Workflow management in problem solving environments (PSEs) is an emerging topic that aims to combine both data-oriented and execution-oriented views of scientific experiments, and closely integrate the processes underlying the practice of computational science with the software artifacts constituted by the PSE. This thesis presents a workflow management solution called BREW (BetteR Experiments through Workflow management) that provides functionality along four dimensions: components and installation management, experiment execution management, data management, and (full fledged) workflow management. BREW builds upon EMDAG, a first generation experiment management system designed at Virginia Tech which provided rudimentary facilities for supporting (only) the first two functionalities. BREW provides a complete dynamic workflow management solution wherein the PSE user can compose arbitrary scientific experiments and specify intended dynamic behavior of these experiments to an extent not previously possible. Along with the design details of the BREW system, this thesis identifies important tradeoffs underlying workflow management for PSEs, and presents two case studies involving large-scale data assimilation in bioinformatics experiments. / Master of Science

Automatic speech recognition (ASR) technology has matured over the past few decades and has made significant impacts in a variety of fields, from assistive technologies to commercial products. However, ASR system development is a resource intensive activity and requires language resources in the form of text annotated audio recordings and pronunciation dictionaries. Unfortunately, many languages found in the developing world fall into the resource-scarce category and due to this resource scarcity the deployment of ASR systems in the developing world is severely inhibited. In this thesis we present research into developing techniques and tools to (1) harvest audio data, (2) rapidly adapt ASR systems and (3) select “useful” training samples in order to assist with resource-scarce ASR system development. We demonstrate an automatic audio harvesting approach which efficiently creates a speech recognition corpus by harvesting an easily available audio resource. We show that by starting with bootstrapped acoustic models, trained with language data obtain from a dialect, and then running through a few iterations of an alignment-filter-retrain phase it is possible to create an accurate speech recognition corpus. As a demonstration we create a South African English speech recognition corpus by using our approach and harvesting an internet website which provides audio and approximate transcriptions. The acoustic models developed from harvested data are evaluated on independent corpora and show that the proposed harvesting approach provides a robust means to create ASR resources. As there are many acoustic model adaptation techniques which can be implemented by an ASR system developer it becomes a costly endeavour to select the best adaptation technique. We investigate the dependence of the adaptation data amount and various adaptation techniques by systematically varying the adaptation data amount and comparing the performance of various adaptation techniques. We establish a guideline which can be used by an ASR developer to chose the best adaptation technique given a size constraint on the adaptation data, for the scenario where adaptation between narrow- and wide-band corpora must be performed. In addition, we investigate the effectiveness of a novel channel normalisation technique and compare the performance with standard normalisation and adaptation techniques. Lastly, we propose a new data selection framework which can be used to design a speech recognition corpus. We show for limited data sets, independent of language and bandwidth, the most effective strategy for data selection is frequency-matched selection and that the widely-used maximum entropy methods generally produced the least promising results. In our model, the frequency-matched selection method corresponds to a logarithmic relationship between accuracy and corpus size; we also investigated other model relationships, and found that a hyperbolic relationship (as suggested from simple asymptotic arguments in learning theory) may lead to somewhat better performance under certain conditions. / Thesis (PhD (Computer and Electronic Engineering))--North-West University, Potchefstroom Campus, 2013.

Automatic speech recognition (ASR) technology has matured over the past few decades and has made significant impacts in a variety of fields, from assistive technologies to commercial products. However, ASR system development is a resource intensive activity and requires language resources in the form of text annotated audio recordings and pronunciation dictionaries. Unfortunately, many languages found in the developing world fall into the resource-scarce category and due to this resource scarcity the deployment of ASR systems in the developing world is severely inhibited. In this thesis we present research into developing techniques and tools to (1) harvest audio data, (2) rapidly adapt ASR systems and (3) select “useful” training samples in order to assist with resource-scarce ASR system development. We demonstrate an automatic audio harvesting approach which efficiently creates a speech recognition corpus by harvesting an easily available audio resource. We show that by starting with bootstrapped acoustic models, trained with language data obtain from a dialect, and then running through a few iterations of an alignment-filter-retrain phase it is possible to create an accurate speech recognition corpus. As a demonstration we create a South African English speech recognition corpus by using our approach and harvesting an internet website which provides audio and approximate transcriptions. The acoustic models developed from harvested data are evaluated on independent corpora and show that the proposed harvesting approach provides a robust means to create ASR resources. As there are many acoustic model adaptation techniques which can be implemented by an ASR system developer it becomes a costly endeavour to select the best adaptation technique. We investigate the dependence of the adaptation data amount and various adaptation techniques by systematically varying the adaptation data amount and comparing the performance of various adaptation techniques. We establish a guideline which can be used by an ASR developer to chose the best adaptation technique given a size constraint on the adaptation data, for the scenario where adaptation between narrow- and wide-band corpora must be performed. In addition, we investigate the effectiveness of a novel channel normalisation technique and compare the performance with standard normalisation and adaptation techniques. Lastly, we propose a new data selection framework which can be used to design a speech recognition corpus. We show for limited data sets, independent of language and bandwidth, the most effective strategy for data selection is frequency-matched selection and that the widely-used maximum entropy methods generally produced the least promising results. In our model, the frequency-matched selection method corresponds to a logarithmic relationship between accuracy and corpus size; we also investigated other model relationships, and found that a hyperbolic relationship (as suggested from simple asymptotic arguments in learning theory) may lead to somewhat better performance under certain conditions. / Thesis (PhD (Computer and Electronic Engineering))--North-West University, Potchefstroom Campus, 2013.

Les zones blanches étendues sont de vastes espaces géographiques (forêts, déserts), sans ou ayant très peu d’infrastructures telles que les routes, les réseaux électriques ou de télécommunication. Cependant, très souvent, dans ces zones se développent de nombreuses activités économiques ou environnementales telles que le monitoring de l’environnement, la surveillance d’une frontière ou d’une installation de pipeline, ou encore la prévention des feux de forêt. Grâce aux techniques de télédétection et de communication, une fonction clé de ces activités repose sur la collecte d’informations issues de capteurs qui sont transmises à un centre d’analyse distant. Nous proposons des solutions réseau afin d’effectuer la collecte de ces données dans les zones blanches étendues grâce à des technologies de communication longue distance et faible énergie, de type LoRaWAN. Pour le problème du déploiement du réseau de capteurs sans fil dans ces zones difficiles, nous avons proposé une heuristique inspirée de la croissance biologique d’un champignon, le physarum. Le physarum est capable de créer un corps complexe de liens pour trouver de la nourriture nécessaire à sa survie tout en optimisant ses propres ressources corporelles lors des périodes de disette. Ce principe d’optimisation a été adapté au domaine des réseaux pour déployer un réseau tolérant aux fautes, tout en minimisant le nombre de ressources ou relais à placer sur la zone d’intérêt. Nous nous sommes ensuite intéressés à la collecte opportuniste de données dans les zones blanches afin de pouvoir collecter l’information des nœuds trop éloignés d’une station relais. Nous avons développé une méthode de collecte basée sur les avions de ligne qui survole le territoire. Durant une fenêtre de communication, l’avion est à portée d’un capteur et peut ainsi collecter les données stockées qui seront livrées au serveur à l’atterrissage de l’avion. Notre dernière contribution utilise conjointement les deux méthodes précédentes, pour permettre à la fois le déploiement du réseau et la collecte des capteurs isolés. / Although wide white areas are not equipped or sparsely equipped with any infrastructure (energy, roads ...), strategic human activities are being carried out such as mines, forest, pipeline... To tackle the problem of deploying sensor networks in a very large area where few infrastructures are available, we propose a network deployment algorithm which aims at efficiently linking sparse points of interest in a very wide white area. The originality of the proposed method is that it mimics the evolution of a type of mold called physarum. Secondly, we aim at overcoming the deployment problem in wide white areas by using long range communication between an aircraft and earth. The new data collection scheme he proposes is based on the use of commercial flights to collect data while they cross over an area of interest. It investigates the feasibility of such a scheme by determining the collection capacity of commercial aircraft in different locations of the desert. Finally, we mixed both solutions do repatriate data from sensors not covered by any flight to a covered data sink that relays data to the aircraft.

In recent years, Unmanned aerial vehicles (UAVs) have developed rapidly due to advances in aerospace technology, and wireless communication systems. As a result of their versatility, cost-effectiveness, and flexibility of deployment, UAVs have been developed to accomplish a variety of large and complex tasks without terrain restrictions, such as battlefield operations, search and rescue under disaster conditions, monitoring, etc. Data collection and offloading missions in The internet of thingss (IoTs) networks can be accomplished with the use of UAVs as network edge nodes. The fundamental challenge in such scenarios is to develop a UAV movement policy that enhances the quality of mission completion and avoids collisions. Real-time learning based on neural networks has been proven to be an effective method for solving decision-making problems in a dynamic, unknown environment. In this thesis, we assume a real-life scenario in which a UAV collects data from Ground base stations (GBSs) without knowing the information of the environment. A UAV is responsible for the MOO including collecting data, avoiding obstacles, path planning, and conserving energy. Two Deep reinforcement learnings (DRLs) approaches were implemented in this thesis and compared. / Under de senaste åren har UAV utvecklats snabbt på grund av framsteg inom flygteknik och trådlösa kommunikationssystem. Som ett resultat av deras mångsidighet, kostnadseffektivitet och flexibilitet i utbyggnaden har UAV:er utvecklats för att utföra en mängd stora och komplexa uppgifter utan terrängrestriktioner, såsom slagfältsoperationer, sök och räddning under katastrofförhållanden, övervakning, etc. Data insamlings- och avlastningsuppdrag i IoT-nätverk kan utföras med användning av UAV:er som nätverkskantnoder. Den grundläggande utmaningen i sådana scenarier är att utveckla en UAV-rörelsepolicy som förbättrar kvaliteten på uppdragets slutförande och undviker kollisioner. Realtidsinlärning baserad på neurala nätverk har visat sig vara en effektiv metod för att lösa beslutsfattande problem i en dynamisk, okänd miljö. I den här avhandlingen utgår vi från ett verkligt scenario där en UAV samlar in data från GBS utan att känna till informationen om miljön. En UAV är ansvarig för MOO inklusive insamling av data, undvikande av hinder, vägplanering och energibesparing. Två DRL-metoder implementerades i denna avhandling och jämfördes.