240106 research outputs found
Sort by
Synthesizing Accurate Relational Data under Differential Privacy
No full textMedical data is sensitive personal data which, according to GDPR and HIPAA, necessitates regulations concerning their use. Anonymizing this data prior to research would allow for broader access, due to a lower sensitivity. Privacy-aware data synthesis has been proposed as a solution. However, current algorithms face difficulties in synthesizing medical data while maintaining privacy and utility. This is due to the structure of medical data which consists of multiple interlinked tables with high dimensional columns containing sequential aspects of the patient trajectory. The resulting number of correlations is intractable to model naively and, if relational correlations are not accounted for, the resulting data has poor utility (e.g., leads to invalid patient trajectories). In this paper, we present MARE, a relational synthesis algorithm which focuses on a set of core correlations found in relational data while pruning others. The resulting lower computational complexity allows MARE to produce accurate relational data. We showcase that MARE can synthesize multiple medical datasets, which contain sequential aspects, while maintaining utility in form of inter-table and inter-row correlations and privacy guarantees
Comparative Analysis of Indexing Techniques for Table Search in Data Lakes
No full textData lakes store vast amount of datasets of various forms collected from various sources. In this context, efficient table search is essential for identifying and integrating data to support business intelligence and machine learning pipelines. This paper explores effective methods for finding related tables using advanced table representation learning. Representation learning generates dense vector representations for tables at different levels (row, column, cell), enabling the use of advanced indexing techniques such as LSH, HNSW, and DiskANN, which speed up the core operation of approximate k-NN search within vector spaces. However, while several indexing techniques have been proposed so far, a thorough study and comparison of their effectiveness versus performance trade-offs is still missing. In this paper, we aim at shedding light on this gap. We begin by reviewing advanced vector-search techniques for table search in data lakes, followed by a detailed analysis of k-ANN indexes. Next, we present a comparison of the HNSW and DiskANN indexing techniques, comparing their internal structure, effectiveness, efficiency, and scalability. Additionally, we explore the impact of model accuracy on index performance. Our experiments include four datasets of various sizes and complexity. This study allows us to explore indexing design options, revealing the strengths and weaknesses of each, and also to identify potentially interesting future research directions.</p
Innovating Wireless Education: Visualization as a Catalyst for Student Engagement:NSF grantees poster session
No full text
Humans can use multimodal osseoperceptive sensory feedback to enhance their sensorimotor control of a robotic hand
Get PDFThe somatosensory function is essential for hand motor control, but it is lost following hand amputation. This study explores the potential of osseoperception — auditory and vibrotactile sensations evoked through bone stimulation — to be used as a multimodal supplementary sensory feedback modality for myoelectric robotic prostheses during object manipulation. By stimulating the wrist at the pisiform bone, the feedback system developed in this study conveyed to the user the two types of information from the digits of a robotic hand prosthesis: discrete contact events and continuous but transitory force feedback, evoking either auditory or vibrotactile sensations by varying the stimulation frequency. Non-disabled participants used the prosthetic hand to perform pick-and-lift tasks of a virtually fragile instrumented object with a predefined breaking threshold. The results demonstrated that when osseoperceptive feedback was provided, participants completed significantly more successful trials with improved motor coordination, enhanced task intuitiveness, and reduced cognitive workload. These findings suggest that osseoperceptive feedback, achieved by stimulating distal bones, is a reliable method to improve force control and overall subjective experience when using upper-limb prostheses, showcasing the potential of osseoperception in assistive robotic technologies.</p
The Social Zipper:Redefining the Role of Streets in Disadvantaged Housing Estates
Get PDFDanish postwar non‐profit housing estates reflect the rise of the welfare state by providing quality housing for all, regardless of income. Typically built on the outskirts of cities, these estates were shaped by modernist ideals of traffic separation and functional zoning. Today, several states face criticism for their physical and social fragmentation. In response, the Danish government introduced the Parallel Society Act in 2018, mandating mixed forms of ownership in selected estates to promote greater social and functional diversity. The Parallel Society Act has led to extensive physical changes, including the creation of new internal streets designed to reduce isolation, increase “eyes on the street,” invite visitors, and foster social interaction. These interventions represent a new planning paradigm, in which streets are reframed as “social zippers.” This article explores how such transformations are envisioned and experienced in two Danish estates: Gadehavegaard and Gellerupparken. Drawing on methodological approaches inspired by architectural anthropology and based on an ongoing long‐term study conducted by an interdisciplinary team since 2019, we examine how the role of streets as “social zippers” shapes perceptions of connectivity, safety, child‐friendliness, and livability among residents and visitors. Findings reveal ambiguous outcomes: While streets are intended to connect people and spaces, residents often perceive them as intrusions into established social structures and spatial routines. This raises critical questions about whom such interventions are designed to serve and whose everyday lives they aim to reshape. The study underscores the need for participatory, context‐sensitive approaches to avoid reproducing the fragmentation these policies seek to addres
In search of alternative futures:Co-designing a situated futuring game with an indigenous San community
No full textMainstream design futuring methods regrettably direct the practice of designing and envisioning futures towards a narrow set of possibilities shaped by a prevailing worldview. This tendency limits the exploration of a wider and more diverse range of futures. To address this limitation, alternative perspectives of the future are needed. Therefore, in this article, we present the co-design of a design futuring game, /Xoa-!khaia, with an indigenous San community in Southern Africa. The game, as a situated tool for Community-based Futuring, aims to facilitate new visions of the future by broadening and grounding technology possibilities within the local context. We illustrate how its co-design contributes to a transition from and contentation with the status quo, opening up alternative trajectories that leverage indigenous values and practices.</p
Closing the data gap:leveraging pretrained neural networks for robotic surgical assessment on limited clinical data
Get PDFBackground: In robot-assisted surgery (RAS), surgical assessment is critical for ensuring competence and achieving optimal surgical outcomes. Artificial intelligence (AI)-based assessment offers an alternative to expert-based assessment but often requires large datasets, which are challenging to obtain. Transfer learning with pretrained algorithms may offer a potential solution and could reduce the need for clinical data. This study explores the use of transfer learning with preclinical porcine data to reduce the clinical data needed for action recognition (AC) and skills assessment (SA) in RAS. Methods: Abdominal, thoracic and urologic RAS procedures were video recorded. A convolutional neural network (CNN) with a Long Short-Term Memory (LSTM) layer, initially trained using preclinical data, was applied to the clinical dataset through three strategies; (1) direct application on the clinical dataset, (2) only training the LSTM and dense layers, and (3) retraining the entire network. For comparison, a baseline model was trained from scratch on clinical data. Results: Recordings from 15 procedures were included. The baseline clinical model achieved accuracies of 82.7% (AC) and 40.8% (SA). Direct application of the pretrained network resulted in accuracies of 84.8% (AC) and 51.6% (SA). Fine-tuning the LSTM and dense layers of the pretrained network yielded accuracies of 90.1% (AC) and 60.4 (SA), while retraining all layers achieved 90.5% (AC) and 57.6% (SA). Ablation analysis demonstrated higher accuracies with less data using transfer learning, 87.9% vs. 81.6%. Conclusions: Using pretrained preclinical AI models increases the accuracy of models trained on limited clinical data and reduces the need for clinical data. Public trial registry: www.clinicaltrials.gov (ID: NCT06612606).</p
European regions transitioning to green markets: the role of related capabilities and public procurement policies
No full textThe sustainability transition remains high on the European policy agenda, with an emerging understanding that focusing on green technologies is not enough to achieve disruptive sustainability. An overall green transformation of current systems of production and consumption also requires market formation processes whereby green markets become viable economic opportunities for regions to specialize in. In this study, we draw on insights from evolutionary economic geography and geography of transitions to understand how regions develop green market specializations. To do so, we investigate two key sets of factors. First, we consider the evolutionary capability development process whereby new specializations emerge from existing related regional capabilities, in a path-dependent way. Second, we account for green public procurement initiatives to capture path-creation efforts in the form of deliberate regional policy directed towards green market formation. Our empirical analysis focuses on European regions in the period 2000–2020. We employ original trademark-based metrics to capture regional specializations in green markets and combine them with patent data to construct relatedness linkages between technologies and markets. Our results reveal that only a few regions were able to develop specializations in green markets. We find that both prior capabilities in related technological domains and markets were positively associated with the emergence of these regional specializations. In addition, we also find that green public procurement was positively associated with the emergence of regional green market specializations. Our findings bear relevance for policy and research alike.</p