abstract: In many biological research studies, including speech analysis, clinical research, and prediction studies, the validity of the study is dependent on the effectiveness of the training data set to represent the target population. For example, in speech analysis, if one is performing emotion classification based on speech, the performance of the classifier is mainly dependent on the number and quality of the training data set. For small sample sizes and unbalanced data, classifiers developed in this context may be focusing on the differences in the training data set rather than emotion (e.g., focusing on gender, age, and dialect). This thesis evaluates several sampling methods and a non-parametric approach to sample sizes required to minimize the effect of these nuisance variables on classification performance. This work specifically focused on speech analysis applications, and hence the work was done with speech features like Mel-Frequency Cepstral Coefficients (MFCC) and Filter Bank Cepstral Coefficients (FBCC). The non-parametric divergence (D_p divergence) measure was used to study the difference between different sampling schemes (Stratified and Multistage sampling) and the changes due to the sentence types in the sampling set for the process. / Dissertation/Thesis / Masters Thesis Electrical Engineering 2020

The basic aim of this mini-dissertation was to explore the prospects of digitisation at the University of Ghana Library System (UGLS). The research followed a qualitative approach and a case study research design was adopted. A thorough literature study was conducted. The primary purpose of the literature was to inform the research on the questions and objectives raised on the aspects of digitisation to understand the dynamics and complexity of digitisation. Six staff from the UGLS digitisation programme were purposively sampled for in-depth interviews for their direct involvement in the UGLS digitisation programme. Semi-structured interviews were then conducted with UGLS digitisation managers and the current IT and digitisation operations staff to collect data. During the interviews, the following issues associated with digitisation were addressed: policy, planning, goals and priorities, selection criteria, skills and expertise, digital preservation and long-term access, issues and challenges; solutions and recommendations to digitisation constraints of the UGLS. These issues were used to identify and develop themes where thematic analyses of research data were done. Specific recommendations were also develop to share with UGLS to address the digitisation policy and planning issues, selection issues, human resource and skills requirements; the critical challenges, resource requirements, long-term preservation and access of digitised content and the sustainability of digitisation programme. Final recommendations were made based on the findings and conclusions of the research to advise on the sustainability of the UGLS digitisation programme. / Mini Dissertation (MIT)--University of Pretoria, 2016. / Information Science / MIT / Unrestricted

Orientador: André Luiz Andreoli / Resumo: Os transformadores são equipamentos de suma importância no sistema elétrico de potência. Qualquer falha destes equipamentos, pode acarretar prejuízos e consequências graves tanto para as concessionárias de energia, quanto aos seus consumidores. A ocorrência de descargas parciais em transformadores de potência está diretamente relacionada às condições de seu sistema de isolamento. Elas podem se desenvolver ao longo do tempo e culminar na perda total do sistema de isolamento. Desta forma, o monitoramento das descargas parciais torna-se essencial para o planejamento de uma manutenção e a garantia do funcionamento adequado destes equipamentos. Neste contexto, o objetivo desta pesquisa é apresentar um método de localização e identificação de descargas parciais em transformador de potência operando com carregamento nominal, através de sensores piezelétricos de baixo custo e ferramentas como a transformada wavelet e os sistemas de inferência neuro-difuso adaptativos. / Mestre

The objective of this thesis is to identify, which hazards and failures in operation process will affect Reliability, Availability, Maintainability and Safety of floating offshore structures. The focus is on Dynamic Positioning (DP) system that has the responsibility of keeping the offshore structure in the upright position operation. DP system is one of the most critical subsystems on these types of structures in terms of safety of operation and failure risk costs. Reliability of the system in this thesis has been analyzed in qualitative and quantitativeb methods. In qualitative method to find the effective parameters on the reliability of the DPb system, Reliability Centered Maintenance (RCM ) and its application as a main tool have been used. To achieve the aim it has been tried to define the events and accidents which could be generated by the identified hazards then tried to determine the consequences of the realized accidents. In this step three categories are taken in to account including, safety, operation, and equipment. Next phase should be concentrated on considering and analyzing the relevant processes and the root causes which result in the identified hazard. After clarifying all probable root causes it has been tried to prioritize the root causes and specifying the necessary preventive actions. The aim of this step is either decreasing the occurrence of root causes or increasing the detectability of hazards. In the last part quantitative method has been used to measure the amounts of Reliability, Availability and Maintainability of the system, based on MTBF and MTTR of different components of the system and it has been tried to present the solutions to improve system reliability based on components RCM tables. Further, assuming DP system as human- machine system safety assessment has been included to indicate human factors in the reliability of the system beside probable failure of the components of the system.

<p><br></p><p>How to set privacy parameters is a crucial problem for the consistent application of DP in practice. The current privacy parameters do not provide direct suggestions for this problem. On the other hand, different databases may have varying degrees of information leakage, allowing attackers to enhance their attacks with the available information. This dissertation provides an additional interpretation of the current DP notions by introducing a framework that directly considers the worst-case average failure probability of attackers under different levels of knowledge. </p><p><br></p><p>To achieve this, we introduce a novel measure of attacker knowledge and establish a dual relationship between (type I error, type II error) and (prior, average failure probability). By leveraging this framework, we propose an interpretable paradigm to consistently set privacy parameters on different databases with varying levels of leaked information. </p><p><br></p><p>Furthermore, we characterize the minimax limit of private parameter estimation, driven by $1/(n(1-2p))^2+1/n$, where $p$ represents the worst-case probability risk and $n$ is the number of data points. This characterization is more interpretable than the current lower bound $\min{1/(n\epsilon^2),1/(n\delta^2)}+1/n$ on $(\epsilon,\delta)$-DP. Additionally, we identify the phase transition of private parameter estimation based on this limit and provide suggestions for protocol designs to achieve optimal private estimations. </p><p><br></p><p>Last, we consider a federated learning setting where the data are stored in a distributed manner and privacy-preserving interactions are required. We extend the proposed interpretation to federated learning, considering two scenarios: protecting against privacy breaches against local nodes and protecting privacy breaches against the center. Specifically, we consider a non-convex sparse federated parameter estimation problem and apply it to the generalized linear models. We tackle two challenges in this setting. Firstly, we encounter the issue of initialization due to the privacy requirements that limit the number of queries to the database. Secondly, we overcome the heterogeneity in the distribution among local nodes to identify low-dimensional structures.</p>

The object of this study was to investigate the micro-structural response of DP 600 subjected to high strain rate, ballistic impact tests. The ballistic tests were conducted using normal impact of a hardened steel penetrator into a 2 mm thick sheet of DP 600. The average strain rates produced from this test method are on the order of 10^5 s-1. Multiple methods were used to investigate the micro-structure before and after high strain rate deformation including optical microscopy, electron microscopy, and X-ray diffraction. A large variation in material response was observed between tests conducted at 0.8 x 10^5 and 2.5 x 10^5 s-1.

No description available.

No description available.

No description available.

Les premières cellules progénitrices lympho-myéloïdes (LMPP) entrent dans le thymus et commencent leur processus de différenciation au stade double négatif (DN, CD4-CD8-). Après un réarrangement fonctionnel de la chaine bêta de leur récepteur des cellules T (RCT), les cellules reçoivent des signaux de différenciation, de prolifération, de survie et de sélection et passent au stade double positif (DP, CD4+CD8+). Ensuite, la chaine alpha du RCT est réarrangée et testée via les sélections positive et négative. Si le RCT reçoit un signal ni trop fort, ni trop faible, les cellules passent au stade simple positif où elles exprimeront soit la molécule CD4 ou CD8. ERK3, une MAPK non-conventionnelle, joue un rôle important dans le développement thymique. Des études précédentes ont démontré qu’une déficience en ERK3 diminue de 50 % la cellularité thymique et de 75% le nombre de cellules simples positives CD4 (CD4SP). Nous avons posé comme hypothèses qu’il y a une augmentation de l’apoptose chez les thymocytes DP de souris Erk3-/- et que cette déficience chez les thymocytes DP affecterait la sélection positive des cellules DP, réduisant ainsi le nombre de thymocytes CD4SP. Afin de vérifier la première hypothèse, nous avons regardé les niveaux d’apoptose grâce à la cytométrie en flux et par immunohistochimie. Dans les deux cas, nous étions incapables de détecter une différence du niveau d’apoptose chez les thymocytes DP entre les souris Erk3+/+ et Erk3-/-. Ensuite, nous nous sommes posés la question suivante : La demi-vie des thymocytes DP Erk3-/- était-elle plus courte que le type sauvage? La demi-vie des thymocytes DP a été mesurée à l’aide de l’étude des réarrangements secondaires de la chaine alpha du RCT par PCR semi-quantitatif et à l’aide de cultures de thymus fœtaux. En effet, ERK3 semble important pour prolonger la demi-vie des thymocytes DP. Ensuite, nous avons utilisé des marqueurs cellulaires différentiels (CD69, CD5 et RCT) pour regarder si les thymocytes DP sont capables de passer la sélection positive. En effet, il y a moins de thymocytes DP Erk3-/- qui sont CD69fort, CD5fort et RCTfort. Finalement, nous voulons savoir si les fonctions de ERK3 passent par MK5, son seul partenaire d’interaction connu à ce jour. Après la caractérisation du thymus de la souris Mk5-/-, nous observons que seulement la réduction du nombre de thymocytes CD4SP est identique à celle des thymocytes CD4SP de la souris Erk3-/-. En conclusion, ces résultats révèlent des fonctions importantes pour la molécule ERK3 lors du processus de sélection positive, le maintient de la demi-vie des thymocytes DP et lors de la régulation de développement thymique de manière MK5-dépendante et -indépendante. / Early thymic progenitor cells (ETP) enter the thymus and begin their differentiation process at the double negative (DN) stage, having no CD4 and CD8 expression. After a functional rearrangement of the beta chain of the T-cell receptor (TCR), the cells receive differentiation, proliferation, survival, and selection signals and pass to the double positive (DP) stage and acquire the expression of CD4 and CD8 molecules. Then, the TCR alpha chain is rearranged and the fully composed TCR is tested through positive and negative selection. If the TCR receives a signal that is just right (not too strong or too weak), the cells pass to the single-positive stage by acquiring either the CD4 or CD8 molecule. ERK3, an unconventional MAPK, plays an important role in thymic development. Previous studies have shown that a deficiency in ERK3 decreases by 50% the total thymic cellularity and by 75% the number of CD4 single- positive cells (CD4SP). We hypothesized that there is an increase in apoptosis in DP thymocytes of Erk3-/- mice for and that this deficiency would affect the positive selection of DP cells, resulting in a reduced number of CD4SP. To test the first hypothesis, we looked at apoptosis levels by flow cytometry and immunohistochemistry. In both cases, we were unable to detect a difference in the level of apoptosis in DP thymocytes from Erk3+/+ and Erk3-/- mice. We then made a second assumption, that the half-life of Erk3-/- DP thymocytes for was shorter, which was measured by secondary rearrangements of the RCT alpha chain by semi-quantitative PCR and cultures of fetal thymi. Indeed, ERK3 seems important to extend the half- life of DP thymocytes. Next, we used differential cell markers (CD69, CD5, and TCR) to see if the DP thymocytes are able to pass positive selection. Indeed, there are less DP thymocytes Erk3-/- that are CD69hi, CD5hi, and TCRhi. Finally, we wanted to know whether the functions of ERK3 passes through MK5, its only interacting partner known to date. After characterization of the thymus of Mk5-/- mice, we observe that only the reduction of CD4SP thymocytes is identical to that of CD4SP thymocytes of Erk3-/-mice. In conclusion, these results reveal important functions for the molecule ERK3 for maintaining the half-life of DP thymocytes, for the process of positive selection and for regulating T-cell development in a MK5-dependent and - independent manner.