Zum Hauptinhalt springen
Konzeption und Entwicklung einer mobilen Messenger-Anwendung im Kontext einer Hochschule
The aim of the present bachelor thesis is to design and implement a messenger for the FH-Erfurt. For this purpose, comprehensive requirements for a messenger in the area of a university are defined in the concept, which are partly based on the analysis of the functions, the design and the structure of existing messengers. In order to implement the requirements, various communication protocols, architecture principles, storage options and mobile operating systems are presented and compared. The results are extensively represented by UML diagrams, errors are shown and possible improvements are named. The author hopes to gain knowledge in the conception and development of messenger applications in the area of a university, as well as an application that offers added value compared to existing messengers and promotes communication and networking between students and professors.
Unentdeckte Mehrwerte skalierter agiler Frameworks herausfinden und darstellen
The goal of the work is to find out which undiscovered added values scaled agility has. Undiscovered in the sense that scaled agility is not used as often as agile per se. So the question is: What added values does scaled agility have? And by what means do I use it? To answer the questions, agile was put in relation with scaled agile to highlight the added values of scaled agile. The result showed that regardless of whether Scaled Agile or Agile, it has a common core of benefits that one would not want to miss. However, it was clearly shown that in a larger constellation, scaling makes a lot of sense. With the help of scaling, large, complex problems are solved better. This allows the grouped teams to work faster and more efficiently by working in parallel on tasks that were previously broken down.
Empirische Analyse zur Portfolio-Entwicklung eines Managed Service Security Providers anhand eines qualitativen und quantitativen Kriterienkatalogs
This bachelor thesis deals with the development of a product portfolio for a Managed Service Security Provider (MSSP). A MSSP has great responsibility towards its customers, since the clients IT security is entrusted to the MSSP. This results in special and unique requirements for the products that a Managed Service Security Provider must utilise in its clients networks. In this paper, the three areas of firewall, endpoint protection and mobile device management are considered more closely. Suitable products are evaluated for each area, using a catalog of both quantitative and qualitative criteria. The evaluation criteria include technical and sales aspects, but also softer factors such as user-friendliness, support and other area-specific properties. In the end a recommendation for a product portfolio, optimized for a Managed Service Security Provider, is made on the basis of the analyses performed.
Ein Convolutional Neural Network zur Klassifizierung perforierter Zeichen
This bachelor thesis is about the development of a convolutional neural network for the recognition of perforated characters. As perforated characters can be found on passports and therefore only few images of such characters are available, synthetic images shall be used for training the network. After considering the theoretical fundamentals, the development process of the convolutional neural network is described. Thereby will be shown how an optimal structure of the network for the classification of perforated characters could be found by various modifications and tests. Furthermore, the development and generation of the required synthetic images is described. The resulting synthetic images can hardly be distinguished from real images. With these images it could be proven that a network trained only with synthetic data can correctly classify real data. The result is a network, which correctly classifies 99.47% of the real test data.
Reinforcement Learning - Schaffung einer Beobachtungsumgebung anhand von Tic-Tac-Toe
The bachelor thesis with the topic "Reinforcement Learning - Creation of an Observation Environment based on Tic Tac Toe" deals with the development of an interactive user environment, in oder to provide the opportunity for exploration of reinforcement learning approaches. An AI agent is presented, which interacts with a Tic Tac Toe game environment. As the implemented Reinforcement Learning approach QLearning was chosen. The deployed AI agent takes the values stored in the QTable to compute the next move in the considered game state. After each game, the game result is used as reward to update the values in the QTable. This update process is continuing until the learning phase will end. The insights gained are used to give an outlook to different deployment scenarios for reinforcement learning.
Webex-Link zur öffentlichen Verteidigung: