Ingeniero Civil en Computación / La creación y manipulación de colecciones de valores está ampliamente soportada por los lenguajes de programación modernos. A pesar de que cada lenguaje viene con su propia implementación de colecciones, la mayoría de las implementaciones encontradas son muy similares en estructura y funcionalidad ofrecida. Una excepción notable es el caso de Lua, pues presenta una sola colección: la tabla. Esta estructura es la unificación de una colección expansible secuencial y un arreglo asociativo o diccionario. La tabla es una estructura híbrida, compuesta por una parte de hash y una parte de arreglo, y hará uso de sus estructuras internas según su uso: si es utilizada como lista hará uso del arreglo y si es usada como diccionario usará el hash. Esto permite flexibilidad y eficiencia pues las estructuras internas sólo se crean cuando requieren ser usadas. Este comportamiento adaptativo es interesante en cuanto sugiere buen rendimiento mientras que se muestra muy simple para el usuario. A pesar de que Lua ha logrado bastante popularidad, las tablas no han sido cuidadosamente estudiadas por la comunidad científica. La eficiencia no solo depende de la implementación de las colecciones, sino también de si se escoge la colección correcta, con los parámetros adecuados, para cada situación particular. La selección inadecuada de colecciones puede resultar en un sobrecosto severo. Por este motivo, contar con una forma automática de asignar la colección más adecuada para cada caso sería muy valioso en la práctica. El experimento Chameleon aborda este problema sobre la JVM, mediante la identificación automática de elecciones inadecuadas de colecciones a través de análisis en tiempo de ejecución. El reemplazo de las colecciones identificadas por otras más adecuadas permitió reducir el consumo de memoria hasta en un 55% y el tiempo de ejecución en un 60% en algunas aplicaciones. Este trabajo enfrenta el problema de determinar la aplicabilidad de la tabla de Lua, o un derivado de esta, en lenguajes con bibliotecas ricas de colecciones. Para esto se desarrollan tres colecciones en el lenguaje Pharo: SLua, SmartCollection y SmartCollection2. Estas luego se evalúan en comparación con las colecciones normales de Pharo y con una versión lazy de las mismas mediante distintos tipos de benchmarks que miden su tiempo de ejecución y consumo de memoria, y se realiza análisis dinámico para entender las diferencias en los resultados. Además, se aborda el problema de la selección adaptativa de colecciones y se replica el experimento Chameleon. Luego se usa la herramienta de selección y reemplazo automático de colecciones para evaluar las colecciones desarrolladas. Los resultados de los experimentos realizados mediante instrumentación muestran reducciones de hasta un 30% en consumo de memoria y hasta un 15% en tiempo de ejecución en algunos escenarios al reemplazar las colecciones tradicionales por las desarrolladas. Sin embargo, la evaluación posterior mediante la herramienta de reemplazo de colecciones a nivel de código fuente no reveló diferencias estadísticamente significativas entre los tiempos de ejecución y uso de memoria entre las distintas colecciones. Esto indica una amplificación de los resultados introducida por la instrumentación. Las colecciones de las librerías externas utilizadas por las aplicaciones estudiadas no fueron reemplazadas, por lo que se plantea estudiar el impacto del reemplazo de colecciones sobre estas como trabajo futuro.

Growth phenotype profiling of genome-wide gene-deletion strains over stress conditions can offer a clear picture that the essentiality of genes depends on environmental conditions. Systematically identifying groups of genes from such recently emerging high-throughput data that share similar patterns of conditional essentiality and dispensability under various environmental conditions can elucidate how genetic interactions of the growth phenotype are regulated in response to the environment. In this dissertation, we first demonstrate that detecting such “co-fit” gene groups can be cast as a less well-studied problem in biclustering, i.e., constant-column biclustering. Despite significant advances in biclustering techniques, very few were designed for mining in growth phenotype data. Here, we propose Gracob, a novel, efficient graph-based method that casts and solves the constant-column biclustering problem as a maximal clique finding problem in a multipartite graph. We compared Gracob with a large collection of widely used biclustering methods that cover different types of algorithms designed to detect different types of biclusters. Gracob showed superior performance on finding co-fit genes over all the existing methods on both a variety of synthetic data sets with a wide range of settings, and three real growth phenotype data sets for E. coli, proteobacteria, and yeast.

Personalized recommender systems aim to assist users in retrieving and accessing interesting items by automatically acquiring user preferences from the historical data and matching items with the preferences. In the last decade, recommendation services have gained great attention due to the problem of information overload. However, despite recent advances of personalization techniques, several critical issues in modern recommender systems have not been well studied. These issues include: (1) understanding the accessing patterns of users (i.e., how to effectively model users' accessing behaviors); (2) understanding the relations between users and other objects (i.e., how to comprehensively assess the complex correlations between users and entities in recommender systems); and (3) understanding the interest change of users (i.e., how to adaptively capture users' preference drift over time). To meet the needs of users in modern recommender systems, it is imperative to provide solutions to address the aforementioned issues and apply the solutions to real-world applications. The major goal of this dissertation is to provide integrated recommendation approaches to tackle the challenges of the current generation of recommender systems. In particular, three user-oriented aspects of recommendation techniques were studied, including understanding accessing patterns, understanding complex relations and understanding temporal dynamics. To this end, we made three research contributions. First, we presented various personalized user profiling algorithms to capture click behaviors of users from both coarse- and fine-grained granularities; second, we proposed graph-based recommendation models to describe the complex correlations in a recommender system; third, we studied temporal recommendation approaches in order to capture the preference changes of users, by considering both long-term and short-term user profiles. In addition, a versatile recommendation framework was proposed, in which the proposed recommendation techniques were seamlessly integrated. Different evaluation criteria were implemented in this framework for evaluating recommendation techniques in real-world recommendation applications. In summary, the frequent changes of user interests and item repository lead to a series of user-centric challenges that are not well addressed in the current generation of recommender systems. My work proposed reasonable solutions to these challenges and provided insights on how to address these challenges using a simple yet effective recommendation framework.

Raw material accounts for a large proportion (approximately 75 percent) of a sawmill’s operating costs. However, about 15 percent of raw material ends up as low valued sawdust and planer shavings due to inaccurate cutting. Sizable financial benefits can be realized through maximizing conversion of raw material into valuable solid wood. Advanced process control in a sawmill can help achieve straighter cuts closer to final product dimensions and reduce loss of valuable raw material. A novel and practical method for enhanced process control in a sawmill is presented. A laser arrangement consisting of industrial point and line scanners is used to obtain a surface profile of the entire (two-dimensional) top and bottom surfaces of a lumber board. Each surface profile is independent of the other and free of data contamination caused by relative motions between the measured surface and sensors. Point scanners and line scanners simultaneously record 1-D and 2-D height data, respectively, along the length of the board. One-dimensional height data are used to identify relative motions through a mathematical technique based on linear inverse theory. Subtracting relative motion information from raw line scanner data provides an accurate 2-D surface profile. A second line scanner placed below the board can be used to obtain a separate 2-D profile of the bottom lumber surface. Separate-sided profiling is advantageous because typically a different saw or machine mills each side of a board. Thus, knowing the surface profile of each side of a board is crucial not only in diagnosing a deficiency in the milling process but also in determining the location of this deficiency. Results demonstrate that two-dimensional surface profiling can identify common surface defects such as step, washboard and knot tear-out with an accuracy of 0.3mm. Reproduction of each surface is rapid (approximately 0.2 seconds) and stable. / Applied Science, Faculty of / Mechanical Engineering, Department of / Graduate

A key question that marketers and salespeople face is based on the dilemma of which customer to pursue and which not. This choice has far reaching effects in both company success. Investigating a client’s seriousness to make a purchase at the time of an enquiry is key. This will assist the firm in determining which future clients offer the firm the most potential value in sales turnover and company profits. By taking the client’s background and history into consideration the marketer has the opportunity to determine which clients have or have not previously committed to the firm and which did not. By profiling the customer, providing a yardstick against which one can measure probability of a purchase can be developed. New or prospective customer against those attributes of customers can then be benchmarked against the attributes that attest to whether interest and intention to purchase can be converted into a sale. Identifying these customers who will be most likely to make a purchase is a very important area of study for any business that focuses on the marketing and sales of products or services. Having an accurate profile of potential customers, of who will buy and who will not, will play a key role in the success of any organization. The focus of this study will be on investigating the profile of a potential customer/company as a benchmark for sales lead qualification. The profile will also be a precursor to determining the lifetime value of the prospective customer. Profiling of high sales potential customers is a management issue. It directly influences the number of hours key personnel in the sales department spend on key tasks, which can alternatively be spent more productively on other projects.

Hepatitis C virus (HCV) is a growing health concern in Canada and around the world, as it currently infects 3% of the global population. While there is no vaccine available against this virus, novel and effective treatment regimens have improved prospects for the cure of HCV. Complications caused by HCV can lead to severe liver disease and even death. The limited viral proteome forces HCV to rely heavily on various host factors for its replication. Additionally HCV modulates the host physiology to facilitate its pathogenesis; consequently, the in dept study of essential host-virus interactions expands our understandingof how the virus and related species commandere host cell machinery. This understanding can help create new therapeutic strategies, which may have applications towards HCV and other related RNA viruses. While numerous studies have demonstrated that HCV modulates the abundance of various host proteins, the systematic study of the virus’s effect on the enzymatic activity has been relatively unexplored. For this reason, activity-based protein profiling (ABPP) was applied to study the changes in the activity of host enzymes during HCV replication. ABPP is a functional proteomics technique that employs active site-directed probe (ABP) to report on the activity of enzymes within complex proteomes, such as living cells. Herein, directed and non-directed ABPs were employed for specific as well as global profiling of the alterations in the activity of cellular enzymes during HCV replication. As a result, essential host enzymes that are differentially active during HCV infection were identified. Furthermore, I have developed a quantitative ABPP method for relative quantification of the cellular enzymes activity during HCV infection. These results contribute to the discovery of disease-associated biomarkers, with diagnostic significance, and aid in the identification of potential targets for therapeutic interventions. In addition to developing protein-based tools to study host-virus interactions, I employed a novel technique to investigate the interactions of micro-RNA 122 (miR-122), an essential HCV host factor, with the viral RNA genome. This in vitro screening approach, interrogates the folding of HCV RNA using viral RNA-coated magnetic bead (VRB) to determine target site accessibility for RNA silencing. This method predicts the relative affinity of small RNAs towards HCV genomic RNA that are not easily predicted by informatic means, and led to discovery of potent miR-122 interaction site within the large, highly-structured HCV RNA genome. For that reason, VRB assay may represent an attractive tool for the examination of target site accessibility for RNA silencing.

O presente estudo investiga a possibilidade do uso da perfilagem geofísica de Gama Natural e Resistividade na discriminação de estratos carbonosos, em relação a outras litologias presentes em quatro depósitos de carvão localizados no Rio Grande do Sul (áreas B3, CALOMBO, CERRO e SEIVAL). Uma vez observada a possibilidade de identificação dos estratos carbonosos com estes dois perfis geofísicos apenas, fez-se uma verificação da acuracidade na determinação das espessuras das camadas de carvão e também um exame da existência de correlação entre os registros geofísicos e parâmetros de qualidade do carvão (especificamente cinzas, enxofre e voláteis). Neste contexto, as espessuras das camadas de carvão foram comparadas com as espessuras observadas em testemunhos de sondagem, sendo que o reconhecimento das assinaturas das camadas pode ser feito com boa acuracidade para todos os depósitos, exceto no depósito do Cerro, onde o sinal do carvão confunde-se arenitos, inviabilizando muitas vezes a discriminação. Os parâmetros de qualidade do carvão foram determinados por meio de análise laboratorial, para posterior comparação com as leituras de perfilagem. Os resultados mostraram uma boa correlação entre a emissão de radiação gama natural em camadas de carvão e o teor de cinzas para os depósitos do CALOMBO e CERRO, mas uma correlação apenas razoável para os depósitos de B3 e SEIVAL. Além disso, os resultados indicaram que não há correlação entre teor de enxofre ou voláteis e os registros geofísicos. Vale ressaltar a importância deste tipo de avaliação, especialmente quanto às espessuras de camadas de carvão no planejamento de mina de curto prazo, onde não é exequível a sondagem com recuperação de testemunhos em bancadas de lavra simultaneamente ao processo extração de carvão, em função da demora da atividade de sondagem. / This study investigates the possibility of using geophysical logging of natural gamma and resistivity in the identification of carbonaceous strata in relation to other lithologies present in four coal deposits located in Rio Grande do Sul (areas B3,CALOMBO, CERRO and SEIVAL). Once observed, the possibility of identifying carbonaceous strata with these two geophysical logs, the degree of accuracy in the determination of the thicknesses of the coal seams and the existence of correlation between the geophysical logs and coal quality parameters (specifically ash, sulfur and volatile) were determined. In this context, the thicknesses of the coal seams were compared with the thickness of the core samples recovered in the boreholes and the result showed that the recognition of the coal seam signatures can be obtained with good accuracy for all deposits except the CERRO deposit, where the coal seam signature can be confused for that of sandstone which often precludes identification. The coal quality parameters were determined by laboratory analysis and later compared with the readings from geophysical logging. The results showed good correlation between natural gamma radiation emitted by the coal seams and the ash content in the CALOMBO and CERRO deposits but weak correlation in the B3 and SEIVAL deposits. Furthermore, the results showed no correlation between sulfur or volatile and the geophysical data. It is worth mentioning, the importance of this type of evaluation, especially in regards to the thickness of coal seams in short term mine planning, where it is not possible to recover core samples during the coal extraction process as a result of excessive time delay with respect to core sample recovery in boreholes.

Although cryptography based remote voting protocols have been researched since 1981, most of the previous protocols [9], [5], [13] assume the existence of public bul- letins or, in other words, a publicly readable, tamper-proof, append-only log. As blockchain or distributed ledger technology (DLT) offers properties like irreversibil- ity, transparency and decentralization, it is suitable for realization of public bulletin board for the voting system. We see a gap in the research of blockchain based voting systems because there either exists work on just the protocol aspect of the voting system or the performance aspect of the blockchain. As blockchain is a general purpose tool, we believe that there lies opportunities for micro-optimizations that could specifically benefit the voting system. This ushered us to focus our effort on the performance aspect of integration of voting protocol with blockchain. Hence, in this thesis, we first introduce a homomorphic encryption based voting protocol that uses blockchain, Hyperledger Fabric (HLF), as bulletin board. The protocol is de- signed such that it leverages the transaction processing characteristics of underlying DLT. We then created an experiment where we designed a smart contract, set up a blockchain network and exposed the system to 40k concurrent voting transactions to profile the code of HLF. From the profile data, it was found that execution of cryptographic operations constitutes most of the transaction processing time. This led us to benchmark cryptographic libraries for SHA256 and digital signature algo- rithm and integrate the faster library into HLF for better performance. We also found that the transaction manager of HLF does not need read-write locks to ensure transaction isolation in special scenarios, which alleviates the performance drop due to lock contention. Altogether we were able to improve the throughput and latency of the baseline system by more than 30%. Lastly, we make a comparison between public and permissioned DLT based remote voting system and discuss the suitability of permissioned blockchain for the application of voting systems. / Graduate

Thesis (Ed.D.)--Boston University / The focus of this study was to examine the controversial issue of racial profiling. This study examined the perceptions, attitudes, and behaviors of 112 police officers from four separate police departments. The study began with the premise that racial profiling was a widespread issue in law enforcement. It looked to identify attributes, similarities, and differences of contrasting police agencies that would explain why some groups engage in racial profiling and others do not. The research was also undertaken to determine what might mitigate racial profiling practices. This study examined the issue from three different positions. The first step was to determine whether police officers believed that racial profiling was problematic in their respective departments. The cohort overwhelmingly reported that they did not believe racial profiling was a serious problem. The second goal of the study was to determine if a department's organizational culture (attitudes, values, and behaviors) contributed to racial profiling attitudes and practices. This study found, through an examination of organizational commitment, that organizational culture did affect racial profiling practices. Police departments with more committed officers showed lower search disparity rates between minority and majority motorists. Finally, racial profiling attitudes and practices were examined controlling for training. Police departments where a majority of officers received racial profiling training showed lower search rate disparities between majority and minority motorists. They also showed higher levels of commitment to the organization. A statistical analysis found that training and Organizational Commitment were shown to have a statistically significant effect on racial profiling practices. Attitudes, however, appeared to be unaffected. Quantity of training appeared to be the critical factor; quality of training was not examined as a part of this research. Police departments that committed to department-wide training showed lower search rate disparities between minority and majority motorists. Further research is needed to identify the connection between training, organizational culture, and officers' individualized behavior.

Magister Pharmaceuticae - MPharm / The growing resistance of Methicillin-Resistant Staphylococcus aureus (MRSA) to currently prescribed drugs has resulted in the failure of prevention and treatment of different infections caused by the superbug. Therefore, to keep pace with the resistance, there is a pressing need for novel antimicrobial agents, especially from non-conventional sources. Several natural products (NPs) have displayed varying in vitro activities against the pathogen but few of these natural compounds have been studied for their prospects to be potential antimicrobial drug candidates. This may be due to the high cost, tedious, and time-consuming process of conducting the important preclinical tests on these compounds. Hence, there is a need for cost-effective strategies for mining the available data on these natural compounds. This would help to get the knowledge that may guide rational prioritization of “likely to succeed” natural compounds to be developed into potential antimicrobial drug candidates.