67471 research outputs found
Sort by
Exploring the Role of Digital Platforms in Logistics Flexibility: A Qualitative Analysis in the German Manufacturing Industry
No full textIn an era of business volatility, logistics flexibility has become crucial. However, the role of digital platforms in this context remains under-explored. This study undertakes an investigation into how digital platforms enhance logistics flexibility. It employs a theory elaboration approach grounded in dynamic capabilities. Drawing from a series of 16 semi-structured interviews conducted within the context of the German manufacturing sector, the study offers insights into how digital platforms support dynamic and structural flexibility. Specifically, the study identifies relationships between digital platforms, dimensions of logistics flexibility, and dynamic capabilities, thereby offering propositions that could guide further research. In terms of practical implications, the study offers guidance for leveraging digital platforms to enhance flexibility in supply chains. This research contributes to a more profound understanding of digital platforms’ strategic role in supply chain management
Railroad Track Defect Detection Using YOLO Models: A Comparative Study
Get PDFDetecting substructure defects is critical to ensuring track stability, preventing derailments, and avoiding costly slow orders. This study explores the use of surficial thermal and optical imagery captured by uncrewed aerial vehicle (UAV)-mounted sensors to detect subsurface issues such as ballast contamination, mud pumping, and differential settlement. We evaluate real-time object detection models You Only Look Once (YOLO)v5 through YOLOv10 across optical, thermal, and combined datasets. Results show that YOLOv10 achieves state-of-the-art performance, reaching 0.915 in precision and 0.918 in recall on optical imagery, and leading across metrics on thermal and combined modalities as well. Zero-shot cross-modal inference experiments, however, reveal limited transferability: models trained on thermal imagery generalize poorly to optical data, and vice versa, with mAP@50 below 0.25 in both directions. Overall, our results establish both a strong baseline for UAV-based defect detection and a roadmap for future research into more generalizable multimodal railway inspection models
When Surgeons Skip Water: Exploring the Feasibility of Real-Time Sweat Sensor Data and Its Potential Value in the Operating Room
Get PDFSurgeons maintain intense focus and accuracy while completing complex tasks under significant stress, opening the door for research and interventions to optimize surgical performance. Dehydration is a potential target as it is common among surgeons and linked to worsened mood, focus, and psychomotor skills. Objectively monitoring sweat rate and relative dehydration has potential to elucidate relationships between hydration and cognitive performance in the peri-operative environment. This mixed-methods approach sought to assess the feasibility and value of continuously monitoring surgeon hydration with biosensors. We found that current hydration tracking technology is unable to reliably determine skin-based fluid loss in the operating room. Additionally, survey data from 24 surgeons revealed an average daily non-caffeinated fluid intake of 47 ounces per day, which is below recommended amounts. Semi-structured interviews (n=14) revealed desire for improved real-time hydration data due to the frequency of reported “purposeful dehydration”, negative dehydration symptoms, and self-imposed break avoidance
Spatio-Temporal Energy Flexibility of Data Centers: Modeling the Impact on the Western Interconnection
Get PDFData centers are rapidly becoming major electricity consumers due to the growth of digital infrastructure and AI, but they can transition from a grid challenge to a grid flexibility solution through workload scheduling, distributed power resources, and advanced cooling systems. This paper presents an analysis of the impacts of data center energy flexibility on the operation of the Western Interconnection. The energy flexibility of data centers is modeled through coordinated IT workloads with spatio-temporal flexibility, cooling systems enhanced with thermal energy storage, and integrated on-site energy resources including solar generation, small modular reactors, and battery storage systems. A multi-regional unit commitment framework is developed to integrate data center energy flexibility into the Western Interconnection while enforcing inter-regional transmission limits and operational constraints. The numerical results show that spatio-temporal flexibility of data centers significantly reduces the operation cost, while supporting a diverse generation mix in the Western Interconnection
Designing Conversational AI for Social Robots in Corporate Contexts: A Case Study on Customizing LLMs through Action Research
Get PDFThis paper presents a case study on customizing Large Language Models (LLMs) for social robots in corporate environments. Over a year-long collaboration with an enterprise, we explored how LLMs can be integrated into multimodal assistants that operate across both embodied robot platforms and flexible digital interfaces. Using three iterative action research cycles, we developed and evaluated a scalable customization framework. We introduce Knowledge Interaction Distillation (KID), a method for generating synthetic training data from simulated user interactions, enabling efficient finetuning of task-specific models. Through participatory design, controlled experiments, and field deployment, we identified core requirements and validated a distributed system architecture that supports multimodal interaction. Our findings offer actionable insights for deploying conversational AI for social robots not as isolated systems, but as part of a broader ecosystem aligned with real-world workflows and organizational constraints
Optimizing Class Distributions for Bias-Aware Multi-Class Learning
Get PDFWe propose BiCDO (Bias-Controlled Class Distribution Optimizer), an iterative, data-centric framework that identifies Pareto-optimized class distributions for multi-class image classification. BiCDO enables performance prioritization for specific classes, which is useful in safety-critical scenarios (e.g. prioritizing 'Human' over 'Dog'). Unlike uniform distributions, BiCDO determines the optimal number of images per class to enhance reliability and minimize bias and variance in the objective function. BiCDO can be incorporated into existing training pipelines with minimal code changes and supports any labelled multi-class dataset. We have validated BiCDO using EfficientNet, ResNet and ConvNeXt on CIFAR-10 and iNaturalist21 datasets, demonstrating improved, balanced model performance through optimized data distribution
Influenceable HR Personnel: An Empirical Investigation of Social Influence when Recruiting Female IT Professionals
Get PDFHuman resource (HR) personnel operate within a complex network of organizational stakeholders whose diverse perspectives can influence recruitment strategies, including those aimed at attracting female IT professionals. This study contributes to the existing body of research by identifying the key stakeholders who influence human resource personnel’s decision-making. Employing a fuzzy-set qualitative comparative analysis, the findings reveal that different configurations of influence impact HR personnel’s intention to recruit female IT professionals. Based on the findings, the study offers both research contributions and practical implications
Trust Anchors as Pillars of Humanistic Governance in AI-Driven Knowledge Management
Get PDFAs knowledge management systems increasingly rely on artificial intelligence, it becomes imperative to reassert the significance of human agency in shaping, guiding, and governing these technologies. This paper introduces the concept of Trust Anchors, a novel theoretical framework developed which identifies critical human touchpoints that stabilize and ethically ground AI-driven knowledge processes. Drawing from traditions in self-monitoring ethics, organizational culture, and social governance, Trust Anchors function as multi-level mechanisms that embed individual, organizational, and societal principles into AI systems enabling responsible intervention, contextual judgement, and alignment with human values. This paper proposes a humanistic stewardship governance approach to digital knowledge ecosystems by identifying the relationship between humans and machines. It contributes to the discourse on innovation of AI in KM by situating knowledge translation as a dynamic, relational process in which trust, culture, and care are essential for transformation and sustainable knowledge management