Graduate School of Informatics

This thesis explores the critical role of people and culture-related capabilities in the success of Advanced Data Analytics (ADA) projects, addressing a gap in current literature that predominantly focuses on technical aspects. By conducting a systematic literature review and semi-structured interviews, the study identifies and categorizes these capabilities, integrating them into structured processes tailored from the People Capability Maturity Model (Curtis et al., 2009). The research contributes actionable frameworks and practices to enhance workforce readiness, collaboration, and organizational culture, enabling businesses to align ADA initiatives with strategic goals and achieve sustainable success.

Date: 10.01.2025 / 13:30 Place: A-212

Somatic mutations in non-coding regions can disrupt transcription factor (TF)-DNA interactions, affecting gene regulation and contributing to cancer. This thesis introduces an in silico pipeline to assess the impact of these mutations on TF binding affinities. Using k-mer-based linear regression models trained on ChIP-seq and PBM data for 403 TFs, we analyzed somatic mutations in 560 breast cancer samples. Predicted TF binding changes were classified as gain or loss of function and linked to oncogene and tumor suppressor dysregulation using enhancer-target gene maps. Signature-specific and statistical analyses highlight distinct patterns, providing insights into the regulatory role of mutations in cancer.

Date: 07.01.2025 / 10:00 Place: A-212

The thesis discusses optimizing alarm management systems in manufacturing. It introduces a novel data-driven method using the Tennessee Eastman Process to enhance alarm parameters, aiming to reduce missed critical alarms and improve process safety. Key contributions include associating disturbances with alarms, creating an alarm simulation platform, and optimizing alarm parameters. The study highlights the trade-off between alarm reaction delay and the number of alarms, emphasizing the importance of timely operator responses.

Date: 10.01.2025 / 14:45 Place: II-06

This thesis introduces a methodology for training context-invariant autoencoders using unsupervised domain adaptation to enhance model generalizability under varying contexts. By employing domain-adversarial training and data augmentation, the approach extracts domain-invariant representations while disregarding contextual variations. Experiments utilize the CARLA simulator, generating diverse image datasets across weather conditions and times of day. The proposed framework improves reconstruction loss and feature robustness, demonstrating its efficacy in achieving reliable machine learning performance in dynamic environments. The study emphasizes the utility of domain adaptation techniques in addressing domain shifts, offering a foundation for robust applications in autonomous systems.

Date: 06.01.2025 / 14:30 Place: A-212

The thesis study "Investigating The Semantic Similarity Effect on Delayed Free Recall Using Word Embeddings," investigates how the semantic proximity effect, alongside the temporal proximity effect on delayed free recall. The current study uses fastText and word2vec for methodological purposes to outline the underlying cognitive mechanisms leading to the process of memory retrieval. By investigating the interplay between word meanings and memory performance, this study contributes to Cognitive Science and Psychology specifically in investigating language processing and human memory.

Date: 10.01.2025 / 13:00 Place: B-116

This thesis presents a Deepfake Detection System that leverages public blockchain and collective intelligence to address the growing threat of digital misinformation. Implemented on the Ethereum Sepolia testnet, the system combines human collaboration and decentralized technology to detect deepfakes independent of their generation methods. Using smart contracts ensure transparency, fairness, and scalability by automating voting processes and adjusting user credibility based on voting accuracy. The system builds trust and accuracy by normalizing user influence and promoting open participation. This study demonstrates the system’s robustness, scalability, and ability to combat misinformation, while laying the foundation for blockchain-based verification in other fields.

Date: 07.01.2025 / 14:00 Place: A-212

This thesis examines clozapine effects on cognitive function and brain connectivity in schizophrenia with the M1 muscarinic receptor polymorphism (rs2067477). Differences in cortical activity and connectivity between genotypes were assessed using an N-back working memory task and functional near infrared spectroscopy (fNIRS). Wild-type individuals exhibited higher cortical activation during the task, but had lower functional connectivity in the frontotemporal network compared to non-wild-types. The findings suggest different compensatory mechanisms by highlighting the genetic effects of clozapine-related neural responses and provide valuable information for personalized treatment approaches in schizophrenia.

Date: 10.01.2025 / 10:00 Place: A-212

This thesis addresses the challenge of concept drift in machine learning, where evolving data patterns reduce model relevance and performance. This research proposes a dynamic system that detects and adapts to new concepts by developing tailored models for each concept. It includes leveraging ensemble strategies and mitigating class imbalances with synthetic data. By using detection techniques based on differences between datasets and performance metrics, and different prediction techniques that take account of the concept of the datasets that will be predicted this research aims to enhance model adaptability in dynamic environments, providing a comprehensive framework to tackle concept drift.

Date: 27.12.2024 / 14:00 Place: A-212

Date: 20.12.2024 / 13:00 Place: A-212