This study introduces a testing framework for Deep Neural Network (DNN) models to identify fault-revealing data and understand the causes of failures. We prioritized test inputs based on model uncertainty, and with the proposed meta-model-based approach, we enhanced the effectiveness of test data prioritization. Moreover, distribution-aware test datasets are generated by initially focusing on in-distribution data and subsequently including out-of-distribution data. Finally, we employed post-hoc explainability methods to pinpoint the causes of incorrect predictions after test executions. Evaluations in the image classification domain show that uncertainty-based test selection significantly improves the detection rate of DNN model failures.

Date: 10.07.2024 / 15:30 Place: A-212

The most interactive field of digital transformation is data science, as it entails a longtime active collaboration among multiple partners. Data scientists seek domain expertise to understand the structure and environment of the data while business users take pains with concepts to exploit analytical solutions. This thesis presents the conceptual design and implementation of CoDS (Collaborative Data Science Framework) as a knowledge management system on which business and data details, modeling procedures, and deployment steps are shared. It mediates and scales ongoing projects, enriches knowledge transfer among stakeholders, facilitates ideation of new products, and supports the onboarding of new developers.

Date: 06.06.2024 / 11:00 Place: II-06

The study aimed to understand the impact of IT governance mechanisms on IT agility. Results indicated that relational IT governance mechanisms significantly enhance IT agility, unlike structures and processes, which were not significant. Relational mechanisms involve top management support, cross-functional training, and a clear IT role within the firm, fostering mutual understanding and integration between IT and business units. While structures and processes are essential for compliance, they may not independently achieve agility. Future research should investigate how these formal mechanisms enable or inhibit agility. Additionally, a hybrid IT governance model may balance centralization and decentralization, offering both efficiency and flexibility.

Date: 22.05.2024 / 13:00 Place: A-212

Due to the increased use of hyperspectral remote sensing payloads, there has been a rise in the number of hyperspectral remote sensing image archives, resulting in a massive amount of collected data. This highlights the need for a content-based image retrieval system that can manage and enable the use hyperspectral remote-sensing images efficiently. A novel CBHIR system is proposed that aims to define global hyperspectral image representations based on a semantic approach to differentiate background and foreground image content considering both spatial and spectral information. In this way, two spectral content dictionaries are used in the process of modeling hyperspectral images.

Date: 24.01.2024 / 14:00 Place: B-116

It is significant to comprehend the basics of decision making behaviour because people make decisions in their everyday lives. The purpose of this research is to understand individuals’ decision making behaviour under risk and uncertainty using computational cognitive modeling and neuroscience perspectives. Results related to behavioural and neural data analyses and computational cognitive modeling utilizing the collected data from experiments provide explanations for the mechanisms behind decision making under risk and uncertainty cases.

Date: 26.01.2024 / 11:00 Place: A-212

The most interactive field of digital transformation is data science, as it entails a longtime active collaboration among multiple partners. Data scientists seek domain expertise to understand the structure and environment of the data while business users take pains with concepts to exploit analytical solutions. This thesis presents the conceptual design and implementation of CoDS (Collaborative Data Science Framework) as a knowledge management system on which business and data details, modeling procedures, and deployment steps are shared. It mediates and scales ongoing projects, enriches knowledge transfer among stakeholders, facilitates ideation of new products, and supports the onboarding of new developers.

Date: 22.01.2024 / 13:00 Place: II-06

Organizations strive to improve their digital transformation (DX) maturity for market success, utilizing maturity structures such as maturity index. However, these structures face limitations, revealing a research gap. Therefore, this thesis introduces a novel self-diagnostic tool called the DX maturity index (DX-MI) using design science research. DX-MI assists organizations in measuring and advancing their DX maturity. It has a hierarchical structure that includes dimensions, sub-dimensions, and metrics, all underpinned by an assessment approach grounded in evidence or objective quantifiable metrics. Multiple case studies were conducted to check the applicability and usability of the DX-MI, confirming its effectiveness and practicality.

Date: 22.01.2024 / 14:30 Place: II-06

Steel production serves as the backbone of countless infrastructure projects and industrial applications worldwide. In order to maintain and improve its productivity, quality and environmental sustainability, hyperspectral imaging is a promising technology for steel industry.  A novel, non-destructive approach is presented to quantify the free lime content in steel slag by utilizing an integrated algorithm applied to hyperspectral images. This method includes spectral unmixing for mixture component quantification and endmember extraction of mixture. Methodology involved various experiments with both fresh and six-month-aged steel slag, demonstrating its accuracy compared to the Rietveld Analysis of X-ray Diffraction patterns.

Date: 11.01.2024 / 09:30 Place: B-116

State-of-the-art omics technologies use network-based contextualization methods to give molecular information about different biological contexts, like disease states, patients, and drug changes. In the beginning, this thesis identified challenging issues such as missing points in contextualization, hidden knowledge in omics datasets, bias in reference networks, and noisy interactions with highly connected nodes or hubs. Subsequently, to address these challenges, we developed pyPARAGON (PAgeRAnk-flux on Graphlet-guided network for multi-Omics data integratioN). Also, a novel tool, pyPARAGON, contextualized patient datasets by inferring patient-specific networks and complex diseases by constructing disease models, namely breast cancer and autism spectrum disorders.

Date: 22.01.2024 / 14:00 Place: A-108

Despite impressive achievements made by machine learning algorithms (especially in deep learning), they are easily tricked by modified input data. Adversarial attacks target machine learning models severely. Adversarial training is an effective method against adversarial attacks, but it is not suitable for network domains due to network flow characteristics. In this thesis, the autoencoder's reconstruction error is used for detecting adversarial attacks. The IDS model, RAIDS is proposed and achieves respectful results against adversarial attacks.

Date: 17.01.2024 / 14:00 Place: A-108