Publications
Research publications from the FIM community.
Felicitas Kuch, Christina Nicole Lane, Anna Maria Oberländer, Manuel Johannes Sauer
(2025). ICIS 2025 Proceedings. 2.
Sparking Digital Innovation: A Framework for Employee and Generative AI Involvement
Business environments are becoming increasingly complex due to the pervasiveness of digital technologies and socio-technical interactions, complicating the initiation of digital innovations. To navigate these complexities, incumbent firms draw on insights from employees working with core products or services, referred to as Employee-Driven Digital Innovation (EDDI). However, many employers face quiet quitting (e.g., 78% in Germany), leading to untapped innovation potential. Research on Generative Artificial Intelligence (GenAI) shows it can enhance employee engagement and produce higher-quality ideas more efficiently. This interview study, therefore, explores how employees and GenAI interact during ideation in incumbents. Based on current literature and semi-structured interviews with employees, managers, and researchers, an Employee-GenAI Involvement framework with three types of GenAI and employee involvement was developed. This research contributes theoretically by deepening the understanding of the initiation phase of digital innovation and practically by identifying drivers and barriers when integrating GenAI into employee-driven ideation.
Publikation öffnen
Zeno Woywood, Jasper I. Wiltfang, Julius Luy, Tobias Enders, Maximilian Schiffer
LION 2025, Part I, LNCS 15744
Multi-Agent Soft Actor-Critic with Coordinated Loss for Autonomous Mobility-on-Demand Fleet Control
We study a sequential decision-making problem for a profit-maximizing operator of an autonomous mobility-on-demand system. Optimizing a central operator's vehicle-to-request dispatching policy requires efficient and effective fleet control strategies. To this end, we employ a multi-agent Soft Actor-Critic algorithm combined with weighted bipartite matching. We propose a novel vehicle-based algorithm architecture and adapt the critic's loss function to appropriately consider coordinated actions. Furthermore, we extend our algorithm to incorporate rebalancing capabilities. Through numerical experiments, we show that our approach outperforms state-of-the-art benchmarks by up to 12.9% for dispatching and up to 38.9% with integrated rebalancing.
Publikation öffnen
Bokstaller, Jonas; Altheimer, Julia; Dormehl, Julian Armin; Buss, Alina; Wiltfang, Jasper I.; Schneider, Johannes; and Röglinger, Maximilian
(2025). ECIS 2025 Proceedings. 3.
Enhancing ML Model Interpretability: Leveraging Fine-Tuned Large Language Models for Better Understanding of AI
Across various sectors applications of eXplainableAI (XAI) gained momentum as the increasing black-boxedness of prevailing Machine Learning (ML) models became apparent. In parallel, Large Language Models (LLMs) significantly developed in their abilities to understand human language and complex patterns. By combining both, this paper presents a novel reference architecture for the interpretation of XAI through an interactive chatbot powered by a fine-tuned LLM. We instantiate the reference architecture in the context of State-of-Health (SoH) prediction for batteries and validate its design in multiple evaluation and demonstration rounds. The evaluation indicates that the implemented prototype enhances the human interpretability of ML, especially for users with less experience with XAI.
Publikation öffnen
Franziska Burghard, Laura Heim, Thomas Kreuzer, Jana Wozar
(2025). ECIS 2025 Proceedings. 10.
Twin to win: A resource orchestration perspective on twin transformation
Organizations today need to drive both a digital transformation and a sustainability transformation. Twin transformation (TT) puts forward the idea suggests that these transformations should be integrated to leverage synergies and optimize resource utilization. While previous research has identified novel resources necessary for TT, such as dynamic capabilities, little is known about how organizations can effectively create and exploit them. We adopt a resource orchestration lens on TT to address this shortcoming. To address this shortcoming, we adopt a resource orchestration lens on TT and analyze how organizations structure, bundle, and leverage their resources for TT. Based on 20 in-depth interviews with TT industry experts, we present the TT resource orchestration pyramid, through which we unfold the processes and sub-processes of resource orchestration for TT. Our findings enhance our understanding of TT resources and contribute to the emerging body of knowledge on how organizations can drive TT. In doing so, we also provide guidance for practitioners to better manage the complexity of TT.
Publikation öffnen
Heiko Hoppe, Tobias Enders, Quentin Cappart, Maximilian Schiffer
Proceedings of the 6th Annual Learning for Dynamics & Control Conference, PMLR 242:260-272, 2024.
Global rewards in multi-agent deep reinforcement learning for autonomous mobility on demand systems
We study vehicle dispatching in autonomous mobility on demand (AMoD) systems, where a central operator assigns vehicles to customer requests or rejects these with the aim of maximizing its total profit. Recent approaches use multi-agent deep reinforcement learning (MADRL) to realize scalable yet performant algorithms, but train agents based on local rewards, which distorts the reward signal with respect to the system-wide profit, leading to lower performance. We therefore propose a novel global-rewards-based MADRL algorithm for vehicle dispatching in AMoD systems, which resolves so far existing goal conflicts between the trained agents and the operator by assigning rewards to agents leveraging a counterfactual baseline. Our algorithm shows statistically significant improvements across various settings on real-world data compared to state-of-the-art MADRL algorithms with local rewards. We further provide a structural analysis which shows that the utilization of global rewards can improve implicit vehicle balancing and demand forecasting abilities. An extended version of our paper, including an appendix, can be found at https://arxiv.org/abs/2312.08884. Our code is available at https://github.com/tumBAIS/GR-MADRL-AMoD.
Publikation öffnen