Gilbert Fridgen

Joaquin Delgado Fernandez, Martin Brennecke, Tom Barbereau et al.

Restrictive rules for data sharing in many industries have led to the development of federated learning. Federated learning is a machine-learning technique that allows distributed clients to train models collaboratively without the need to share their respective training data with others. In this paper, we first explore the technical foundations of federated learning and its organizational opportunities. Second, we present a conceptual framework for the adoption of federated learning, mapping four types of organizations by their artificial intelligence capabilities and limits to data sharing. We then discuss why exemplary organizations in different contexts - including public authorities, financial service providers, manufacturing companies, as well as research and development consortia - might consider different approaches to federated learning. To conclude, we argue that federated learning presents organizational challenges with ample interdisciplinary opportunities for information systems researchers.

Afshin Khadangi, Amir Sartipi, Igor Tchappi et al.

Art, as a universal language, can be interpreted in diverse ways, with artworks embodying profound meanings and nuances. The advent of Large Language Models (LLMs) and the availability of Multimodal Large Language Models (MLLMs) raise the question of how these transformative models can be used to assess and interpret the artistic elements of artworks. While research has been conducted in this domain, to the best of our knowledge, a deep and detailed understanding of the technical and expressive features of artworks using LLMs has not been explored. In this study, we investigate the automation of a formal art analysis framework to analyze a high-throughput number of artworks rapidly and examine how their patterns evolve over time. We explore how LLMs can decode artistic expressions, visual elements, composition, and techniques, revealing emerging patterns that develop across periods. Finally, we discuss the strengths and limitations of LLMs in this context, emphasizing their ability to process vast quantities of art-related data and generate insightful interpretations. Due to the exhaustive and granular nature of the results, we have developed interactive data visualizations, available online https://cognartive.github.io/, to enhance understanding and accessibility.

Afshin Khadangi, Amir Sartipi, Igor Tchappi et al.

The tension between data privacy and model utility has become the defining bottleneck for the practical deployment of large language models (LLMs) trained on sensitive corpora including healthcare. Differentially private stochastic gradient descent (DP-SGD) guarantees formal privacy, yet it does so at a pronounced cost: gradients are forcibly clipped and perturbed with noise, degrading sample efficiency and final accuracy. Numerous variants have been proposed to soften this trade-off, but they all share a handicap: their control knobs are hard-coded, global, and oblivious to the evolving optimization landscape. Consequently, practitioners are forced either to over-spend privacy budget in pursuit of utility, or to accept mediocre models in order to stay within privacy constraints. We present RLDP, the first framework to cast DP optimization itself as a closed-loop control problem amenable to modern deep reinforcement learning (RL). RLDP continuously senses rich statistics of the learning dynamics and acts by selecting fine-grained per parameter gradient-clipping thresholds as well as the magnitude of injected Gaussian noise. A soft actor-critic (SAC) hyper-policy is trained online during language model fine-tuning; it learns, from scratch, how to allocate the privacy budget where it matters and when it matters. Across more than 1,600 ablation experiments on GPT2-small, Llama-1B, Llama-3B, and Mistral-7B, RLDP delivers perplexity reductions of 1.3-30.5% (mean 5.4%) and an average 5.6% downstream utility gain. RLDP reaches each baseline's final utility after only 13-43% of the gradient-update budget (mean speed-up 71%), all while honoring the same ($ε$, $δ$)-DP contract and exhibiting equal or lower susceptibility to membership-inference and canary-extraction attacks.

Timothée Hornek Amir Sartipi, Igor Tchappi, Gilbert Fridgen

Accurate electricity price forecasting (EPF) is crucial for effective decision-making in power trading on the spot market. While recent advances in generative artificial intelligence (GenAI) and pre-trained large language models (LLMs) have inspired the development of numerous time series foundation models (TSFMs) for time series forecasting, their effectiveness in EPF remains uncertain. To address this gap, we benchmark several state-of-the-art pretrained models--Chronos-Bolt, Chronos-T5, TimesFM, Moirai, Time-MoE, and TimeGPT--against established statistical and machine learning (ML) methods for EPF. Using 2024 day-ahead auction (DAA) electricity prices from Germany, France, the Netherlands, Austria, and Belgium, we generate daily forecasts with a one-day horizon. Chronos-Bolt and Time-MoE emerge as the strongest among the TSFMs, performing on par with traditional models. However, the biseasonal MSTL model, which captures daily and weekly seasonality, stands out for its consistent performance across countries and evaluation metrics, with no TSFM statistically outperforming it.

Nils Messerschmidt, Kilian Sprenkamp, Amir Sartipi et al.

E-participation platforms are an important asset for governments in increasing trust and fostering democratic societies. By engaging public and private institutions and individuals, policymakers can make informed and inclusive decisions. However, current approaches of primarily static nature struggle to integrate citizen feedback effectively. Drawing on the Media Richness Theory and applying the Design Science Research method, we explore how a chatbot can address these shortcomings to improve the decision-making abilities for primary stakeholders of e-participation platforms. Leveraging the "Have Your Say" platform, which solicits feedback on initiatives and regulations by the European Commission, a Large Language Model-based chatbot, called AskThePublic is created, providing policymakers, journalists, researchers, and interested citizens with a convenient channel to explore and engage with citizen input. Evaluating AskThePublic in 11 semi-structured interviews with public sector-affiliated experts, we find that the interviewees value the interactive and structured responses as well as enhanced language capabilities.

Joaquin Delgado Fernandez, Sergio Potenciano Menci, Charles Lee et al.

With high levels of intermittent power generation and dynamic demand patterns, accurate forecasts for residential loads have become essential. Smart meters can play an important role when making these forecasts as they provide detailed load data. However, using smart meter data for load forecasting is challenging due to data privacy requirements. This paper investigates how these requirements can be addressed through a combination of federated learning and privacy preserving techniques such as differential privacy and secure aggregation. For our analysis, we employ a large set of residential load data and simulate how different federated learning models and privacy preserving techniques affect performance and privacy. Our simulations reveal that combining federated learning and privacy preserving techniques can secure both high forecasting accuracy and near-complete privacy. Specifically, we find that such combinations enable a high level of information sharing while ensuring privacy of both the processed load data and forecasting models. Moreover, we identify and discuss challenges of applying federated learning, differential privacy and secure aggregation for residential short-term load forecasting.

Johannes Sedlmeir, Hans Ulrich Buhl, Gilbert Fridgen et al.

The enormous power consumption of Bitcoin has led to undifferentiated discussions in science and practice about the sustainability of blockchain and distributed ledger technology in general. However, blockchain technology is far from homogeneous - not only with regard to its applications, which now go far beyond cryptocurrencies and have reached businesses and the public sector, but also with regard to its technical characteristics and, in particular, its power consumption. This paper summarizes the status quo of the power consumption of various implementations of blockchain technology, with special emphasis on the recent 'Bitcoin Halving' and so-called 'zk-rollups'. We argue that although Bitcoin and other proof-of-work blockchains do indeed consume a lot of power, alternative blockchain solutions with significantly lower power consumption are already available today, and new promising concepts are being tested that could further reduce in particular the power consumption of large blockchain networks in the near future. From this we conclude that although the criticism of Bitcoin's power consumption is legitimate, it should not be used to derive an energy problem of blockchain technology in general. In many cases in which processes can be digitised or improved with the help of more energy-efficient blockchain variants, one can even expect net energy savings.

Tobias Guggenberger, Johannes Sedlmeir, Gilbert Fridgen et al.

Private permissioned blockchains are deployed in ever greater numbers to facilitate cross-organizational processes in various industries, particularly in supply chain management. One popular example of this trend is Hyperledger Fabric. Compared to public permissionless blockchains, it promises improved performance and provides certain features that address key requirements of enterprises. However, also permissioned blockchains are still not as scalable as centralized systems, and due to the scarcity of theoretical results and empirical data, their real-world performance cannot be predicted with the necessary precision. We intend to address this issue by conducting an in-depth performance analysis of Hyperledger Fabric. The paper presents a detailed compilation of various performance characteristics using an enhanced version of the Distributed Ledger Performance Scan (DLPS). Researchers and practitioners alike can use the various performance properties identified and discussed as guidelines to better configure and implement their Hyperledger Fabric network. Likewise, they are encouraged to use the DLPS framework to conduct their measurements.

Gilbert Fridgen, Julia Heidemann

Despite several scientific achievements in the last years, there are still a lot of IT projects that fail. Researchers found that one out of five IT-projects run out of time, budget or value. Major reasons for this failure are unexpected economic risk factors that emerge during the runtime of projects. In order to be able to identify emerging risks early and to counteract reasonably, financial methods for a continuous IT-project-steering are necessary, which as of today to the best of our knowledge are missing within scientific literature.