Dissertations / Theses on the topic 'Embedding Analysis'

This Licentiate Thesis consists of four chapters, which deal with a new Sobolev type function space called the space with multiweighted derivatives. This space is a generalization of the usual one dimensional Sobolev space. Chapter 1 is an introduction, where, in particular, the importance to study function spaces with weights is discussed and motivated. In Chapter 2 we consider and analyze some results of L. D. Kudryavtsev, where he investigated one dimensional Sobolev spaces. Moreover, in this chapter we present and prove analogous results by B. L. Baidel'dinov for generalized Sobolev spaces. These results are crucially for the proofs of the main results of this Licentiate Thesis. In Chapter 3 we prove some embedding theorems for these new generalized Sobolev spaces. The main results of Kudryavtsev and Baidel'dinov about characterization of the behavior of functions at a singularity take place in weak degeneration of spaces. However, with the help of our new embedding theorems we can extend these results to the case of strong degeneration. In Chapter 4 we prove some new estimates for each function in a Tchebychev system. In order to be able to study also compactness of the embeddings from Chapter 3 such estimates are crucial. I plan to study this question in detail in my further PhD studies.

Godkänd; 2007; 20071107 (ysko)

Steganography is the technique that hides secret messages in seemingly inoccuous data. Steganalysis is the technique that is used to discover the use of steganography. Usuallyit is difficult to extract the exact hidden message because itis usually encrypted. Though often it is enough to merely uncover the use of steganography to reveal important information.In this report we discuss some common methods for steganography in images, as well as some methods of analysis,both general and targeted towards specific methods. We test how well the methods fare against each other and what one should consider when making use of steganography. According to our results, many steganographic methods need targeted analysis attacks to be discovered. Our advice is to avoid using the full potential capacity when embeddingdata. The hidden data should also be distributed randomly in the cover file.

E-learning has acquired the status of a "radical innovation" in higher education over the past decade. This claim is contestable, but certainly as the latest educational innovation, it can be attributed with introducing significant disruption into many facets of university life, reaching well beyond the traditional activities associated with the classroom pedagogies. In Australian universities, there are many now who simply take e-learning for granted as accepted teaching and learning practice (Oliver, 2004). Conversely, there are others who forecast its demise, claiming that, like previous educational technological innovations, it is another passing fad (Noble, 1998b). This thesis does not primarily engage this debate. Instead the purpose of this thesis is to gain insight into how universities can realise sustained benefits from the considerable investments to date that have been made in educational technological innovations. The inquiry seeks to understand better change within contemporary universities, in particular the process of embedding the e-learning innovation effectively. The intention is to produce an analysis useful to university executives, managers, teachers and researchers, as well as to make a more general contribution to knowledge about innovations in organisations. The research literature on change and innovation in organisations is relevant but is reviewed and assessed as of limited value to the enquiry. This is because: * the literature mainly focuses on the objective characteristics of an innovative product which cannot encompass the socially constructed value of e-learning * it fails to differentiate between the concept of "embedding" and other change phases and constructs, mostly examining the precursory and innovation-producing processes * the context of research into innovation has been primarily industrial, not university-based * its variable analytic paradigm fails to produce holistic analyses which can be appreciated and enacted on by decision makers and practising managers. For these reasons and because suitable research on innovation in universities is lacking, an introductory investigation based on grounded theory building was undertaken. To this end, four qualitative, descriptive case studies of contrasting Australian universities embedding e-learning were compiled. The four case universities (their identities protected through use of pseudonyms) assessed were: * Gamma University - a multi-campus institution, geographically spread across urban and regional locations * Lambda University - an established university, with the majority of students located at a single urban campus * Epsilon University - a younger, multi-campus amalgamated university with a strong reputation for distance education * Delta University - a relatively young multi-campus, urban university, although its parent bodies provide a longer history. The cases were based on interviews and focus group sessions with 74 participants, and electronic resource and document analyses over two phases; the first conducted in 1998-1999 and the second in 2002-2003. These analyses provided holistic pragmatic accounts that encapsulate a number of issues. One issue was about the importance of creativity in the innovating process. A second set of issues centred on the theme of complexity and the multifarious nature of the e-learning innovation. Other themes included the significance of the innovation context, partnerships and collaborations, and the emerging polarisation of issues such as standardisation versus diversification. These issues provoked three major propositions about the process of embedding and prompted the development of two systems-based analytical frameworks; one focusing on the nature of system relationships and interactions and the second providing a longitudinal perspective of system change. The propositions are: * the ability of a university to negotiate system intersections and transitions influences the degree to which e-learning can be embedded in that university * complexity is an integral part of an innovation, therefore cannot be ignored or eliminated without destroying the kernel of the innovation itself, and its longterm viability * the efficacy of the innovation is related, in some measure, to the ability to sustain partnerships and collaborations. The analysis suggested that there are number of key influences which affect the embedding process and the ability of an organisation, such as a university, to manage the processes associated with the e-learning innovation. The key system influences which affect embedding include: * the nature of the interactions and transactions occurring within the system, at the boundaries and between the phases of transition * the importance of organisational context (cultural, technological, strategic, geographic) * the pervasive impact of complexity on all dimensions of the research problem (the e-learning innovation, the change process and the university environment) * the necessity for collaboration. The implications of this study for university executives, managers and beyond are far reaching, and in some respects contradict accepted contemporary management practice. They include: seeking ways to maximise organisational tensions to achieve positive outcomes; enhancing decision making by allowing more flexibility and personal judgement into the process; developing greater tolerance for system fuzziness and uncertainty; and encouraging better utilisation of previous knowledge gained about innovation practices and processes.

E-learning has acquired the status of a "radical innovation" in higher education over the past decade. This claim is contestable, but certainly as the latest educational innovation, it can be attributed with introducing significant disruption into many facets of university life, reaching well beyond the traditional activities associated with the classroom pedagogies. In Australian universities, there are many now who simply take e-learning for granted as accepted teaching and learning practice (Oliver, 2004). Conversely, there are others who forecast its demise, claiming that, like previous educational technological innovations, it is another passing fad (Noble, 1998b). This thesis does not primarily engage this debate. Instead the purpose of this thesis is to gain insight into how universities can realise sustained benefits from the considerable investments to date that have been made in educational technological innovations. The inquiry seeks to understand better change within contemporary universities, in particular the process of embedding the e-learning innovation effectively. The intention is to produce an analysis useful to university executives, managers, teachers and researchers, as well as to make a more general contribution to knowledge about innovations in organisations. The research literature on change and innovation in organisations is relevant but is reviewed and assessed as of limited value to the enquiry. This is because: * the literature mainly focuses on the objective characteristics of an innovative product which cannot encompass the socially constructed value of e-learning * it fails to differentiate between the concept of "embedding" and other change phases and constructs, mostly examining the precursory and innovation-producing processes * the context of research into innovation has been primarily industrial, not university-based * its variable analytic paradigm fails to produce holistic analyses which can be appreciated and enacted on by decision makers and practising managers. For these reasons and because suitable research on innovation in universities is lacking, an introductory investigation based on grounded theory building was undertaken. To this end, four qualitative, descriptive case studies of contrasting Australian universities embedding e-learning were compiled. The four case universities (their identities protected through use of pseudonyms) assessed were: * Gamma University - a multi-campus institution, geographically spread across urban and regional locations * Lambda University - an established university, with the majority of students located at a single urban campus * Epsilon University - a younger, multi-campus amalgamated university with a strong reputation for distance education * Delta University - a relatively young multi-campus, urban university, although its parent bodies provide a longer history. The cases were based on interviews and focus group sessions with 74 participants, and electronic resource and document analyses over two phases; the first conducted in 1998-1999 and the second in 2002-2003. These analyses provided holistic pragmatic accounts that encapsulate a number of issues. One issue was about the importance of creativity in the innovating process. A second set of issues centred on the theme of complexity and the multifarious nature of the e-learning innovation. Other themes included the significance of the innovation context, partnerships and collaborations, and the emerging polarisation of issues such as standardisation versus diversification. These issues provoked three major propositions about the process of embedding and prompted the development of two systems-based analytical frameworks; one focusing on the nature of system relationships and interactions and the second providing a longitudinal perspective of system change. The propositions are: * the ability of a university to negotiate system intersections and transitions influences the degree to which e-learning can be embedded in that university * complexity is an integral part of an innovation, therefore cannot be ignored or eliminated without destroying the kernel of the innovation itself, and its longterm viability * the efficacy of the innovation is related, in some measure, to the ability to sustain partnerships and collaborations. The analysis suggested that there are number of key influences which affect the embedding process and the ability of an organisation, such as a university, to manage the processes associated with the e-learning innovation. The key system influences which affect embedding include: * the nature of the interactions and transactions occurring within the system, at the boundaries and between the phases of transition * the importance of organisational context (cultural, technological, strategic, geographic) * the pervasive impact of complexity on all dimensions of the research problem (the e-learning innovation, the change process and the university environment) * the necessity for collaboration. The implications of this study for university executives, managers and beyond are far reaching, and in some respects contradict accepted contemporary management practice. They include: seeking ways to maximise organisational tensions to achieve positive outcomes; enhancing decision making by allowing more flexibility and personal judgement into the process; developing greater tolerance for system fuzziness and uncertainty; and encouraging better utilisation of previous knowledge gained about innovation practices and processes.

This thesis is devoted to the study of mixed norm spaces that arise in connection with embeddings of Sobolev and Besov type spaces. We study different structural, integrability, and smoothness properties of functions satisfying certain mixed norm conditions. Conditions of this type are determined by the behaviour of linear sections of functions. The work in this direction originates in a paper due to Gagliardo (1958), and was further developed by Fournier (1988), by Blei and Fournier (1989), and by Kolyada (2005). Here we continue these studies. We obtain some refinements of known embeddings for certain mixed norm spaces introduced by Gagliardo, and we study general properties of these spaces. In connection with these results, we consider a scale of intermediate mixed norm spaces, and prove intrinsic embeddings in this scale. We also consider more general, fully anisotropic, mixed norm spaces. Our main theorem states an embedding of these spaces to Lorentz spaces. Applying this result, we obtain sharp embedding theorems for anisotropic Sobolev-Besov spaces, and anisotropic fractional Sobolev spaces. The methods used are based on non-increasing rearrangements, and on estimates of sections of functions and sections of sets. We also study limiting relations between embeddings of spaces of different type. More exactly, mixed norm estimates enable us to get embedding constants with sharp asymptotic behaviour. This gives an extension of the results obtained for isotropic Besov spaces by Bourgain, Brezis, and Mironescu, and for anisotropic Besov spaces by Kolyada. We study also some basic properties (in particular the approximation properties) of special weak type spaces that play an important role in the construction of mixed norm spaces, and in the description of Sobolev type embeddings. In the last chapter, we study mixed norm spaces consisting of functions that have smooth sections. We prove embeddings of these spaces to Lorentz spaces. From this result, known properties of Sobolev-Liouville spaces follow.

Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2000.
Includes bibliographical references (p. 139-142).
Digital watermarking, information embedding, and data hiding systems embed information, sometimes called a digital watermark, inside a host signal, which is typically an image, audio signal, or video signal. The host signal is not degraded unacceptably in the process, and one can recover the watermark even if the composite host and watermark signal undergo a variety of corruptions and attacks as long as these corruptions do not unacceptably degrade the host signal. These systems play an important role in meeting at least three major challenges that result from the widespread use of digital communication networks to disseminate multimedia content: (1) the relative ease with which one can generate perfect copies of digital signals creates a need for copyright protection mechanisms, (2) the relative ease with which one can alter digital signals creates a need for authentication and tamper-detection methods, and (3) the increase in sheer volume of transmitted data creates a demand for bandwidth-efficient methods to either backwards-compatibly increase capacities of existing legacy networks or deploy new networks backwards-compatibly with legacy networks. We introduce a framework within which to design and analyze digital watermarking and information embedding systems. In this framework performance is characterized by achievable rate-distortion-robustness trade-offs, and this framework leads quite naturally to a new class of embedding methods called quantization index modulation (QIM). These QIM methods, especially when combined with postprocessing called distortion compensation, achieve provably better rate-distortion-robustness performance than previously proposed classes of methods such as spread spectrum methods and generalized low-bit modulation methods in a number of different scenarios, which include both intentional and unintentional attacks. Indeed, we show that distortion-compensated QIM methods can achieve capacity, the information-theoretically best possible rate-distortion-robustness performance, against both additive Gaussian noise attacks and arbitrary squared error distortion-constrained attacks. These results also have implications for the problem of communicating over broadcast channels. We also present practical implementations of QIM methods called dither modulation and demonstrate their performance both analytically and through empirical simulations.
by Brian Chen.
Ph.D.

The analysis of signals stemming from a physical system is crucial for the experimental investigation of the underlying dynamics that drives the system itself. The field of time series analysis comprises a wide variety of techniques developed with the purpose of characterizing signals and, ultimately, of providing insights on the phenomena that govern the temporal evolution of the generating system. A renowned example in this field is given by spectral analysis: the use of Fourier or Laplace transforms to bring time-domain signals into the more convenient frequency space allows to disclose the key features of linear systems. A more complex scenario turns up when nonlinearity intervenes within a system's dynamics. Nonlinear coupling between a system's degrees of freedom brings about interesting dynamical regimes, such as self-sustained periodic (though anharmonic) oscillations ("limit cycles"), or quasi-periodic evolutions that exhibit sharp spectral lines while lacking strict periodicity ("limit tori"). Among the consequences of nonlinearity, the onset of chaos is definitely the most fascinating one. Chaos is a dynamical regime characterized by unpredictability and lack of periodicity, despite being generated by deterministic laws. Signals generated by chaotic dynamical systems appear as irregular: the corresponding spectra are broad and flat, prediction of future values is challenging, and evolutions within the systems' state spaces converge to strange attractor sets with noninteger dimensionality. Because of these properties, chaotic signals can be mistakenly classified as noise if linear techniques such as spectral analysis are used. The identification of chaos and its characterization require the assessment of dynamical invariants that quantify the complex features of a chaotic system's evolution. For example, Lyapunov exponents provide a marker of unpredictability; the estimation of attractor dimensions, on the other hand, highlights the unconventional geometry of a chaotic system's state space. Nonlinear time series analysis techniques act directly within the state space of the system under investigation. However, experimentally, full access to a system's state space is not always available. Often, only a scalar signal stemming from the dynamical system can be recorded, thus providing, upon sampling, a scalar sequence. Nevertheless, by virtue of a fundamental theorem by Takens, it is possible to reconstruct a proxy of the original state space evolution out of a single, scalar sequence. This reconstruction is carried out by means of the so-called embedding procedure: m-dimensional vectors are built by picking successive elements of the scalar sequence delayed by a lag L. On the other hand, besides posing some necessary conditions on the integer embedding parameters m and L, Takens' theorem does not provide any clue on how to choose them correctly. Although many optimal embedding criteria were proposed, a general answer to the problem is still lacking. As a matter of fact, conventional methods for optimal embedding are flawed by several drawbacks, the most relevant being the need for a subjective evaluation of the outcomes of applied algorithms. Tackling the issue of optimally selecting embedding parameters makes up the core topic of this thesis work. In particular, I will discuss a novel approach that was pursued by our research group and that led to the development of a new method for the identification of suitable embedding parameters. Rather than most conventional approaches, which seek a single optimal value for m and L to embed an input sequence, our approach provides a set of embedding choices that are equivalently suitable to reconstruct the dynamics. The suitability of each embedding choice m, L is assessed by relying on statistical testing, thus providing a criterion that does not require a subjective evaluation of outcomes. The starting point of our method are embedding-dependent correlation integrals, i.e. cumulative distributions of embedding vector distances, built out of an input scalar sequence. In the case of Gaussian white noise, an analytical expression for correlation integrals is available, and, by exploiting this expression, a gauge transformation of distances is introduced to provide a more convenient representation of correlation integrals. Under this new gauge, it is possible to test—in a computationally undemanding way—whether an input sequence is compatible with Gaussian white noise and, subsequently, whether the sequence is compatible with the hypothesis of an underlying chaotic system. These two statistical tests allow ruling out embedding choices that are unsuitable to reconstruct the dynamics. The estimation of correlation dimension, carried out by means of a newly devised estimator, makes up the third stage of the method: sets of embedding choices that provide uniform estimates of this dynamical invariant are deemed to be suitable to embed the sequence.The method was successfully applied to synthetic and experimental sequences, providing new insight into the longstanding issue of optimal embedding. For example, the relevance of the embedding window (m-1)L, i.e. the time span covered by each embedding vector, is naturally highlighted by our approach. In addition, our method provides some information on the adequacy of the sampling period used to record the input sequence.The method correctly distinguishes a chaotic sequence from surrogate ones generated out of it and having the same power spectrum. The technique of surrogate generation, which I also addressed during my Ph. D. work to develop new dedicated algorithms and to analyze brain signals, allows to estimate significance levels in situations where standard analytical algorithms are unapplicable. The novel embedding approach being able to tell apart an original sequence from surrogate ones shows its capability to distinguish signals beyond their spectral—or autocorrelation—similarities.One of the possible applications of the new approach concerns another longstanding issue, namely that of distinguishing noise from chaos. To this purpose, complementary information is provided by analyzing the asymptotic (long-time) behaviour of the so-called time-dependent divergence exponent. This embedding-dependent metric is commonly used to estimate—by processing its short-time linearly growing region—the maximum Lyapunov exponent out of a scalar sequence. However, insights on the kind of source generating the sequence can be extracted from the—usually overlooked—asymptotic behaviour of the divergence exponent. Moreover, in the case of chaotic sources, this analysis also provides a precise estimate of the system's correlation dimension. Besides describing the results concerning the discrimination of chaotic systems from noise sources, I will also discuss the possibility of using the related correlation dimension estimates to improve the third stage of the method introduced above for the identification of suitable embedding parameters. The discovery of chaos as a possible dynamical regime for nonlinear systems led to the search of chaotic behaviour in experimental recordings. In some fields, this search gave plenty of positive results: for example, chaotic dynamics was successfully identified and tamed in electronic circuits and laser-based optical setups. These two families of experimental chaotic systems eventually became versatile tools to study chaos and its possible applications. On the other hand, chaotic behaviour is also looked for in climate science, biology, neuroscience, and even economics. In these fields, nonlinearity is widespread: many smaller units interact nonlinearly, yielding a collective motion that can be described by means of few, nonlinearly coupled effective degrees of freedom. The corresponding recorded signals exhibit, in many cases, an irregular and complex evolution. A possible underlying chaotic evolution—as opposed to a stochastic one—would be of interest both to reveal the presence of determinism and to predict the system's future states. While some claims concerning the existence of chaos in these fields have been made, most results are debated or inconclusive. Nonstationarity, low signal-to-noise ratio, external perturbations and poor reproducibility are just few among the issues that hinder the search of chaos in natural systems. In the final part of this work, I will briefly discuss the problem of chasing chaos in experimental recordings by considering two example sequences, the first one generated by an electronic circuit and the second one corresponding to recordings of brain activity. The present thesis is organized as follows. The core concepts of time series analysis, including the key features of chaotic dynamics, are presented in Chapter 1. A brief review of the search for chaos in experimental systems is also provided; the difficulties concerning this quest in some research fields are also highlighted. Chapter 2 describes the embedding procedure and the issue of optimally choosing the related parameters. Thereupon, existing methods to carry out the embedding choice are reviewed and their limitations are pointed out. In addition, two embedding-dependent nonlinear techniques that are ordinarily used to characterize chaos, namely the estimation of correlation dimension by means of correlation integrals and the assessment of maximum Lyapunov exponent, are presented. The new approach for the identification of suitable embedding parameters, which makes up the core topic of the present thesis work, is the subject of Chapter 3 and 4. While Chapter 3 contains the theoretical outline of the approach, as well as its implementation details, Chapter 4 discusses the application of the approach to benchmark synthetic and experimental sequences, thus illustrating its perks and its limitations. The study of the asymptotic behaviour of the time-dependent divergent exponent is presented in Chapter 5. The alternative estimator of correlation dimension, which relies on this asymptotic metric, is discussed as a possible improvement to the approach described in Chapters 3, 4. The search for chaos out of experimental data is discussed in Chapter 6 by means of two examples of real-world recordings. Concluding remarks are finally drawn in Chapter 7.

The analysis of signals stemming from a physical system is crucial for the experimental investigation of the underlying dynamics that drives the system itself. The field of time series analysis comprises a wide variety of techniques developed with the purpose of characterizing signals and, ultimately, of providing insights on the phenomena that govern the temporal evolution of the generating system. A renowned example in this field is given by spectral analysis: the use of Fourier or Laplace transforms to bring time-domain signals into the more convenient frequency space allows to disclose the key features of linear systems. A more complex scenario turns up when nonlinearity intervenes within a system's dynamics. Nonlinear coupling between a system's degrees of freedom brings about interesting dynamical regimes, such as self-sustained periodic (though anharmonic) oscillations ("limit cycles"), or quasi-periodic evolutions that exhibit sharp spectral lines while lacking strict periodicity ("limit tori"). Among the consequences of nonlinearity, the onset of chaos is definitely the most fascinating one. Chaos is a dynamical regime characterized by unpredictability and lack of periodicity, despite being generated by deterministic laws. Signals generated by chaotic dynamical systems appear as irregular: the corresponding spectra are broad and flat, prediction of future values is challenging, and evolutions within the systems' state spaces converge to strange attractor sets with noninteger dimensionality. Because of these properties, chaotic signals can be mistakenly classified as noise if linear techniques such as spectral analysis are used. The identification of chaos and its characterization require the assessment of dynamical invariants that quantify the complex features of a chaotic system's evolution. For example, Lyapunov exponents provide a marker of unpredictability; the estimation of attractor dimensions, on the other hand, highlights the unconventional geometry of a chaotic system's state space. Nonlinear time series analysis techniques act directly within the state space of the system under investigation. However, experimentally, full access to a system's state space is not always available. Often, only a scalar signal stemming from the dynamical system can be recorded, thus providing, upon sampling, a scalar sequence. Nevertheless, by virtue of a fundamental theorem by Takens, it is possible to reconstruct a proxy of the original state space evolution out of a single, scalar sequence. This reconstruction is carried out by means of the so-called embedding procedure: m-dimensional vectors are built by picking successive elements of the scalar sequence delayed by a lag L. On the other hand, besides posing some necessary conditions on the integer embedding parameters m and L, Takens' theorem does not provide any clue on how to choose them correctly. Although many optimal embedding criteria were proposed, a general answer to the problem is still lacking. As a matter of fact, conventional methods for optimal embedding are flawed by several drawbacks, the most relevant being the need for a subjective evaluation of the outcomes of applied algorithms. Tackling the issue of optimally selecting embedding parameters makes up the core topic of this thesis work. In particular, I will discuss a novel approach that was pursued by our research group and that led to the development of a new method for the identification of suitable embedding parameters. Rather than most conventional approaches, which seek a single optimal value for m and L to embed an input sequence, our approach provides a set of embedding choices that are equivalently suitable to reconstruct the dynamics. The suitability of each embedding choice m, L is assessed by relying on statistical testing, thus providing a criterion that does not require a subjective evaluation of outcomes. The starting point of our method are embedding-dependent correlation integrals, i.e. cumulative distributions of embedding vector distances, built out of an input scalar sequence. In the case of Gaussian white noise, an analytical expression for correlation integrals is available, and, by exploiting this expression, a gauge transformation of distances is introduced to provide a more convenient representation of correlation integrals. Under this new gauge, it is possible to test—in a computationally undemanding way—whether an input sequence is compatible with Gaussian white noise and, subsequently, whether the sequence is compatible with the hypothesis of an underlying chaotic system. These two statistical tests allow ruling out embedding choices that are unsuitable to reconstruct the dynamics. The estimation of correlation dimension, carried out by means of a newly devised estimator, makes up the third stage of the method: sets of embedding choices that provide uniform estimates of this dynamical invariant are deemed to be suitable to embed the sequence.The method was successfully applied to synthetic and experimental sequences, providing new insight into the longstanding issue of optimal embedding. For example, the relevance of the embedding window (m-1)L, i.e. the time span covered by each embedding vector, is naturally highlighted by our approach. In addition, our method provides some information on the adequacy of the sampling period used to record the input sequence.The method correctly distinguishes a chaotic sequence from surrogate ones generated out of it and having the same power spectrum. The technique of surrogate generation, which I also addressed during my Ph. D. work to develop new dedicated algorithms and to analyze brain signals, allows to estimate significance levels in situations where standard analytical algorithms are unapplicable. The novel embedding approach being able to tell apart an original sequence from surrogate ones shows its capability to distinguish signals beyond their spectral—or autocorrelation—similarities.One of the possible applications of the new approach concerns another longstanding issue, namely that of distinguishing noise from chaos. To this purpose, complementary information is provided by analyzing the asymptotic (long-time) behaviour of the so-called time-dependent divergence exponent. This embedding-dependent metric is commonly used to estimate—by processing its short-time linearly growing region—the maximum Lyapunov exponent out of a scalar sequence. However, insights on the kind of source generating the sequence can be extracted from the—usually overlooked—asymptotic behaviour of the divergence exponent. Moreover, in the case of chaotic sources, this analysis also provides a precise estimate of the system's correlation dimension. Besides describing the results concerning the discrimination of chaotic systems from noise sources, I will also discuss the possibility of using the related correlation dimension estimates to improve the third stage of the method introduced above for the identification of suitable embedding parameters. The discovery of chaos as a possible dynamical regime for nonlinear systems led to the search of chaotic behaviour in experimental recordings. In some fields, this search gave plenty of positive results: for example, chaotic dynamics was successfully identified and tamed in electronic circuits and laser-based optical setups. These two families of experimental chaotic systems eventually became versatile tools to study chaos and its possible applications. On the other hand, chaotic behaviour is also looked for in climate science, biology, neuroscience, and even economics. In these fields, nonlinearity is widespread: many smaller units interact nonlinearly, yielding a collective motion that can be described by means of few, nonlinearly coupled effective degrees of freedom. The corresponding recorded signals exhibit, in many cases, an irregular and complex evolution. A possible underlying chaotic evolution—as opposed to a stochastic one—would be of interest both to reveal the presence of determinism and to predict the system's future states. While some claims concerning the existence of chaos in these fields have been made, most results are debated or inconclusive. Nonstationarity, low signal-to-noise ratio, external perturbations and poor reproducibility are just few among the issues that hinder the search of chaos in natural systems. In the final part of this work, I will briefly discuss the problem of chasing chaos in experimental recordings by considering two example sequences, the first one generated by an electronic circuit and the second one corresponding to recordings of brain activity. The present thesis is organized as follows. The core concepts of time series analysis, including the key features of chaotic dynamics, are presented in Chapter 1. A brief review of the search for chaos in experimental systems is also provided; the difficulties concerning this quest in some research fields are also highlighted. Chapter 2 describes the embedding procedure and the issue of optimally choosing the related parameters. Thereupon, existing methods to carry out the embedding choice are reviewed and their limitations are pointed out. In addition, two embedding-dependent nonlinear techniques that are ordinarily used to characterize chaos, namely the estimation of correlation dimension by means of correlation integrals and the assessment of maximum Lyapunov exponent, are presented. The new approach for the identification of suitable embedding parameters, which makes up the core topic of the present thesis work, is the subject of Chapter 3 and 4. While Chapter 3 contains the theoretical outline of the approach, as well as its implementation details, Chapter 4 discusses the application of the approach to benchmark synthetic and experimental sequences, thus illustrating its perks and its limitations. The study of the asymptotic behaviour of the time-dependent divergent exponent is presented in Chapter 5. The alternative estimator of correlation dimension, which relies on this asymptotic metric, is discussed as a possible improvement to the approach described in Chapters 3, 4. The search for chaos out of experimental data is discussed in Chapter 6 by means of two examples of real-world recordings. Concluding remarks are finally drawn in Chapter 7.

Noise and linearity are the two key concerns in RF transceiver systems. However, the impact of circuit topology and device technology on systems noise and linearity behaviors is poorly understood because of the complexity and diversity involved. There are two general questions that are addressed by the RF device and circuit designers: for a given device technology, how best to optimize the circuit topology; and for a given circuit topology, how best to optimize the device technology to improve the noise and linearity performance. In this dissertation, a systematic noise and linearity calculation method is proposed. This approach offers simple and analytical solutions to optimize the noise and linearity characteristics of integrated circuits. Supported by this approach, the physics of state-of-the-art SiGe HBT technology devices can be decoupled and studied. The corresponding impact on noise and linearity is investigated. New optimization methodologies for noise and linearity at both the device and circuit level are presented. In addition, this thesis demonstrates a technique that accurately extracts ac and noise parameters of devices/circuits in the millimeter-wave range. The extraction technique supports and verifies the device/circuit noise analysis from a measurement standpoint.

Microblog has become a central platform where people express their thoughts and opinions toward public events in China. With the sudden outbreak of coronavirus, the posts related to coronavirus are usually followed by a burst immediately in microblog volume, which provides a great opportunity to explore public sentiment about the events. In this context, sentiment analysis is helpful to explore how coronavirus affects public opinions. Deep learning has become a very popular technique for sentiment analysis. This thesis uses Bidirectional Encoder Representations from Transformers (BERT), a pre-trained unsupervised language representation model based on deep learning, to generate initial token embeddings that are further tuned by a neural network model on a supervised corpus, a sentiment classifier is constructed. We utilize data recently made available by the government of Beijing which contains 1 million blog posts from January 1 to February 20, 2020. Also, the model developed in this thesis can be used to track the sentiment variation with Weibo microblog data in the future. At the final stage, the variation of public sentiment is analyzed and presented with visualization charts of preformed people sentiment variation with the development of coronavirus in China. Comparison of the results between labeled data and all data is performed in order to explore how thoughts and opinions evolve in time. The result shows a significant growth of the negative sentiment on January 20 when a lockdown started in Wuhan, and afterward the growth becomes slower. Around February 7 when doctor Wenliang Li died, the number of negative sentiments reached its peak.

L'assistance informatique est devenue une partie indispensable pour la réalisation de procédures chirurgicales modernes. Le désir de créer une nouvelle génération de blocs opératoires intelligents a incité les chercheurs à explorer les problèmes de perception et de compréhension automatique de la situation chirurgicale. Dans ce contexte de prise de conscience de la situation, un domaine de recherche en plein essor adresse la reconnaissance automatique du flux chirurgical. De grands progrès ont été réalisés pour la reconnaissance des phases et des gestes chirurgicaux. Pourtant, il existe encore un vide entre ces deux niveaux de granularité dans la hiérarchie du processus chirurgical. Très peu de recherche se concentre sur les activités chirurgicales portant des informations sémantiques vitales pour la compréhension de la situation. Deux facteurs importants entravent la progression. Tout d'abord, la reconnaissance et la prédiction automatique des activités chirurgicales sont des tâches très difficiles en raison de la courte durée d'une activité, de leur grand nombre et d'un flux de travail très complexe et une large variabilité. Deuxièmement, une quantité très limitée de données cliniques ne fournit pas suffisamment d'informations pour un apprentissage réussi et une reconnaissance précise. À notre avis, avant de reconnaître les activités chirurgicales, une analyse soigneuse des éléments qui composent l'activité est nécessaire pour choisir les bons signaux et les capteurs qui faciliteront la reconnaissance. Nous avons utilisé une approche d'apprentissage profond pour évaluer l'impact de différents éléments sémantiques de l'activité sur sa reconnaissance. Grâce à une étude approfondie, nous avons déterminé un ensemble minimum d'éléments suffisants pour une reconnaissance précise. Les informations sur la structure anatomique et l'instrument chirurgical sont de première importance. Nous avons également abordé le problème de la carence en matière de données en proposant des méthodes de transfert de connaissances à partir d'autres domaines ou chirurgies. Les méthodes de ''word embedding'' et d'apprentissage par transfert ont été proposées. Ils ont démontré leur efficacité sur la tâche de prédiction d'activité suivante offrant une augmentation de précision de 22%. De plus, des observations pertinentes
Computer assistance became indispensable part of modern surgical procedures. Desire of creating new generation of intelligent operating rooms incited researchers to explore problems of automatic perception and understanding of surgical situations. Situation awareness includes automatic recognition of surgical workflow. A great progress was achieved in recognition of surgical phases and gestures. Yet, there is still a blank between these two granularity levels in the hierarchy of surgical process. Very few research is focused on surgical activities carrying important semantic information vital for situation understanding. Two important factors impede the progress. First, automatic recognition and prediction of surgical activities is a highly challenging task due to short duration of activities, their great number and a very complex workflow with multitude of possible execution and sequencing ways. Secondly, very limited amount of clinical data provides not enough information for successful learning and accurate recognition. In our opinion, before recognizing surgical activities a careful analysis of elements that compose activity is necessary in order to chose right signals and sensors that will facilitate recognition. We used a deep learning approach to assess the impact of different semantic elements of activity on its recognition. Through an in-depth study we determined a minimal set of elements sufficient for an accurate recognition. Information about operated anatomical structure and surgical instrument was shown to be the most important. We also addressed the problem of data deficiency proposing methods for transfer of knowledge from other domains or surgeries. The methods of word embedding and transfer learning were proposed. They demonstrated their effectiveness on the task of next activity prediction offering 22% increase in accuracy. In addition, pertinent observations about the surgical practice were made during the study. In this work, we also addressed the problem of insufficient and improper validation of recognition methods. We proposed new validation metrics and approaches for assessing the performance that connect methods to targeted applications and better characterize capacities of the method. The work described in this these aims at clearing obstacles blocking the progress of the domain and proposes a new perspective on the problem of surgical workflow recognition

Sentiment analysis aims at figuring out the opinions of the users towards a certain service or product. In this research, the aim is at classifying the sentiments of users based on the comments they have posed on Douban movie website. In this thesis, I try two different ways to classify the sentiments: with the first one classifying comments into five classes of ratings from 1 to 5, and with the second one classifying comments into three classes of ratings: negative, neutral and positive. For the latter, the ratings of 1 and 2 are grouped as negative, the ratings of 3 neutral and the ratings of 4 and 5 positive. First, Term Frequency Inverse Document Frequency (TF-IDF) is used as the feature extraction technique for machine learning algorithms. Chi Square and Mutual Information are used for feature selection. The selected features are fed into different machine learning methods: Logistic Regression, Linear SVC, SGD classifier and Multinomial Naive Bayes. The performance of models with feature selection will be compared with the performance of models without feature selection for 5-class classification as well as 3-class classification. Also, fastText and Skip-Gram are used as embedding methods for deep learning algorithms LSTM and BILSTM. FastText will also be used for both embedding as well as being a classifier. The aim is to compare different machine learning and deep learning algorithms using different vectorization methods to see which model performs the best regarding both 5-class and 3-class classification. The two classification strategies will be compared with each other in terms of error analysis. The aim is to figure out the similarities and differences of misclassifications made by two different classification strategies.

The use of mobile devices to support students’ learning experiences is a growing area of interest in higher education (Wankel & Blessinger, 2013). This study adopts an ‘umbrella’ term of m-learning to consider the use of mobile and wireless technologies to support students in a blended learning environment. Whilst m-learning pedagogy has received considerable attention (e.g. Attewell, 2005, Sharples et. al. 2007, Kukulska-Hulme, 2012), the process of adopting this potentially disruptive innovation within universities has been neglected. This study addresses this gap by attempting to answer the research question: How do university organizations (business models, modes of operation, people and processes) adapt to a potentially disruptive innovation like m-learning and what factors and working practices support or hinder embedding? Possible frameworks for studying innovation are reviewed, including Rogers’ innovation diffusion framework (Rogers, 1962), Actor-Network Theory (Latour, 2005) , Activity Theory ngestr m 1987), Structuration Theory (Giddens, 1984), theories of disruptive innovation (Christensen, 1997) and the Technology Acceptance Model (Venkatesh and Davis, 2000), Actor-Network Theory (ANT) is chosen as the most promising theoretical lens for an in-depth investigation of m-learning embedding, and a participative fieldwork approach is developed that uses Law and Callon’s ANT notion of ‘points of passage’ between local and global networks (Law and Callon, 1991) to illuminate factors and working practices that affect embedding. A framework based on Law and Callon’s work is developed through a year-long study of competing text messaging projects within a university and developed further through a three-year, longitudinal case study involving five universities using smartphone applications to assess students in medical practice situations. Several institutional issues are identified that help or hinder embedding, such as fragmentation of IT strategy and decision-making, and the need to provide students with a compelling offer of multiple institutional services on their mobiles. The role of people and artefacts in forming a link, or ‘point of passage’ between m-learning projects ‘local networks’) and institutional IT strategies and services ‘global networks’) is found to be of central interest for understanding processes of embedding. A clear path to an ANT analysis is demonstrated starting from interview and observation data, using coding techniques borrowed from grounded theory (Schatzman and Strauss, 1973) and finishing with Law and Callon’s local-global network model, which is used to compare and contrast embedding trajectories of the case study institutions. Systematic comparison enables a three dimensional model of embedding trajectories to be built, which extends Law and Callon’s work and places in sharper focus the importance of establishing a path by which local initiatives can be evaluated strategically and, where appropriate, incorporated in a timely manner into a university’s IT strategy. Grounded in extensive longitudinal research, the study offers a contribution to methodology through its demystification of ANT; a contribution to theory through its three dimensional model for mapping embedding trajectories; and a contribution to practice by highlighting specific issues that affect mobile technology adoption in higher education, such as having a compelling, multi-service offer, appropriate mobile tariffs for undertaking mandatory assessment and guidelines for incorporating knowledge gained from technology experiments into institutional strategies and decision-making. The study concludes by highlighting opportunities for using its model to explore challenges of embedding faced not only by formal projects but by ‘maverick’ innovators and for potentially disruptive technologies other than m-learning, such as Web 2.0 services.

High-dimensional data take place from digital image processing, gene expression micro arrays, neuronal population activities to financial time series. Dimensionality Reduction - extracting low dimensional structure from high dimension - is a key problem in many areas like information processing, machine learning, data mining, information retrieval and pattern recognition, where we find some data reduction techniques. In this thesis we will give a survey about modern data reduction techniques, representing the state-of-the-art of theory, methods and application, by introducing the language of mathematics there. This needs a special care concerning the questions of, e.g., how to understand discrete structures as manifolds, to identify their structure, preparing the dimension reduction, and to face complexity in the algorithmically methods. A special emphasis will be paid to Principal Component Analysis, Locally Linear Embedding and Isomap Algorithms. These algorithms are studied by a research group from Vilnius, Lithuania and Zeev Volkovich, from Software Engineering Department, ORT Braude College of Engineering, Karmiel, and others. The main purpose of this study is to compare the results of the three of the algorithms. While the comparison is beeing made we will focus the results and duration.

This work aims at combining the Chaos theory postulates and Artificial Neural Networks classification and predictive capability, in the field of financial time series prediction. Chaos theory, provides valuable qualitative and quantitative tools to decide on the predictability of a chaotic system. Quantitative measurements based on Chaos theory, are used, to decide a-priori whether a time series, or a portion of a time series is predictable, while Chaos theory based qualitative tools are used to provide further observations and analysis on the predictability, in cases where measurements provide negative answers. Phase space reconstruction is achieved by time delay embedding resulting in multiple embedded vectors. The cognitive approach suggested, is inspired by the capability of some chartists to predict the direction of an index by looking at the price time series. Thus, in this work, the calculation of the embedding dimension and the separation, in Takens‘ embedding theorem for phase space reconstruction, is not limited to False Nearest Neighbor, Differential Entropy or other specific method, rather, this work is interested in all embedding dimensions and separations that are regarded as different ways of looking at a time series by different chartists, based on their expectations. Prior to the prediction, the embedded vectors of the phase space are classified with Fuzzy-ART, then, for each class a back propagation Neural Network is trained to predict the last element of each vector, whereas all previous elements of a vector are used as features.

A central operation of users of the text analysis tool Gavagai Explorer is to look through a list of words and arrange them in groups. This thesis explores the use of word clustering to automatically arrange the words in groups intended to help users. A new word clustering algorithm is introduced, which attempts to produce word clusters tailored to be small enough for a user to quickly grasp the common theme of the words. The proposed algorithm computes similarities among words using word embeddings, and clusters them using hierarchical graph clustering. Multiple variants of the algorithm are evaluated in an unsupervised manner by analysing the clusters they produce when applied to 110 data sets previously analysed by users of Gavagai Explorer. A supervised evaluation is performed to compare clusters to the groups of words previously created by users of Gavagai Explorer. Results show that it was possible to choose a set of hyperparameters deemed to perform well across most data sets in the unsupervised evaluation. These hyperparameters also performed among the best on the supervised evaluation. It was concluded that the choice of word embedding and graph clustering algorithm had little impact on the behaviour of the algorithm. Rather, limiting the maximum size of clusters and filtering out similarities between words had a much larger impact on behaviour.

Ductal carcinoma in situ (DCIS); is thought directly to precede invasive breast cancer (IBC). Screening mammography has driven the incidence of this key precursor lesion to >65,000 cases per year. However, little is known about the factors controlling the natural history or risk for recurrence following treatment of a particular patients DCIS. Though the heterogeneity of the disease is well established, no histologic or demographic criteria have been able to stratify DCIS for treatment. We hypothesize that at initial diagnosis there exist biologically distinct subsets of DCIS with associated prognoses that may be recognized by molecular markers. Molecular approaches have been limited by technical design issues related to the types of tissue available for analysis, namely degraded formalin-fixed paraffin embedded (FFPE) specimens and small core biopsy samples. However, new technologies promise to overcome these issues. In the first phase of our investigation, we aimed a) to pilot feasibility studies on the use of FFPE DCIS for molecular analyses including gene expression microarray and b) to pilot feasibility study of selective, high throughput sequencing through the use of "exon capture" on small input material that simulated expected DCIS core biopsy amounts. The results of this work offer specific technical guidelines for the molecular study of DCIS. Moreover, they have enabled the initiation of the second phase of this study, which aims to assess molecular profiles of DCIS recurrence and progression.

The fundamental theorem of algebra (FTA) is an important theorem in algebra. This theorem asserts that the complex field is algebracially closed. This thesis will include historical research of proofs of the fundamental theorem of algebra and provide information about the first proof given by Gauss of the theorem and the time when it was proved.

The aim of this thesis is to explore and develop measures for image comparison in two main areas of medical image analysis: image registration and manifold learning for population analysis. In particular, the main contributions of the thesis are the following: (i) the development of a multimodal similarity measure using the commutativity of the image graph Laplacians as a criterion of image structure preservation, (ii) the application of such similarity measure to fetal ultrasound-magnetic resonance registration, (iii) the development of a framework for optimal kernel-based manifold embeddings for medical image data, and (iv) the development of a method to learn and combine heterogeneous pairwise image similarities induced by application-specific distance functions for manifold learning. The methods developed in this thesis were evaluated both on synthetic and clinical data, here in particular for brain analysis and classification, both during early childhood and in aging adults.
El objetivo de esta tesis es explorar y desarrollar métodos para la comparación de imágenes en dos de las principales áreas del análisis de imagen: el registro de imágenes y manifold learning por el análisis de población. En particular, las contribuciones pueden resumirse principalmente en tres: (i) se propone una medida de similitud multimodal para el registro de imágenes usando la commutatividad del grafo Laplaciano como criterio para preservar la estructura de la imagen, (ii) la aplicación de esta medida de similitud para el registro de resonancias magnética-ultrasonido de feto, (iii) se presenta un método para un optimo kernel-based manifold learning de imágenes médicas, y (iv) se propone aprender y combinar diferentes aproximaciones de vecindarios de las imágenes por manifold learning. Los métodos desarrollados en este tesis fueron evaluados en datos sintéticos y clínicos, en particular por el análisis de imágenes cerebrales de fetos, neonatos y adultos.

Diese Dissertation präsentiert neue Techniken der Integration für verschiedene Probleme der Finanzmathematik und einige Anwendungen in der Wahrscheinlichkeitstheorie. Zu Beginn entwickeln wir zwei Zugänge zur robusten stochastischen Integration. Der erste, ähnlich der Ito’schen Integration, basiert auf einer Topologie, erzeugt durch ein äußeres Maß, gegeben durch einen minimalen Superreplikationspreis. Der zweite gründet auf der Integrationtheorie für rauhe Pfade. Wir zeigen, dass das entsprechende Integral als Grenzwert von nicht antizipierenden Riemannsummen existiert und dass sich jedem "typischen Preispfad" ein rauher Pfad im Ito’schen Sinne zuordnen lässt. Für eindimensionale "typische Preispfade" wird sogar gezeigt, dass sie Hölder-stetige Lokalzeiten besitzen. Zudem erhalten wir Verallgemeinerungen von Föllmer’s pfadweiser Ito-Formel. Die Integrationstheorie für rauhe Pfade kann mit dem Konzept der kontrollierten Pfade und einer Topologie, welche die Information der Levy-Fläche enthält, entwickelt werden. Deshalb untersuchen wir hinreichende Bedingungen an die Kontrollstruktur für die Existenz der Levy-Fläche. Dies führt uns zur Untersuchung von Föllmer’s Ito-Formel aus der Sicht kontrollierter Pfade. Para-kontrollierte Distributionen, kürzlich von Gubinelli, Imkeller und Perkowski eingeführt, erweitern die Theorie rauher Pfade auf den Bereich von mehr-dimensionale Parameter. Wir verallgemeinern diesen Ansatz von Hölder’schen auf Besov-Räume, um rauhe Differentialgleichungen zu lösen, und wenden die Ergebnisse auf stochastische Differentialgleichungen an. Zum Schluß betrachten wir stark gekoppelte Systeme von stochastischen Vorwärts-Rückwärts-Differentialgleichungen (FBSDEs) und erweitern die Theorie der Existenz, Eindeutigkeit und Regularität der sogenannten Entkopplungsfelder auf Markovsche FBSDEs mit lokal Lipschitz-stetigen Koeffizienten. Als Anwendung wird das Skorokhodsche Einbettungsproblem für Gaußsche Prozesse mit nichtlinearem Drift gelöst.
In this thesis new robust integration techniques, which are suitable for various problems from stochastic analysis and mathematical finance, as well as some applications are presented. We begin with two different approaches to stochastic integration in robust financial mathematics. The first one is inspired by Ito’s integration and based on a certain topology induced by an outer measure corresponding to a minimal superhedging price. The second approach relies on the controlled rough path integral. We prove that this integral is the limit of non-anticipating Riemann sums and that every "typical price path" has an associated Ito rough path. For one-dimensional "typical price paths" it is further shown that they possess Hölder continuous local times. Additionally, we provide various generalizations of Föllmer’s pathwise Ito formula. Recalling that rough path theory can be developed using the concept of controlled paths and with a topology including the information of Levy’s area, sufficient conditions for the pathwise existence of Levy’s area are provided in terms of being controlled. This leads us to study Föllmer’s pathwise Ito formulas from the perspective of controlled paths. A multi-parameter extension to rough path theory is the paracontrolled distribution approach, recently introduced by Gubinelli, Imkeller and Perkowski. We generalize their approach from Hölder spaces to Besov spaces to solve rough differential equations. As an application we deal with stochastic differential equations driven by random functions. Finally, considering strongly coupled systems of forward and backward stochastic differential equations (FBSDEs), we extend the existence, uniqueness and regularity theory of so-called decoupling fields to Markovian FBSDEs with locally Lipschitz continuous coefficients. These results allow to solve the Skorokhod embedding problem for a class of Gaussian processes with non-linear drift.

Time domain analysis in vector network analyzers (VNAs) is a method to represent the frequency response, stated by the S-parameters, in time domain with apparent high resolution. Among other utilities time domain option from Agilent allows to measure microwave devices into a specific frequency range and down till DC as well with the two time domain mode: band-pass and low-pass mode. A special feature named gating is of important as it allows representing a portion of the time domain representation in frequency domain.   This thesis studies the time domain option 010 from Agilent; its uncertainties and sensitivity. The task is to find the best method to measure the antenna element matching taking care to reduce the influence of measurement errors on the results.   The Agilent 8753ES is the instrument used in the thesis. A specific matching problem in the antenna electric down-tilt (AEDT) previously designed by Powerwave Technologies is the task to be solved. This is because it can not be measured directly with 2-port VNAs. It requires adapters, extra coaxial cables and N-connectors, all of which influences the accuracy. The AEDT connects to the array antenna through cable-board-connectors (CBCs). The AEDT and the CBCs were designed before being put into the antenna-system. Their S-parameters do not coincide with the ones measured after these devices were put in the antenna block.   Time domain gating and de-embedding algorithms are two methods proposed in this thesis to measure the S-parameters of the desired antenna element while reducing the influence of measurement errors due to cables CBCs and other connectors. The aim is to find a method which causes less error and gives high confidence measurements.   For the time domain analysis, reverse engineering of the time domain option used in the Agilent VNA 8753ES is implemented in a PC for full control of the process. The results using time-domain are not sufficiently reliable to be used due to the multiple approximations done in the design. The methodology that Agilent uses to compensate the gating effects is not reliable when the gate is not centered on the analyzed response. Big errors are considered due to truncation and masking effects in the frequency response.   The de-embedding method using LRL is implemented in the AEDT measurements, taking away the influences of the CBCs, coaxial cables and N-connector. It is found to have sufficient performance, comparable to the mathematical model. Error analysis of both methods has been done to explaine the different in measurements and design.

Design space exploration (DSE) is a tool used to evaluate and compare designs as part of the design selection process. While evaluating every possible design in a design space is infeasible, understanding design behavior and response throughout the design space may be accomplished by evaluating a subset of designs and interpolating between them using surrogate models. Surrogate modeling is a technique that uses low-cost calculations to approximate the outcome of more computationally expensive calculations or analyses, such as finite element analysis (FEA). While surrogates make quick predictions, accuracy is not guaranteed and must be considered. This research addressed the need to improve the accuracy of surrogate predictions in order to improve DSE of structural responses. This was accomplished by performing comparative analyses of linear and nonlinear dimensionality-reduction-based radial basis function (RBF) surrogate models for emulating various FEA nodal results. A total of four dimensionality reduction methods were investigated, namely principal component analysis (PCA), kernel principal component analysis (KPCA), isometric feature mapping (ISOMAP), and locally linear embedding (LLE). These methods were used in conjunction with surrogate modeling to predict nodal stresses and coordinates of a compressor blade. The research showed that using an ISOMAP-based dual-RBF surrogate model for predicting nodal stresses decreased the estimated mean error of the surrogate by 35.7% compared to PCA. Using nonlinear dimensionality-reduction-based surrogates did not reduce surrogate error for predicting nodal coordinates. A new metric, the manifold distance ratio (MDR), was introduced to measure the nonlinearity of the data manifolds. When applied to the stress and coordinate data, the stress space was found to be more nonlinear than the coordinate space for this application. The upfront training cost of the nonlinear dimensionality-reduction-based surrogates was larger than that of their linear counterparts but small enough to remain feasible. After training, all the dual-RBF surrogates were capable of making real-time predictions. This same process was repeated for a separate application involving the nodal displacements of mode shapes obtained from a FEA modal analysis. The modal assurance criterion (MAC) calculation was used to compare the predicted mode shapes, as well as their corresponding true mode shapes obtained from FEA, to a set of reference modes. The research showed that two nonlinear techniques, namely LLE and KPCA, resulted in lower surrogate error in the more complex design spaces. Using a RBF kernel, KPCA achieved the largest average reduction in error of 13.57%. The results also showed that surrogate error was greatly affected by mode shape reversal. Four different approaches of identifying reversed mode shapes were explored, all of which resulted in varying amounts of surrogate error. Together, the methods explored in this research were shown to decrease surrogate error when performing DSE of a turbomachine compressor blade. As surrogate accuracy increases, so does the ability to correctly make engineering decisions and judgements throughout the design process. Ultimately, this will help engineers design better turbomachines.

Milioni di commenti ogni giorno vengono scritti ed aggiunti su internet da migliaia di utenti, essi riguardano ristoranti, hotel, acquisti, luoghi da visitare e molto altro. Quotidianamente moltissime persone fano affidamento su tali opinioni, ad esempio preferendo l’acquisto di un prodotto rispetto ad un altro su un sito di e-commerce, rendendo così il valore di questi dati molto elevato ed una loro analisi un processo molto importante da effettuare. Estrarre le informazioni di interesse da essi è un procedimento difficilmente eseguibile da operatori umani data la loro enorme quantità, di conseguenza sarà necessario trovare dei sistemi informatici in grado di eseguire questo processo. Questo elaborato propone un’infrastruttura per eseguire il processo comunemente chiamato Aspect Based Sentiment Analysis (ABSA), il cui scopo finale sarà, dato un commento, estrarne le categorie da esso affrontate con le relative polarità. Ad esempio dal commento “buona la pizza ma il conto era alto” le categorie da estrarre potrebbero essere “gusto” e “prezzo” aventi rispettivamente polarità “positiva” e “negativa”. La nostra infrastruttura permetterà di personalizzare agilmente le sue componenti interne, lasciando all’utente la possibilità di aggiungere soluzioni proprie o di utilizzare quelle già implementate. Verranno inoltre sfruttati attivamente diversi servizi della Google Cloud Platform (GCP) grazie anche alla collaborazione con Injenia S.r.l., l’azienda dove è stato svolto il lavoro per questo elaborato. Il sistema è stato testato su due data set, uno dei quali preso dalla competizione italiana ABSITA, conclusa nel dicembre 2018. Le prestazioni su quest’ultimo data set hanno fornito un punteggio di F1-Score pari a 0.694, senza l’utilizzo di tecnologie specifiche al data set e ottimizzazione dei parametri. Tale punteggio avrebbe rappresentato un quinto posto alla competizione ABSITA.

This thesis proposes a method to reconstruct a frontal facial video basedon encoding done with the facial profile of another video sequence.The reconstructed facial video will have the similar facial expressionchanges as the changes in the profile video. First, the profiles for boththe reference video and for the test video are captured by edge detection.Then, asymmetrical principal component analysis is used to model thecorrespondence between the profile and the frontal face. This allows en-coding from a profile and decoding of the frontal face of another video.Another solution is to use dynamic time warping to match the profilesand select the best matching corresponding frontal face frame for re-construction. With this method, we can reconstructed the test frontalvideo to make it have the similar changing in facial expressions as thereference video. To improve the quality of the result video, Local Lin-ear Embedding is used to give the result video a smoother transitionbetween frames.

Le Web est en croissance continue, et une quantité énorme de données est générée par les réseaux sociaux, permettant aux utilisateurs d'échanger une grande diversité d'informations. En outre, les textes au sein des réseaux sociaux sont souvent subjectifs. L'exploitation de cette subjectivité présente au sein des textes peut être un facteur important lors d'une recherche d'information. En particulier, cette thèse est réalisée pour répondre aux besoins de la plate-forme Books de Open Edition en matière d'amélioration de la recherche et la recommandation de livres, en plusieurs langues. La plateforme offre des informations générées par des utilisateurs, riches en sentiments. Par conséquent, l'analyse précédente, concernant l'exploitation de sentiment en recherche d'information, joue un rôle important dans cette thèse et peut servir l'objectif d'une amélioration de qualité de la recherche de livres en utilisant les informations générées par les utilisateurs. Par conséquent, nous avons choisi de suivre une voie principale dans cette thèse consistant à combiner les domaines analyse de sentiment (AS) et recherche d'information (RI), dans le but d'améliorer les suggestions de la recherche de livres. Nos objectifs peuvent être résumés en plusieurs points: • Une approche d'analyse de sentiment, facilement applicable sur différentes langues, peu coûteuse en temps et en données annotées. • De nouvelles approches pour l'amélioration de la qualité lors de la recherche de livres, basées sur l'utilisation de l'analyse de sentiment dans le filtrage, l'extraction et la classification des informations
The web technology is in an on going growth, and a huge volume of data is generated in the social web, where users would exchange a variety of information. In addition to the fact that social web text may be rich of information, the writers are often guided by provoked sentiments reflected in their writings. Based on that concept, locating sentiment in a text can play an important role for information extraction. The purpose of this thesis is to improve the book search and recommendation quality of the Open Edition's multilingual Books platform. The Books plat- form also offers additional information through users generated information (e.g. book reviews) connected to the books and rich in emotions expressed in the users' writings. Therefore, the previous analysis, concerning locating sentiment in a text for information extraction, plays an important role in this thesis, and can serve the purpose of quality improvement concerning book search, using the shared users generated information. Accordingly, we choose to follow a main path in this thesis to combine sentiment analysis (SA) and information retrieval (IR) fields, for the purpose of improving the quality of book search. Two objectives are summarised in the following, which serve the main purpose of the thesis in the IR quality improvement using SA: • An approach for SA prediction, easily applicable on different languages, low cost in time and annotated data. • New approaches for book search quality improvement, based on SA employment in information filtering, retrieving and classifying

This thesis describes a structural health monitoring (SHM) strategy to detect and classify changes in structures that can be equipped with sensors. SHM is an area of great interest, because its main objective is to verify the health of the structure to ensure its correct operation and, in turn, save maintenance costs. This objective is achieved by using algorithms and equipping the structure with a network of sensors that continuously monitor it. Researchers from around the world focus their efforts on the development of new forms of continuous monitoring to know the current state of the structure and to avoid possible failures or catastrophes. In this sense, in this work, a network of piezoelectric sensors (PZTs) is used for the development of the strategy of detection and classification of structural changes. This network of PZTs, attached to the surface of the structure to be diagnosed, applies vibrational excitation signals and, at the same time, collects the responses propagated through the structure. With this collected information, certain mathematical algorithms are developed. To carry out the main task of the proposed methodology, detection and classification of structural changes, the technique called t-distributed stochastic neighbor embedding (t-SNE) is essentially used. This technique is capable of representing the local structure of the high-dimensional data collected by the sensor network in two-dimensional or three-dimensional space. Furthermore, for the classification of structural changes, the detection methodology is expanded by adding the use of three strategies: (a) the smallest point-centroid distance; (b) the majority vote; and (c) the sum of the inverses of the distances. The methodology proposed in this study is tested and validated using an aluminum plate equipped with four PZT sensors and for certain predefined structural changes. The promising results obtained show the great classification capacity and the strong performance of this methodology, successfully classifying about 100% of the cases in various experimental scenarios. The main contribution of this project is the combination of the t-SNE technique with a carefully selected pre-processing of the data and with the three proposed classification strategies. This combination significantly improves the quality of the groups or clusters obtained with the damage detection and classification method, which represent the different structural states. Likewise, said combination diagnoses a structure with a low computational cost and high reliability. Regarding the applicability of the suggested strategy, there is no prescribed field of application: if a network of sensors can be installed in the structure to be diagnosed and several phases of action can be considered, the approach presented here can be, a priori, implemented.
Esta tesis describe una estrategia de monitorización de la salud estructural (SHM, por sus siglas en inglés) para detectar y clasificar fallos en estructuras que pueden ser equipadas con sensores. La SHM es un área de gran interés, ya que su objetivo principal es la verificación de la salud de la estructura para asegurar su correcto funcionamiento y, a su vez, ahorrar costes de mantenimiento. Este objetivo se consigue haciendo uso de algoritmos y equipando a la estructura con una red de sensores que la monitorizan de forma continuada. Investigadores de todo el mundo centran sus esfuerzos en el desarrollo de nuevas formas de monitorización continua para conocer el estado actual de la estructura y evitar posibles fallos o catástrofes. En este sentido, en este trabajo, se utiliza una red de sensores piezoeléctricos (PZT, por sus siglas en inglés) para el desarrollo de la estrategia de detección y clasificación de los cambios estructurales. Esta red de PZT, adherida a la superficie de la estructura a diagnosticar, aplica señales vibracionales de excitación y al mismo tiempo recoge las respuestas propagadas a través de la estructura. Con esta información recopilada se desarrollan ciertos algoritmos matemáticos. Para llevar a cabo la tarea principal de la metodología propuesta, detección y clasificación de fallos, se utiliza esencialmente la técnica denominada t-distributed stochastic neighbor embedding (t-SNE). Dicha técnica es capaz de representar la estructura local de los datos de alta dimensionalidad recopilados por la red de sensores en un espacio bidimensional o tridimensional. Además, para la clasificación de los cambios estructurales, se amplía la metodología de detección añadiendo el uso de tres estrategias: (a) la distancia punto-centroide más pequeña; (b) el voto mayoritario; y (c) la suma de las inversas de las distancias. La metodología propuesta en este estudio se prueba y valida utilizando una placa de aluminio equipada con cuatro sensores PZT y para ciertos daños predefinidos. Los prometedores resultados obtenidos ponen de manifiesto la gran capacidad de clasificación y el fuerte rendimiento de esta metodología, clasificando con éxito cerca del 100% de los casos en varios escenarios experimentales. La principal contribución de este proyecto es la combinación de la técnica t-SNE con un preprocesamiento de los datos cuidosamente seleccionado y con las tres estrategias de clasificación propuestas. Esta combinación mejora significativamente la calidad de los grupos o clústeres obtenidos con el método de detección y clasificación de daños, que representan los diferentes estados estructurales. Asimismo, dicha combinación diagnostica una estructura con un bajo coste computacional y una alta fiabilidad. En cuanto a la aplicabilidad de la estrategia sugerida, no hay un campo de aplicación prescrito: si se puede instalar una red de sensores en la estructura a diagnosticar y se pueden considerar varias fases de actuación, el enfoque aquí presentado puede implementarse a priori.

The main result in this thesis is the generalisation of Bergman's diamond lemma for ring theory to power series rings. This generalisation makes it possible to treat problems in which there arise infinite descending chains. Several results in the literature are shown to be special cases of this diamond lemma and examples are given of interesting problems which could not previously be treated. One of these examples provides a general construction of a normed skew field in which a custom commutation relation holds.

There is also a general result on the structure of totally ordered semigroups, demonstrating that all semigroups with an archimedean element has a (up to a scaling factor) unique order-preserving homomorphism to the real numbers. This helps analyse the concept of filtered structure. It is shown that whereas filtered structures can be used to induce pretty much any zero-dimensional linear topology, a real-valued norm suffices for the definition of those topologies that have a reasonable relation to the multiplication operation.

The thesis also contains elementary results on degree (as of polynomials) functions, norms on algebras (in particular ultranorms), (Birkhoff) orthogonality in modules, and construction of semigroup partial orders from ditto quasiorders.

Submitted by JOSIANE SANTOS DE OLIVEIRA (josianeso) on 2019-03-07T14:41:05Z No. of bitstreams: 1 Roger Alan Stein_.pdf: 476239 bytes, checksum: a87a32ffe84d0e5d7a882e0db7b03847 (MD5)
Made available in DSpace on 2019-03-07T14:41:05Z (GMT). No. of bitstreams: 1 Roger Alan Stein_.pdf: 476239 bytes, checksum: a87a32ffe84d0e5d7a882e0db7b03847 (MD5) Previous issue date: 2018-03-28
CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
Efficient distributed numerical word representation models (word embeddings) combined with modern machine learning algorithms have recently yielded considerable improvement on automatic document classification tasks. However, the effectiveness of such techniques has not been assessed for the hierarchical text classification (HTC) yet. This study investigates application of those models and algorithms on this specific problem by means of experimentation and analysis. Classification models were trained with prominent machine learning algorithm implementations—fastText, XGBoost, and Keras’ CNN—and noticeable word embeddings generation methods—GloVe, word2vec, and fastText—with publicly available data and evaluated them with measures specifically appropriate for the hierarchical context. FastText achieved an LCAF1 of 0.871 on a single-labeled version of the RCV1 dataset. The results analysis indicates that using word embeddings is a very promising approach for HTC.
Modelos eficientes de representação numérica textual (word embeddings) combinados com algoritmos modernos de aprendizado de máquina têm recentemente produzido uma melhoria considerável em tarefas de classificação automática de documentos. Contudo, a efetividade de tais técnicas ainda não foi avaliada com relação à classificação hierárquica de texto. Este estudo investiga a aplicação daqueles modelos e algoritmos neste problema em específico através de experimentação e análise. Modelos de classificação foram treinados usando implementações proeminentes de algoritmos de aprendizado de máquina—fastText, XGBoost e CNN (Keras)— e notórios métodos de geração de word embeddings—GloVe, word2vec e fastText—com dados disponíveis publicamente e avaliados usando métricas especificamente adequadas ao contexto hierárquico. Nesses experimentos, fastText alcançou um LCAF1 de 0,871 usando uma versão da base de dados RCV1 com apenas uma categoria por tupla. A análise dos resultados indica que a utilização de word embeddings é uma abordagem muito promissora para classificação hierárquica de texto.

Bone is a common site for metastases and spine represent the most frequent site. Lytic lesions are associated with the loss of bone tissue, which can compromise the mechanical competence of the vertebra, leading to spine instability. Rigid stabilization is a solution, but it is a complex surgery, that can be very critical for oncologic patients; on the other hand, an untreated metastasis can lead to mechanical failure of the bone, leading to pain, immobilization and in the worst case, paralysis. In this study, a protocol to analyse the strain with simulated lytic metastasis under compressive loading has been developed and optimized using a porcine vertebra. The strain distribution has been measured experimentally using micro-computed tomography (micro-CT) and Digital Volume Correlation (DVC), which provided three-dimensional displacements and strains maps inside the specimen. The ideal parameters for the DVC have been found by analysing two repeated scans in constant strain condition and setting a target of 200 microstrain for the errors (one order of magnitude lower than typical strains in bone subjected to physiological loading conditions). An ideal nodal spacing of 50 voxels (approximately 2 mm) has been chosen and a voxel detection algorithm has been applied to all data to remove regions outside the bone. In order to understand how the presence of the defect could alter the strain distribution, the porcine vertebra has also been subjected to non-destructive compressive load before and after the preparation of a mechanically induced lytic metastasis in the vertebral body. An increase of the 40% of the compressive principal strain after the defect has been found in proximity of the lesion. This protocol will be used in future studies to analyse the effect of size and position of artificially metastatic lesions in the vertebral body of human spines.

This thesis addresses a challenging problem: “how to let the visually impaired ‘see’ others emotions”. We, human beings, are heavily dependent on facial expressions to express ourselves. A smile shows that the person you are talking to is pleased, amused, relieved etc. People use emotional information from facial expressions to switch between conversation topics and to determine attitudes of individuals. Missing emotional information from facial expressions and head gestures makes the visually impaired extremely difficult to interact with others in social events. To enhance the visually impaired’s social interactive ability, in this thesis we have been working on the scientific topic of ‘expressing human emotions through vibrotactile patterns’. It is quite challenging to deliver human emotions through touch since our touch channel is very limited. We first investigated how to render emotions through a vibrator. We developed a real time “lipless” tracking system to extract dynamic emotions from the mouth and employed mobile phones as a platform for the visually impaired to perceive primary emotion types. Later on, we extended the system to render more general dynamic media signals: for example, render live football games through vibration in the mobile for improving mobile user communication and entertainment experience. To display more natural emotions (i.e. emotion type plus emotion intensity), we developed the technology to enable the visually impaired to directly interpret human emotions. This was achieved by use of machine vision techniques and vibrotactile display. The display is comprised of a ‘vibration actuators matrix’ mounted on the back of a chair and the actuators are sequentially activated to provide dynamic emotional information. The research focus has been on finding a global, analytical, and semantic representation for facial expressions to replace state of the art facial action coding systems (FACS) approach. We proposed to use the manifold of facial expressions to characterize dynamic emotions. The basic emotional expressions with increasing intensity become curves on the manifold extended from the center. The blends of emotions lie between those curves, which could be defined analytically by the positions of the main curves. The manifold is the “Braille Code” of emotions. The developed methodology and technology has been extended for building assistive wheelchair systems to aid a specific group of disabled people, cerebral palsy or stroke patients (i.e. lacking fine motor control skills), who don’t have ability to access and control the wheelchair with conventional means, such as joystick or chin stick. The solution is to extract the manifold of the head or the tongue gestures for controlling the wheelchair. The manifold is rendered by a 2D vibration array to provide user of the wheelchair with action information from gestures and system status information, which is very important in enhancing usability of such an assistive system. Current research work not only provides a foundation stone for vibrotactile rendering system based on object localization but also a concrete step to a new dimension of human-machine interaction.
Taktil Video

Die Präsentation stellt kurz die Grundlagen der Berechnung von Schraubenverbindungen in Anlehnung an die VDI-Richtlinie 2230 Teil 1 dar. Auch die vier FEM-Modellklassen, die die VDI 2230 Teil 2 (Entwurf) zur Berechnung von Mehrschraubenverbindungen vorschlägt, werden behandelt, und die in Creo Simulate vorhandenen Softwarefeatures zu deren Umsetzung vorgestellt. Es folgt eine Darstellung, was bei der Linearisierung von Schraubenverbindungen zur vereinfachten Berechnung zu beachten ist, und wieso bei der Berechnung im FEM-System dann nicht notwendigerweise eine Vorspannung benötigt wird. Ausführlich wird das neue Schraubenfeature in Creo Simulate betrachtet, das eine weitgehend automatisierte Modellierung und Berechnung von Standardverschraubungen erlaubt. Weitere Features, wie die neuen Vorspannelemente, werden erläutert, sowie auch die Grenzen der Software aufgezeigt. Abschließend werden zwei anspruchsvolle Anwendungsbeispiele vorgestellt: Eine zentrisch belastete Verschraubung mit Berücksichtigung von Elasto-Plastizität und einer komplexen Lasthistorie (Anziehen durch Anzugsmoment, Setzeffekte, Entfall des Torsionsmomentes durch das Anziehen, Betriebskraft) sowie eine exzentrisch belastete Verschraubung, die wegen eines relativ dünnen Flansches starke Biegezusatzbeanspruchungen erfährt
The presentation shows the foundations of bolt analysis according to VDI-guideline 2230 part 1. In addition, the four FEM model classes proposed in VDI 2230 part 2 (draft) are described, as well as the features available in Creo Simulate to realize these model classes. Next, the presentation shows the requirements for linearizing bolted connections, and why in a FEM analysis with a linearized connection no preload is necessary. The new fastener feature introduced in Creo Simulate is explained in detail. This feature allows the automated modeling and analysis of bolted connections having standard geometry. Further software features, like pretension elements, as well as the current software limitations are shown. Finally, two advanced application examples are shown: A centrically loaded bolted connection taking into account elasto-plasticity and a complex load history (tightening torque, embedding, removal of tightening stress, operational load), and an eccentrically loaded flange connection, which is subjected to high additional bending loads because the flange is relatively thin

This thesis explores connections between homotopy sheaves, manifold calculus of functors and operad theory. We argue that there is a deep overlap between these, and as evidence we give a new operadic description of the homotopy theoretical obstructions to deforming a smooth immersion into a smooth embedding. We then discuss an application which improves on some aspects of recent results of Arone-Turchin and Dwyer-Hess concerning spaces of long knots and high-dimensional variants. Along the way, we define fibrewise complete Segal spaces, a mild generalisation of Rezk's notion of complete Segal spaces. Also in the context of Segal spaces, we define right fibrations and prove a Grothendieck construction theorem for presheaves with values in spaces. Finally, we prove a result of independent interest which states that weakly k-reduced operads (those with contractible space of operations in arity j ? k) can be strictified when k = 0, 1.

We investigate existence of solutions for a singular Liouville equation on S^2 and prove sharp Onofri-type inequalities for a Moser-Trudinger functional in the presence of singular potentials. As a consequence we obtain existence of extremal functions for the Moser-Trudinger embedding on compact surfaces with conical singularities. Finally we study the blow-up behavior for sequences of solutions Liouville-type systems and prove a compactness condition which plays an important role in the variational analysis of Toda systems.

La présente thèse décrit un projet de recherche multidisciplinaire qui porte sur la reconnaissance de l’émotion dans les sons, couvrant les théories psychologiques, l’analyse du signal acoustique,l’apprentissage automatique et la dynamique chaotique.Dans nos interactions et nos relations sociales, nous dépendons considérablement de la communication de l’information et de notre perception des messages transmis. En fait, la communication se produit lorsque les signaux transmettent des informations entre une source et une destination. Le signal peut être verbal, et l’information est ensuite portée par des motifs sonores, tels que des mots. Dans la communication vocale non verbale, cependant,l’information peut être des modèles perceptifs qui véhiculent des indices affectifs, que nous percevons et évaluons sous la forme d’intentions, d’attitudes, d’humeurs et d’émotions.La prévalence de la composante affective peut être observée dans les interactions informatiques humaines (HCI) où le développement d’applications automatisées qui comprennent et expriment les émotions est devenu crucial. De tels systèmes doivent être significatifs et faciles à utiliser pour l’utilisateur final, de sorte que notre interaction avec eux devient une expérience positive. Bien que la reconnaissance automatique des émotions dans les sons ait reçu une attention accrue au cours des dernières années, il s’agit encore d’un jeune domaine de recherche.Non seulement cela contribue à l’informatique affective en général, mais il fournit également une compréhension approfondie de la signification des sons dans notre vie quotidienne.Dans cette thèse, le problème de la reconnaissance des affects est abordé à partir d’une double perspective: nous commençons par adopter une approche standard de l’analyse acoustique du signal, où nous examinons et expérimentons les fonctionnalités existantes pour déterminer leur rôle dans la communication émotionnelle. Ensuite, nous nous tournons vers la dynamique chaotique et la symbolisation des séries temporelles, pour comprendre le rôle de la dynamique inhérente des sons dans l’expressivité affective. Nous menons nos études dans le contexte des sons non verbaux, à savoir les sons vocaux, musicaux et environnementaux.D’un point de vue de l’écoute humaine, une tâche d’annotation est menée pour construire un ground-truth de voix de chant non verbales, marquées par des descriptions catégoriques du modèle bidimensionnel d’émotions. Deux types de sons sont inclus dans l’étude: vocal et glottal.D’un point de vue psychologique, la présente recherche porte sur un débat qui existe depuis longtemps parmi les scientifiques et les psychologues, concernant les origines communes de la musique et de la voix. La question est abordée à partir d’une analyse acoustique ainsi que d’une approche dynamique non linéaire.D’un point de vue de la modélisation, ce travail propose une nouvelle approche dynamique non linéaire pour la reconnaissance de l’affect dans le son, basée sur la dynamique chaotique et la symbolisation adaptative des séries temporelles. Tout au long de cette thèse, les contrastes clés dans l’expressivité de l’émotion sont illustrés parmi les différents types de sons, à travers l’analyse des propriétés acoustiques, les métriques de la dynamique non linéaire et les performances des prédictions.Enfin, d’un point de vue progressif, nous suggérons que les travaux futurs étudient des caractéristiques motivées par les études cognitives. Nous suggérons également d’examiner dans quelle mesure nos caractéristiques reflètent les processus cognitifs. En outre, nous recommandons que nos fonctionnalités dynamiques soient testées dans des études à grande échelle de la reconnaissance d’émotions à travers la participation aux défis expérimentaux, dans le but de vérifier s’ils obtiennent un consensus
The present thesis describes a multidisciplinary research project on emotion recognition in sounds, covering psychological theories, acoustic-based signal analysis, machine learning and chaotic dynamics.In our social interactions and relationships, we rely greatly on the communication of information and on our perception of the messages transmitted. In fact communication happens when signals transmit information between a source and a destination. The signal can be verbal,and the information is then carried by sound patterns, such as words. In non verbal vocal communication however, information can be perceptual patterns that convey affective cues,that we sense and appraise in the form of intentions, attitudes, moods and emotions.The prevalence of the affective component can be seen in human computer interactions(HCI) where the development of automated applications that understand and express emotions has become crucial. Such systems need to be meaningful and friendly to the end user, so thatour interaction with them becomes a positive experience. Although the automatic recognition of emotions in sounds has received increased attention in recent years, it is still a young fieldof research. Not only does it contribute to Affective Computing in general, but it also provides insight into the significance of sounds in our daily life.In this thesis the problem of affect recognition is addressed from a dual perspective: we start by taking a standard approach of acoustic-based signal analysis, where we survey and experiment with existing features to determine their role in emotion communication. Then,we turn to chaotic dynamics and time series symbolization, to understand the role of the inherent dynamics of sounds in affective expressiveness. We conduct our studies in the context of nonverbal sounds, namely voice, music and environmental sounds.From a human listening point of view, an annotation task is conducted to build a ground truth of nonverbal singing voices, labelled with categorical descriptions of the two-dimensional model of affect. Two types of sounds are included in the study: vocal and glottal.From a psychological perspective, the present research addresses a debate that is of long standing among scientists and psychologists, concerning the common origins of music and voice.The question is addressed from an acoustic-based analysis as well as a nonlinear dynamics approach.From a modeling viewpoint, this work proposes a novel nonlinear dynamics approach for the recognition of affect in sound, based on chaotic dynamics and adaptive time series symbolization.Throughout this thesis, key contrasts in the expressiveness of affect are illustrated among the different types of sounds, through the analysis of acoustic properties, nonlinear dynamics metrics and predictions performances. Finally from a progressive perspective, we suggest that future works investigate features that are motivated by cognitive studies. We also suggest to examine to what extent our features reflect cognitive processes. Additionally we recommend that our dynamic features be tested inlarge scale ER studies through the participation in ER challenges, with an aim to verify if they gain consensus

The understanding and treatment of challenging behaviors in individuals with Autism Spectrum Disorder is paramount to enabling the success of behavioral therapy; an essential step in this process being the labeling of challenging behaviors demonstrated in therapy sessions. These manifestations differ across individuals and within individuals over time and thus, the appropriate classification of a challenging behavior when considering purely qualitative factors can be unclear. In this thesis we seek to add quantitative depth to this otherwise qualitative task of challenging behavior classification. We do so through the application of natural language processing techniques to behavioral descriptions extracted from the CARD Skills dataset. Specifically, we construct 3 sets of 50-dimensional document embeddings to represent the 1,917 recorded instances of challenging behaviors demonstrated in Applied Behavior Analysis therapy. These embeddings are learned through three processes: a TF-IDF weighted sum of Word2Vec embeddings, Doc2Vec embeddings which use hierarchical softmax as an output layer, and Doc2Vec which optimizes the original Doc2Vec architecture through Negative Sampling. Once created, these embeddings are initially used as input to a Support Vector Machine classifier to demonstrate the success of binary classification within this problem set. This preliminary exploration achieves promising classification accuracies ranging from 78.2-100% and establishes the separability of challenging behaviors given their neural embeddings. We next construct a multi-class classification model via a Gaussian Process Classifier fitted with Laplace approximation. This classification model, trained on an 80/20 stratified split of the seven most frequently occurring behaviors in the dataset, produces an accuracy of 82.7%. Through this exploration we demonstrate that the semantic queues derived from the language of challenging behavior descriptions, modeled using natural language processing techniques, can be successfully leveraged in classification architectures. This study represents the first of its kind, providing a proof of concept for the application of machine learning to the observations of challenging behaviors demonstrated in ASD with the ultimate goal of improving the efficacy of the behavioral treatments which intrinsically rely on the accurate identification of these behaviors.

Dans le Web des données, des graphes de connaissances de plus en plus nombreux sont simultanément publiés, édités, et utilisés par des agents humains et logiciels. Cette large adoption rend essentielles les tâches d'appariement et de fouille. L'appariement identifie des unités de connaissances équivalentes, plus spécifiques ou similaires au sein et entre graphes de connaissances. Cette tâche est cruciale car la publication et l'édition parallèles peuvent mener à des graphes de connaissances co-existants et complémentaires. Cependant, l'hétérogénéité inhérente aux graphes de connaissances (e.g., granularité, vocabulaires, ou complétude) rend cette tâche difficile. Motivés par une application en pharmacogénomique, nous proposons deux approches pour apparier des relations n-aires représentées au sein de graphes de connaissances : une méthode symbolique à base de règles et une méthode numérique basée sur le plongement de graphe. Nous les expérimentons sur PGxLOD, un graphe de connaissances que nous avons construit de manière semi-automatique en intégrant des relations pharmacogénomiques de trois sources du domaine. La tâche de fouille permet quant à elle de découvrir de nouvelles unités de connaissances à partir des graphes de connaissances. Leur taille croissante et leur nature combinatoire entraînent des problèmes de passage à l'échelle que nous étudions dans le cadre de la fouille de patrons de chemins. Nous proposons également l'annotation de concepts, une méthode d'amélioration des graphes de connaissances qui étend l'Analyse Formelle de Concepts, un cadre mathématique groupant des entités en fonction de leurs attributs communs. Au cours de tous nos travaux, nous nous sommes particulièrement intéressés à tirer parti des connaissances de domaines formalisées au sein d'ontologies qui peuvent être associées aux graphes de connaissances. Nous montrons notamment que, lorsqu'elles sont prises en compte, ces connaissances permettent de réduire l'impact des problèmes d'hétérogénéité et de passage à l'échelle dans les tâches d'appariement et de fouille
In the Web of data, an increasing number of knowledge graphs are concurrently published, edited, and accessed by human and software agents. Their wide adoption makes key the two tasks of matching and mining. First, matching consists in identifying equivalent, more specific, or somewhat similar units within and across knowledge graphs. This task is crucial since concurrent publication and edition may result in coexisting and complementary knowledge graphs. However, this task is challenging because of the inherent heterogeneity of knowledge graphs, e.g., in terms of granularities, vocabularies, and completeness. Motivated by an application in pharmacogenomics, we propose two approaches to match n-ary relationships represented in knowledge graphs: a symbolic rule-based approach and a numeric approach using graph embedding. We experiment on PGxLOD, a knowledge graph that we semi-automatically built by integrating pharmacogenomic relationships from three distinct sources of this domain. Second, mining consists in discovering new and useful knowledge units from knowledge graphs. Their increasing size and combinatorial nature entail scalability issues, which we address in the mining of path patterns. We also propose Concept Annotation, a refinement approach extending Formal Concept Analysis, a mathematical framework that groups entities based on their common attributes. Throughout all our works, we particularly focus on taking advantage of domain knowledge in the form of ontologies that can be associated with knowledge graphs. We show that, when considered, such domain knowledge alleviates heterogeneity and scalability issues in matching and mining approaches

In the field of Natural Language Processing, supervised machine learning is commonly used to solve classification tasks such as sentiment analysis and text categorization. The classical way of representing the text has been to use the well known Bag-Of-Words representation. However lately low-dimensional dense word vectors have come to dominate the input to state-of-the-art models. While few studies have made a fair comparison of the models' sensibility to the text representation, this thesis tries to fill that gap. We especially seek insight in the impact various unsupervised pre-trained vectors have on the performance. In addition, we take a closer look at the Random Indexing representation and try to optimize it jointly with the classification task. The results show that while low-dimensional pre-trained representations often have computational benefits and have also reported state-of-the-art performance, they do not necessarily outperform the classical representations in all cases.

PEREIRA, Rodrigo Mendes.Geometria métrica e topologia de superfícies algebricamente parametrizadas. 2016. 68 f. Tese (Doutorado em Matemática) – Centro de Ciências, Universidade Federal do Ceará, Fortaleza, 2016.
Submitted by Andrea Dantas (pgmat@mat.ufc.br) on 2017-06-09T17:57:35Z No. of bitstreams: 1 2016_tese_rmpereira.pdf: 663687 bytes, checksum: 396b9fef564fece1bd791fcf3677b5e5 (MD5)
Rejected by Rocilda Sales (rocilda@ufc.br), reason: Oi Andrea, Favor informar ao aluno para fazer as seguintes correções: No Resumo e Abstract não tem recuo. Os termos das palavras-chave e abstract são separadas e finalizadas por ponto. nAS Referências, o sobrenome dos autores são em caixa alta. Ex: WHITNEY, Hassler. Dois ou três autores: SOBRENOME, Nome; SOBRENOME, Nome Rocilda on 2017-06-13T12:22:02Z (GMT)
Submitted by Andrea Dantas (pgmat@mat.ufc.br) on 2017-06-19T17:03:38Z No. of bitstreams: 1 2016_tese_rmpereira.pdf: 663687 bytes, checksum: 396b9fef564fece1bd791fcf3677b5e5 (MD5)
Rejected by Rocilda Sales (rocilda@ufc.br), reason: Oi Andrea, Estou devolvendo novamente o arquivo do Rodrigo, pois ele não fez as correções. Elton disse que ele veio ontem e foi dito o quais eram. Estou enviando as mesmas. Favor repassar pra ele. Caso ele não tenha entendido são essas as correções: Sumário a Abstracts não tem o recuo. As palavras-chave e keywords são separados e finalizados por ponto. E Referências : Os sobrenomes dos autores são em caixa alta: BALLETSEROS, J .J .Nuno. BIRBRAIR, L. ; FERNANDES, A.; GRANDJEAN, V. SE FOR MAIS DE 3 AUTORES INDICA-SE APENAS O PRIMEIRO ACRESCENTANDO A EXPRESSÃO ET AL. Birbrair, L ET AL. on 2017-06-21T14:06:19Z (GMT)
Submitted by Andrea Dantas (pgmat@mat.ufc.br) on 2017-06-22T12:40:49Z No. of bitstreams: 1 2016_tese_rmpereira.pdf: 661684 bytes, checksum: c8673d7c6d1002a51d5c45584e60f6b7 (MD5)
Approved for entry into archive by Rocilda Sales (rocilda@ufc.br) on 2017-06-22T14:20:44Z (GMT) No. of bitstreams: 1 2016_tese_rmpereira.pdf: 661684 bytes, checksum: c8673d7c6d1002a51d5c45584e60f6b7 (MD5)
Made available in DSpace on 2017-06-22T14:20:44Z (GMT). No. of bitstreams: 1 2016_tese_rmpereira.pdf: 661684 bytes, checksum: c8673d7c6d1002a51d5c45584e60f6b7 (MD5) Previous issue date: 2016-07-20
In this work, we study the singularities of the surfaces obtained as an image of a plane by an algebraic or analytic application in the spaces R3 and R4. We study the topological and metric properties of the knots that are obtained as a link of this surfaces. Normal embedding criterion are obtained for the surfaces and, in addition, the connections between the topology of the knot and normal embedding are investigated. We will also give a description of the tangent cones of these surfaces at the singular points.
Nesse trabalho, estudamos as singularidades das superfícies obtidas como imagem de um plano por uma aplicação algébrica ou analítica no espaços R3 e R4. Estudamos as propriedades topológicas e métricas dos nós que são obtidos como link de tais superfícies. É obtido critérios de mergulho normal para a superfície-imagem e, além disso, é investigado as ligações entre a topologia do nó e o mergulho normal. Faremos ainda uma descrição dos cones tangentes destas superfícies nos pontos singulares.

The creation of word embeddings is one of the key breakthroughs in natural language processing. Word embeddings allow for words to be represented semantically, opening the way to many new deep learning methods. Understanding what information is in word embeddings will help understanding the behaviour of embeddings in natural language processing tasks, but also allows for the quantitative study of the linguistic features such as grammatical gender. This thesis attempts to explore how grammatical gender is encoded in word embeddings, through analysing the performance of a neural network classifier on the classification of nouns by gender. This analysis is done in three experiments: an analysis of contextualized embeddings, an analysis of embeddings learned from modified corpora and an analysis of aligned embeddings in many languages. The contextualized word embedding model ELMo has multiple output layers with a gradual increasing presence of semantic information in the embedding. This differing presence of semantic information was used to test the classifier's reliance on semantic information. Swedish, German, Spanish and Russian embeddings were classified at all layers of a three layered ELMo model. The word representation layer without any contextualization was found to produce the best accuracy, indicating the noise introduced by the contextualization was more impactful than any potential extra semantic information. Swedish embeddings were learned from a corpus stripped of articles and a stemmed corpus. Both sets of embeddings showed an drop of about 6% in accuracy in comparison with the embeddings from a non-augmented corpus, indicating agreement plays a large role in the classification. Aligned multilingual embeddings were used to measure the accuracy of a grammatical gender classifier in 24 languages. The classifier models were applied to data of other languages to determine the similarity of the encoding of grammatical gender in these embeddings. Correcting the results with a random guessing baseline shows that transferred models can be highly accurate in certain language combinations and in some cases almost approach the accuracy of the model on its source data. A comparison between transfer accuracy and phylogenetic distance showed that the model transferability follows a pattern that resembles the phylogenetic distance.

Nell'era dell'informatizzazione e con l'avvento del web semantico vi è una sempre crescente necessità di classificare e organizzare grandi moli di dati non strutturati posti in linguaggio naturale al fine di poter trarre da essi informazioni utili alla costruzione di una conoscenza. Tuttavia l'analisi e la classificazione di tali tipi di dati rappresenta, per un sistema automatico, un problema non triviale quanto per un essere umano. A tale scopo in tempi odierni assistiamo a una larga diffusione di tecniche di natural language processing, text mining e sentiment analysis grazie anche ai recenti importanti sviluppi nel campo del machine learning. L'obiettivo che questo elaborato si pone è quello di presentare un approccio alternativo alla classificazione automatica di documenti in categorie gerarchiche di argomenti basato sulle relazioni semantiche che intercorrono tra le parole che compongono gli stessi. Tali relazioni vengono attinte da una base di conoscenza semantica mediante algoritmi sviluppati appositamente, rendendo così la determinazione del grado di correlazione completamente indipendente dagli argomenti usati per l'addestramento e persino dalla lingua. Il modello proposto si differenzia da quelli presenti in letteratura classica per le sue caratteristiche di riusabilità e generalizzazione grazie alle quali è in grado di operare, in fase di classificazione, anche su tematiche e categorie non conosciute in fase di addestramento.