arXiv subjects

Niklas Kühl

Publications and source records attributed to Niklas Kühl.

At least 19 recordsLinked to original sources

arXiv2026

Generative AI Use Cases In Real Estate Marketing: Adoption and Constraints in Germany

Generative artificial intelligence (GenAI) is changing how work is organized and performed. Real estate marketing is a prime example of this, yet evidence of GenAI in real estate agents' day-to-day practice remains scarce. In this work, we report on our insights from a German-based empirical study with eleven semi-structured interviews. GenAI is already utilized across different activities, with marketing communication being the most prominent. Concrete use cases are emergent and unevenly adopted, with writing exposé texts being the only widely established one. Interaction is predominantly human-in-the-loop: GenAI drafts, structures, and retrieves, while real estate agents curate, verify, and decide. Constraints stem less from model capability than from integration with listings and documents, data availability, and compliance in sensitive tasks. The study contributes a grounded map of existing and potential use cases and identifies tentative practical implications for adoption.

cs.AI↗

arXiv2026

I Am AdMan: A Pipeline for Automatic Generation of Personalized Advertising Imagery

Personalized marketing can increase customer engagement, satisfaction, and conversion. While existing personalization approaches have become effective at matching the right product to the right customer, the visual representation of advertisements remains generic and only weakly tailored to the individual. Prior research shows that generative artificial intelligence can improve the creation of personalized advertisements, particularly for text, and that image generation models can support scalable advertisement production. However, little research has examined how detailed customer information can be systematically translated into fully AI-generated, personalized advertising imagery at scale on a technical level. To address this gap, we propose AdMan, a multi-agent pipeline that transforms customer data into personas, generates personalized advertisement images conditioned on product reference images, and applies an LLM-based judge agent for automated quality control. We implement the pipeline with two different model configurations and evaluate it across four products, using six celebrity personas for qualitative inspection, and 100 real customer profiles, producing 1745 advertisements. The evaluation combines a qualitative expert focus group and a quantitative artifact-rate assessment. The results show that the pipeline can generate photorealistic and personalized advertisements. At the same time, performance varies substantially by product complexity and model configuration. Our findings extend the literature on AI-based personalized advertising by demonstrating the feasibility and current limitations of fully automated image generation for advertising.

cs.AI↗

arXiv2026

Enabling and Understanding Personalization in AI-Generated Advertising Imagery

Personalized marketing traditionally matches static products to customers, while dynamic creative optimization focuses mainly on AI-driven text personalization or basic product image modifications. We address this gap by developing and implementing an AI-based framework that generates personalized advertising imagery directly from customer data. We evaluate this framework in a two-stage within-subject study with N=100 participants across four products and three levels of personalization, varied by the amount and specificity of customer data used. Participants rated each image on attitude toward the advertisement, attitude toward the product, and purchase intention. Results show that participants perceive differences across personalization levels and evaluate AI-generated advertising imagery most positively at a moderate level of personalization. High personalization increases perceived personalization, which is positively associated with all three outcome measures, but also increases perceived creepiness, which is negatively associated with the outcomes and dominates the total effect.

cs.AI↗

arXiv2026

How Reliable Is the Multi-Input Heuristic for Bitcoin Address Clustering in Law Enforcement Contexts?

Address clustering is an important technique in blockchain forensics, widely employed by law enforcement to trace illicit crypto asset flows. The multi-input heuristic (MIH), which clusters addresses potentially associated with the same entity, is the most widely used. Yet, despite its broad adoption, the MIH has rarely been evaluated against reliable ground truth data. We implement a reusable evaluation framework covering nine established metrics and apply it to ground truth address-to-entity mappings obtained directly from European crypto asset service providers under legally mandated reporting obligations. When evaluation is restricted to reported addresses, the MIH appears strong at dataset level: we observe no mergers between reported services and recover same-service address pairs with recall 0.71. However, this result is driven by one large service and ignores unlabeled addresses absorbed into full clusters. Metrics that assess the full clusters show substantially lower precision and recall (0.36 and 0.44), meaning that services are often only partially recovered or embedded in larger clusters. Entity-level results further reveal near-complete failures for some services. When MIH-based clusters are used to support criminal suspicion, preliminary seizure of crypto assets to secure later forfeiture/ confiscation, or as evidence in trial proceedings, prosecutors and judges must account for the heuristic's metric-dependent and entity-dependent reliability.

cs.CR↗

arXiv2026

CollaFuse: Collaborative Diffusion Models

In the landscape of generative artificial intelligence, diffusion-based models have emerged as a promising method for generating synthetic images. However, the application of diffusion models poses numerous challenges, particularly concerning data availability, computational requirements, and privacy. Traditional approaches to address these shortcomings, like federated learning, often impose significant computational burdens on individual clients, especially those with constrained resources. In response to these challenges, we introduce the novel approach CollaFuse for distributed collaborative diffusion models inspired by split learning. Our approach facilitates collaborative training of diffusion models while alleviating client computational burdens during image synthesis. This reduced computational burden is achieved by retaining data and computationally inexpensive processes locally at each client while outsourcing the computationally expensive processes to shared, more efficient server resources. Through experiments on the common datasets CelebA, CIFAR-10, and Animals-with-Attributes2, our approach demonstrates enhanced performance while decreasing information disclosure as it reduces the necessity for sharing raw data. These capabilities hold significant potential across various application areas, including the design of edge computing solutions. Thus, our work advances distributed machine learning by contributing to the evolution of collaborative diffusion models.

cs.LG↗

arXiv2026

Reading Between the Tokens: Improving Preference Predictions through Mechanistic Forecasting

Large language models are increasingly used to predict human preferences in both scientific and business endeavors, yet current approaches rely exclusively on analyzing model outputs without considering the underlying mechanisms. Using election forecasting as a test case, we introduce mechanistic forecasting, a method that demonstrates that probing internal model representations offers a fundamentally different - and sometimes more effective - approach to preference prediction. Examining over 24 million configurations across 7 models, 6 national elections, multiple persona attributes, and prompt variations, we systematically analyze how demographic and ideological information activates latent party-encoding components within the respective models. We find that leveraging this internal knowledge via mechanistic forecasting (opposed to solely relying on surface-level predictions) can improve prediction accuracy. The effects vary across demographic versus opinion-based attributes, political parties, national contexts, and models. Our findings demonstrate that the latent representational structure of LLMs contains systematic, exploitable information about human preferences, establishing a new path for using language models in social science prediction tasks.

cs.CY↗

arXiv2026

A Modified Moving Reference Frame Method for Propeller Resolution

Accurate resolution of propeller-hull interaction is essential for predicting the self-propulsion point in ship CFD, yet motion-resolving methods such as sliding interfaces (SI) are computationally expensive, while the classical Moving Reference Frame (MRF) approach cannot capture unsteady interaction effects. Partially rotating grid methods bridge this gap by splitting the propeller rotation into a grid-resolved and an MRF component, but the abrupt transition between the rotating and stationary domains introduces discontinuities in the velocity field. This work presents a modified MRF (mMRF) formulation in which the reference-frame rotation rate is scaled by a spatially varying function that decays smoothly from unity near the propeller to zero at the domain interface, restoring velocity and pressure continuity across the boundary. The governing equations are derived and implemented in the RANS solver FreSCo$^+$, verified against the analytical Taylor--Couette solution, and applied to open-water propeller and Japan Bulk Carrier self-propulsion simulations at model scale. Both MRF and mMRF reproduce the principal integral propulsion quantities ($n$, $K_{\mathrm{T}}$, $K_{\mathrm{Q}}$, $1-t$, $1-w_{\mathrm{T}}$, $η_{\mathrm{R}}$) accurately, but the mMRF markedly reduces interface discontinuities and non-physical artifacts in the local flow field, particularly at large MRF fractions, at essentially the same computational cost.

physics.flu-dyn↗

arXiv2026

Adjoint Sensitivity Maps for Passive Flow Control Around Rotating Circular Cylinders Across a Wide Operating Envelope

Rotating circular cylinders are employed in a variety of engineering applications, one prominent example being Flettner rotors for wind-assisted ship propulsion. Besides optimizing the aerodynamic performance of the cylinder itself, passive flow-control devices placed in its vicinity offer additional potential for manipulating the resulting aerodynamic forces. The present work introduces a topology-based adjoint sensitivity analysis for rotating circular cylinders over a wide operating envelope covering Reynolds numbers from 1E+01 to 1E+07 and spinning ratios between 0 and 2 pi. Local sensitivity fields associated with drag, lift, and torque are derived using a porous-medium formulation and validated by dedicated forward simulations employing both distributed Darcy-type source terms and a sensitivity-informed passive flow-control structure. Particular emphasis is placed on the combined sign distribution of the drag and lift sensitivities, yielding intuitive design maps that directly identify regions where local momentum extraction simultaneously improves or deteriorates both objectives. A systematic investigation of the resulting sensitivity spectra reveals that the large-scale topology of the sensitivity fields is governed primarily by the spinning ratio, whereas the influence of the Reynolds number remains comparatively weak over large parts of the investigated operating envelope. The resulting sensitivity atlas provides practical design guidance for passive flow-control concepts and demonstrates that robust solutions may exist over moderate operating ranges.

physics.flu-dyn↗

arXiv2026

Identification of Beneficial and Detrimental Structure Locations Around Flettner Rotors Using Topology-Optimization-Inspired Sensitivity Fields

Flettner rotors are highly sensitive to their surrounding flow field and may be significantly affected by nearby ship structures, deck cargo, and superstructures. Assessing the aerodynamic influence of such structures during early design stages remains challenging, particularly when a large number of potential arrangements must be considered. This paper presents a topology-optimization-inspired numerical sensitivity-analysis approach for identifying beneficial and detrimental locations of additional structures around Flettner rotors. The numerical method is based on a virtual porosity formulation and evaluates the corresponding sensitivity field using a continuous adjoint framework. In contrast to classical topology optimization, the porosity field is not treated as a design variable and no optimization loop is performed. Instead, the resulting sensitivity field is interpreted as a design-support tool that indicates regions where the introduction of material is expected to improve or deteriorate a selected aerodynamic objective. The approach is demonstrated for a full-scale Flettner rotor operating at a diameter-based Reynolds number of ReD = 2E+06 and a spinning ratio of k=3. Sensitivity fields are evaluated for drag, lift, and a combined objective. Their predictive capability is assessed by positioning container stacks at locations identified as beneficial or detrimental by the sensitivity analysis and subsequently re-evaluating the aerodynamic performance of the modified configurations.

physics.flu-dyn↗

arXiv2026

Rapid Aerodynamic Assessment of Flettner Rotor Installations Using an Inviscid CFD Approach

This paper introduces an inviscid Computational Fluid Dynamics (CFD) approach for the rapid aerodynamic assessment of Flettner rotor systems on ships. The method relies on the Euler equations combined with a dynamic momentum source term to enforce rotor circulation. By avoiding near-wall refinement and relaxing time-step constraints, the approach significantly reduces computational effort, making it particularly suitable for early-stage design tasks such as parametric studies and design space exploration. Validation against potential flow theory and viscous reference simulations shows that the method captures lift-induced forces and overall aerodynamic trends reliably. The level of agreement with viscous reference data depends on the operating conditions and numerical setup. While moderate deviations are observed for lower spinning ratios and dissipative convection schemes, larger discrepancies occur at higher spinning ratios and with low-diffusion schemes, particularly in the prediction of drag and peak lift. Three-dimensional simulations, including idealized wind tunnel setups, rotor-rotor interactions, and full-scale ship applications, demonstrate that the method provides consistent qualitative trends and robust force estimates at a fraction of the computational cost of viscous CFD. This makes the approach well-suited as a fast screening tool in early design phases, where large parameter spaces must be evaluated efficiently.

physics.flu-dyn↗

arXiv2026

On the Removal of Solver-Induced Dependencies in Momentum-Weighted Interpolation for Primal and Continuous-Adjoint Flow Solvers

Momentum-Weighted Interpolation (MWI) is a key component in pressure--velocity coupling schemes on collocated cell-centered finite-volume methods for both primal and continuous adjoint formulations. In many practical implementations, MWI relies on diagonal momentum coefficients that include contributions from under-relaxation and time discretization. As a result, both primal quantities of interest and adjoint sensitivities may exhibit a non-physical dependence on solver parameters such as relaxation factors and time-step size, and no well-defined limit is obtained as these parameters approach zero. In this work, building on previous developments in discrete-consistent MWI formulations, a simple correction is proposed that removes solver-induced contributions from the diagonal momentum coefficients in the pressure-driven term. The resulting formulation preserves the original discretization while eliminating artificial dependencies on relaxation and time-stepping parameters and is applied consistently to both primal and adjoint systems. To facilitate its application, the derivation is presented in a structured, recipe-like manner that can be readily followed and transferred to different finite volume-based solver configurations. The proposed modification is assessed for a two-dimensional laminar cylinder flow and a three-dimensional turbulent ship hull flow configuration. In both cases, the uncorrected formulation leads to significant variations in forces, wake-related quantities, and shape sensitivities when solver parameters are altered, despite all simulations being iterated to converged residual levels and stable integral quantities. In contrast, the corrected formulation yields consistent results across a wide range of relaxation factors and time-step sizes.

physics.comp-ph↗

arXiv2026

Integrating Causal Machine Learning into Clinical Decision Support Systems: Insights from Literature and Practice

Current clinical decision support systems (CDSSs) typically base their predictions on correlation, not causation. In recent years, causal machine learning (ML) has emerged as a promising way to improve decision-making with CDSSs by offering interpretable, treatment-specific reasoning. However, existing research often emphasizes model development rather than designing clinician-facing interfaces. To address this gap, we investigated how CDSSs based on causal ML should be designed to effectively support collaborative clinical decision-making. Using a design science research methodology, we conducted a structured literature review and interviewed experienced physicians. From these, we derived eight empirically grounded design requirements, developed seven design principles, and proposed nine practical design features. Our results establish guidance for designing CDSSs that deliver causal insights, integrate seamlessly into clinical workflows, and support trust, usability, and human-AI collaboration. We also reveal tensions around automation, responsibility, and regulation, highlighting the need for an adaptive certification process for ML-based medical products.

cs.HC↗

arXiv2026

Smart But Not Moral? Moral Alignment In Human-AI Decision-Making

In high-stakes AI-supported decisions, considerations are not purely technical but involve moral judgments about fairness, responsibility, and harm. While prior research has focused mainly on functional or behavioral alignment, this paper argues that moral alignment may be a more fundamental dimension of human-AI decision-making. Moral alignment is defined as the perceived congruence between the values embedded in an AI system's decision logic and the moral intuitions of stakeholders. Building on Moral Foundations Theory, the paper adopts a multi-stakeholder perspective and highlights why moral (mis)alignment matters for the meaningful integration of AI in sensitive contexts.

cs.HC↗

arXiv2026

Normative Common Ground Replication (NormCoRe): Replication-by-Translation for Studying Norms in Multi-Agent AI

In the late 2010s, the fashion trend NormCore framed sameness as a signal of belonging, illustrating how norms emerge through collective coordination. Today, similar forms of normative coordination can be observed in systems based on Multi-agent Artificial Intelligence (MAAI), as AI-based agents deliberate, negotiate, and converge on shared decisions in fairness-sensitive domains. Yet, existing empirical approaches often treat norms as targets for alignment or replication, implicitly assuming equivalence between human subjects and AI agents and leaving collective normative dynamics insufficiently examined. To address this gap, we propose Normative Common Ground Replication (NormCoRe), a novel methodological framework to systematically translate the design of human subject experiments into MAAI environments. Building on behavioral science, replication research, and state-of-the-art MAAI architectures, NormCoRe maps the structural layers of human subject studies onto the design of AI agent studies, enabling systematic documentation of study design and analysis of norms in MAAI. We demonstrate the utility of NormCoRe by replicating a seminal experimental study on distributive justice, in which participants negotiate fairness principles under a "veil of ignorance". We show that normative judgments in AI agent studies can differ from human baselines and are sensitive to the choice of the foundation model and the language used to instantiate agent personas. Our work provides a principled pathway for analyzing norms in MAAI and helps to guide, reflect, and document design choices whenever AI agents are used to automate or support tasks formerly carried out by humans.

cs.AI↗

arXiv2026

The Bidirectional Relationship Between XAI and Regulation: Operationalizing XAI for the AI Act

The EU AI Act makes explainability urgent for high-risk AI systems, yet most XAI research focuses on technical metrics rather than regulatory compliance. Understanding how legal requirements reshape XAI method design is challenging: the AI Act regulates organizational relationships (providers, deployers) using legal terminology, specifies obligations without concrete technical requirements, and underrepresents end-users--the very stakeholders whose needs human-centered XAI addresses. As regulations emerge globally, human-centered XAI practitioners face both a challenge and an opportunity: regulations pull XAI research toward real-world deployment, while practitioners can actively shape how explainability enables compliance. This establishes a bidirectional relationship. Our contribution is threefold. First, we provide the first interdisciplinary analysis of XAI's role in the AI Act--conducted by a team comprising AI Act legal experts, ML engineers, and requirements engineers--on a real-world clinical decision support system. Second, we systematically align XAI stakeholder roles with AI Act legal responsibilities, revealing where explainability methods address regulatory requirements versus where additional measures are necessary. Third, we identify three key opportunities for human-centered XAI practitioners: actively defining their roles in regulatory implementation; making the user-to-affected-party relationship explicit where regulations address only provider-deployer obligations; and enabling compliance while building multi-level trust--from regulators to affected parties.

cs.CY↗

arXiv2025

Data Quality Challenges in Retrieval-Augmented Generation

Organizations increasingly adopt Retrieval-Augmented Generation (RAG) to enhance Large Language Models with enterprise-specific knowledge. However, current data quality (DQ) frameworks have been primarily developed for static datasets, and only inadequately address the dynamic, multi-stage nature of RAG systems. This study aims to develop DQ dimensions for this new type of AI-based systems. We conduct 16 semi-structured interviews with practitioners of leading IT service companies. Through a qualitative content analysis, we inductively derive 15 distinct DQ dimensions across the four processing stages of RAG systems: data extraction, data transformation, prompt & search, and generation. Our findings reveal that (1) new dimensions have to be added to traditional DQ frameworks to also cover RAG contexts; (2) these new dimensions are concentrated in early RAG steps, suggesting the need for front-loaded quality management strategies, and (3) DQ issues transform and propagate through the RAG pipeline, necessitating a dynamic, step-aware approach to quality management.

cs.AI↗

arXiv2025

PromptPilot: Improving Human-AI Collaboration Through LLM-Enhanced Prompt Engineering

Effective prompt engineering is critical to realizing the promised productivity gains of large language models (LLMs) in knowledge-intensive tasks. Yet, many users struggle to craft prompts that yield high-quality outputs, limiting the practical benefits of LLMs. Existing approaches, such as prompt handbooks or automated optimization pipelines, either require substantial effort, expert knowledge, or lack interactive guidance. To address this gap, we design and evaluate PromptPilot, an interactive prompting assistant grounded in four empirically derived design objectives for LLM-enhanced prompt engineering. We conducted a randomized controlled experiment with 80 participants completing three realistic, work-related writing tasks. Participants supported by PromptPilot achieved significantly higher performance (median: 78.3 vs. 61.7; p = .045, d = 0.56), and reported enhanced efficiency, ease-of-use, and autonomy during interaction. These findings empirically validate the effectiveness of our proposed design objectives, establishing LLM-enhanced prompt engineering as a viable technique for improving human-AI collaboration.

cs.HC↗

arXiv2025

A Survey of AI Reliance

Although artificial intelligence (AI) systems are becoming increasingly indispensable, research into how humans rely on these systems (AI reliance) is lagging behind. To advance this research, this survey presents a novel, comprehensive sociotechnical perspective on AI reliance, essential to fully understand the phenomenon. To address these challenges, the survey introduces a categorization framework resulting in a morphological box, which guides rigorous AI reliance research. Further, the survey identifies the core influences on AI reliance within the components of a sociotechnical system and discusses current limitations alongside emerging future research avenues to form a research agenda.

cs.HC↗