk data representeras. I det här sammanhanget har 3D-modeller av objektytorstuderats för att ge användbar insikt om objektens form och i slutändan bättre robotperception. Detta innebär flera utmaningar, eftersom sensoriska data ofta innehåller artefakter, såsom brus och brist på data. För att hantera detta problem använder vi oss av Gaussian Process Implicit Surface (GPIS), som är en icke-parametrisk probabilistisk rekonstruktion av ett objekts yta utifrån 3D-punkter. Detta examensarbete undersöker olika konfigurationer av GPIS för att på detta sätt kunna extrahera forminformation. I vår metod tolkar vi ett objekts yta som nivåkurvor hos en underliggande gles variational Gaussian Process (GP) modell. Resultat visar att en gles variational GP möjliggör en tillförlitlig approximation av en komplett GP-lösningen. Experiment utförs på ett syntetisk och ett reellt sensorisk dataset. Vi utvärderar resultat genom att bedöma hur nära de rekonstruerade ytorna är till grundtruth- korrespondenser, och hur väl objektkategorier klustras utifrån den erhållna representationen. Slutligen konstaterar vi att den föreslagna lösningen leder till tillräckligt goda representationer av ytor för tolkning av objektens form och för att diskriminera objekt utifrån forminformation.
45
Bassolet, Cyr Gabin. "Approches connexionnistes du classement en Osiris : vers un classement probabiliste." Université Joseph Fourier (Grenoble), 1998. http://www.theses.fr/1998GRE10086.
Full textAbstract:
Le classement d'instance est une fonction importante des systèmes de représentation de connaissances. Il est présent dans les systèmes de représentation de connaissances centrée objet sous le nom de classification d'objet, dans les logiques terminologiques comme un cas particulier de la classification de concepts, et, de manière implicite, dans les systèmes à base de règles, où les faits inférés peuvent être interprétés comme l'appartenance à une classe. Nous étudions le classement d'instance en Osiris, un système de représentation de connaissances centrée objets où la notion de vue jouent un rôle central. Le classement d'instance consiste à déterminer les vues valides d'un objet, ainsi que ses vues potentielles et invalides lorsqu'il est incomplètement connu. Nous montrons une possibilité de traduction des règles de production en Osiris, explicitant ainsi la fonction de classement des systèmes experts. Les contraintes de domaine jouent un rôle privilégié en Osiris. Elles permettent de réaliser une partition du domaine de chaque attribut, partition qui se prolonge à l'espace des objets pour constituer l'espace de classement, dont les éléments sont appelés eq-classes. Tous les objets d'une eq-classe ont le même comportement vis-à-vis du classement. Nous étudions plusieurs architectures connexionnistes pour le classement en Osiris, en privilégiant la détermination complète des vues valides, invalides et potentielles lors du classement d'objets partiellement connus. Nous proposons une méthode pour le classement probabiliste, sous l'hypothèse d'indépendance des attributs. Pour cela, nous distinguons deux sous-ensembles d'Osiris où cette hypothèse peut être faite. Dans le cas général, l'approche proposée fournit un mécanisme homogène pour la détermination des vues valides, invalides et potentielles, sans valuation probabiliste de ces dernières. Enfin, nous évoquons les possibilités de prise en compte des dépendances pour le classement probabiliste.
46
Emerson, Guy Edward Toh. "Functional distributional semantics : learning linguistically informed representations from a precisely annotated corpus." Thesis, University of Cambridge, 2018. https://www.repository.cam.ac.uk/handle/1810/284882.
Full textAbstract:
The aim of distributional semantics is to design computational techniques that can automatically learn the meanings of words from a body of text. The twin challenges are: how do we represent meaning, and how do we learn these representations? The current state of the art is to represent meanings as vectors - but vectors do not correspond to any traditional notion of meaning. In particular, there is no way to talk about 'truth', a crucial concept in logic and formal semantics. In this thesis, I develop a framework for distributional semantics which answers this challenge. The meaning of a word is not represented as a vector, but as a 'function', mapping entities (objects in the world) to probabilities of truth (the probability that the word is true of the entity). Such a function can be interpreted both in the machine learning sense of a classifier, and in the formal semantic sense of a truth-conditional function. This simultaneously allows both the use of machine learning techniques to exploit large datasets, and also the use of formal semantic techniques to manipulate the learnt representations. I define a probabilistic graphical model, which incorporates a probabilistic generalisation of model theory (allowing a strong connection with formal semantics), and which generates semantic dependency graphs (allowing it to be trained on a corpus). This graphical model provides a natural way to model logical inference, semantic composition, and context-dependent meanings, where Bayesian inference plays a crucial role. I demonstrate the feasibility of this approach by training a model on WikiWoods, a parsed version of the English Wikipedia, and evaluating it on three tasks. The results indicate that the model can learn information not captured by vector space models.
47
Krompaß, Denis [Verfasser], and Volker [Akademischer Betreuer] Tresp. "Exploiting prior knowledge and latent variable representations for the statistical modeling and probabilistic querying of large knowledge graphs / Denis Krompaß. Betreuer: Volker Tresp." München : Universitätsbibliothek der Ludwig-Maximilians-Universität, 2015. http://d-nb.info/1081628847/34.
Full text
48
Sallin, Mathieu. "Approche probabiliste du diagnostic de l'état de santé des véhicules militaires terrestres en environnement incertain." Thesis, Université Clermont Auvergne (2017-2020), 2018. http://www.theses.fr/2018CLFAC099.
Full textAbstract:
Ce travail de thèse est une contribution à l’analyse de santé structurale de la caisse de véhicules militaires terrestres à roues. Appartenant à la gamme 20 - 30 tonnes, de tels véhicules sont déployés dans des contextes opérationnels variés où les conditions de roulage sont sévères et difficilement caractérisables. De plus, faisant face à la concurrence, la fonction mobilité des véhicules est acquise auprès de fournisseurs et n’est plus développée par Nexter Systems. De ce fait, la définition complète de cette fonction n’est plus connue. S’appuyant sur ce contexte, l’objectif principal de la thèse est d’aborder l’état de santé de la structure porteuse par approche probabiliste, afin de maitriser les techniques de calcul permettant la prise en compte de l’aléa intrinsèque des chargements liés à la diversité d’emploi des véhicules militaires terrestres. En particulier, les stratégies les plus pertinentes pour propager les incertitudes de roulage au sein d’un modèle mécanique d’un véhicule terrestre sont définies. Ces travaux décrivent comment il est possible d’exploiter une grandeur d’intérêt au sein du véhicule dans un objectif d’évaluation de la fiabilité par rapport à un critère de dommage donné. Une application sur un démonstrateur entièrement conçu par Nexter Systems illustre l’approche proposéeThis thesis is a contribution to the structural health analysis of the body of ground military vehicles. Belonging to the 20 - 30 tons range, such vehicles are deployed in a variety of operational contexts where driving conditions are severe and difficult to characterize. In addition, due to a growing industrial competition, the mobility function of vehicles is acquired from suppliers and is no longer developed by Nexter Systems. As a result, the complete definition of this function is unknown. Based on this context, the main objective of this thesis is to analyze the health of the vehicle body using a probabilistic approach in order to control the calculation techniques allowing to take into account the random nature of loads related to the use of ground military vehicles. In particular, the most relevant strategies for propagating uncertainties due to the terrain within a vehicle dynamics model are defined. This work describes how it is possible to manage an observation data measured in the vehicle for the purpose of assessing the reliability with respect to a given damage criterion. An application on a demonstrator entirely designed by Nexter Systems illustrates the proposed approach
49
Sayadi, Karim. "Classification du texte numérique et numérisé. Approche fondée sur les algorithmes d'apprentissage automatique." Thesis, Paris 6, 2017. http://www.theses.fr/2017PA066079/document.
Full textAbstract:
Différentes disciplines des sciences humaines telles la philologie ou la paléographie font face à des tâches complexes et fastidieuses pour l'examen des sources de données. La proposition d'approches computationnelles en humanités permet d'adresser les problématiques rencontrées telles que la lecture, l'analyse et l'archivage de façon systématique. Les modèles conceptuels élaborés reposent sur des algorithmes et ces derniers donnent lieu à des implémentations informatiques qui automatisent ces tâches fastidieuses. La première partie de la thèse vise, d'une part, à établir la structuration thématique d'un corpus, en construisant des espaces sémantiques de grande dimension. D'autre part, elle vise au suivi dynamique des thématiques qui constitue un réel défi scientifique, notamment en raison du passage à l'échelle. La seconde partie de la thèse traite de manière holistique la page d'un document numérisé sans aucune intervention préalable. Le but est d'apprendre automatiquement des représentations du trait de l'écriture ou du tracé d'un certain script par rapport au tracé d'un autre script. Il faut dans ce cadre tenir compte de l'environnement où se trouve le tracé : image, artefact, bruits dus à la détérioration de la qualité du papier, etc. Notre approche propose un empilement de réseaux de neurones auto-encodeurs afin de fournir une représentation alternative des données reçues en entréeDifferent disciplines in the humanities, such as philology or palaeography, face complex and time-consuming tasks whenever it comes to examining the data sources. The introduction of computational approaches in humanities makes it possible to address issues such as semantic analysis and systematic archiving. The conceptual models developed are based on algorithms that are later hard coded in order to automate these tedious tasks. In the first part of the thesis we propose a novel method to build a semantic space based on topics modeling. In the second part and in order to classify historical documents according to their script. We propose a novel representation learning method based on stacking convolutional auto-encoder. The goal is to automatically learn plot representations of the script or the written language
50
Mocanu, Stéphane. "Construction et propriétés des représentations monocycliques des lois de type phase." Grenoble INPG, 1999. http://www.theses.fr/1999INPG0103.
Full textAbstract:
L'etude des lois de probabilite du temps d'absorption dans une chaine de markov absorbante presente un grand interet pour la modelisation des systemes de production. Une telle loi est dite de type phase. L'utilisation des lois de type phase pour des resolutions analytiques suppose la connaissance d'une chaine de markov sous-jacente, ou representation de la loi. Generalement la construction de cette chaine est difficile a cause de la complexite et du sur-parametrage du probleme. La contribution principale de cette these est la definition d'une classe des representations a nombre reduit de parametres et la preuve du fait que toute loi de type phase accepte une telle representation. Ces representations - appelees monocycliques - sont caracterisees par les deux proprietes suivantes : (a) un etat de la chaine ne peut pas appartenir a plus d'un cycle du graphe ; (b) a l'interieur d'un cycle tous les taux de transition sont egaux. Ces representations peuvent etre vues comme des combinaisons convexes de convolutions d'exponentielles et de lois d'erlang rebouclees. La deuxieme partie de cette etude est consacree a la recherche des proprietes des lois d'erlang rebouclees. Des formules analytiques pour les densites de probabilite et pour des quantites matricielles usuelles dans les calculs stochastiques ainsi que des caracterisations des graphes des densites de probabilite ont ete obtenues. Une partie des proprietes et formules ont ete generalisees pour les representations monocycliques. Quelques exemples d'applications sont proposes. La derniere partie est dediee a deux applications. La premiere utilise une representation monocyclique pour l'identification des trois premiers moments d'une loi de probabilite dans le cas d'un coefficient de variation inferieur a 1. La deuxieme vise l'approximation du comportement caudal du graphique d'une distribution de type phase.