Fabian Sven Karst
6 results
Now showing 1 - 6 of 6
- Some of the metrics are blocked by yourconsent settings
Item type:Publication, SynDEc: A Synthetic Data Ecosystem(2025-01-25); ; Given the critical role of data availability for growth and innovation in financial services, especially small and mid-sized banks lack the data volumes required to fully leverage AI advancements for enhancing fraud detection, operational efficiency, and risk management. With existing solutions facing challenges in scalability, inconsistent standards, and complex privacy regulations, we introduce a synthetic data sharing ecosystem (SynDEc) using generative AI. Employing design science research in collaboration with two banks, among them UnionBank of the Philippines, we developed and validated a synthetic data sharing ecosystem for financial institutions. The derived design principles highlight synthetic data setup, training configurations, and incentivization. Furthermore, our findings show that smaller banks benefit most from SynDEcs and our solution is viable even with limited participation. Thus, we advance data ecosystem design knowledge, show its viability for financial services, and offer practical guidance for privacy-resilient synthetic data sharing, laying groundwork for future applications of SynDEcs.Type:journal articleJournal:Electronic MarketsVolume:35Issue:1Scopus© Citations 13 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Integrating Brand Identity Into AI-Based Conversational Agents: A Systematic Literature Review(2025-01-07); ; A company's brand identity is vital for fostering meaningful interactions with customers in a digital economy. As generative artificial intelligence (GenAI) becomes increasingly integrated into our daily lives, AI-based assistants are evolving into the primary channel for representing a brand's identity. Consequently, brand departments face the challenge of translating their company's brand identity into highly interactive AI-based assistants effectively. While there is diverse research on single cues that influence user perceptions, the literature lacks a comprehensive understanding of conversational cues and their impact on brand-related outcomes. Hence, we conduct a systematic literature review to identify verbal and visual cues in the design of AI-based chatbots and examine how they influence the perceptions of the assistants and the brand. We contribute a conceptual framework that summarizes the independent and dependent variables as well as their interrelationships and provides researchers and practitioners with a basis for designing and studying conversational design in the age of GenAI.Type:conference paperJournal:Hawaii International Conference on System Sciences (HICSS) - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Not Enough Data to Be Fair? Evaluating Fairness Implications of Data Scarcity Solutions(2025-01-07); ; - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Generative AI for Banks: Benchmarks and Algorithms for Synthetic Financial Transaction DataThe banking sector faces challenges in using deep learning due to data sensitivity and regulatory constraints, but generative AI may offer a solution. Thus, this study identifies effective algorithms for generating synthetic financial transaction data and evaluates five leading models - Conditional Tabular Generative Adversarial Networks (CTGAN), DoppelGANger (DGAN), Wasserstein GAN, Financial Diffusion (FinDiff), and Tabular Variational AutoEncoders (TVAE) - across five criteria: fidelity, synthesis quality, efficiency, privacy, and graph structure. While none of the algorithms is able to replicate the real data's graph structure, each excels in specific areas: DGAN is ideal for privacy-sensitive tasks, FinDiff and TVAE excel in data replication and augmentation, and CTGAN achieves a balance across all five criteria, making it suitable for general applications with moderate privacy concerns. As a result, our findings offer valuable insights for choosing the most suitable algorithm.Type:conference paperJournal:Workshop on Information Technologies and Systems (WITS)Volume:2024Issue:34 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, FinDEx: A Synthetic Data Sharing Platform for Financial Fraud Detection(2024-01-06); ; The rising number of financial frauds inflicted in the last year more than 800 billion USD in damages on the global economy. Although financial institutions possess advanced AI systems for fraud detection, the time required to accumulate a sufficient volume of fraudulent data for training models creates a costly vulnerability. Combined with the inability to share fraud detection training data among institutions due to data and privacy regulations, this poses a major challenge. To address this issue, we propose the concept of a synthetic data-sharing ecosystem platform (FinDEx). This platform ensures data anonymity by generating synthesized training data based on each institution's fraud detection datasets. Various synthetic data generation techniques are employed to rapidly construct a shared dataset for all ecosystem members. Using design science research, this paper leverages insights from financial fraud detection literature, data sharing practices, and modular systems theory to derive design knowledge for the platform architecture. Furthermore, the feasibility of using different data generation algorithms such as generative adversarial networks, variational auto encoder and Gaussian mixture model was evaluated and different methods for the integration of synthetic data into the training procedure were tested. Thus, contributing to the theory at the intersection between fraud detection and data sharing and providing practitioners with guidelines on how to design such systems.Type:conference paperJournal:Hawaii International Conference on System Sciences (HICSS) - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Synthesizing Training Data with Generative Adversarial Networks: Towards the Design of a Data-Sharing Ecosystem Platform for Fraud Detection(2023-06-02); ; Financial fraud has a severe impact on the general population. While financial institutions have technological capabilities for fraud detection using intelligent AI systems, the delay until they have collected a sufficient size of fraudulent data to train their fraud detection models creates a costly vulnerability. One major challenge for quickly training data lies in the inability to share fraud detection training data with other financial institutions, due to data and privacy regulations. Thus, we create the concept for a data-sharing ecosystem platform that addresses data anonymity concerns by creating synthesized training data based on each institution’s fraud detection training data sets. We rely on the advantages of generative adversarial networks (GAN) to quickly construct a shared dataset for all ecosystem members. Applying design science research, this paper derives design knowledge based on financial fraud detection literature, data sharing between financial institutions, GANs and modular systems theory for the design of a plat-form architecture for data-sharing ecosystems.Type:conference paperJournal:International Conference on Design Science Research in Information Systems and Technology (DESRIST)