192815 research outputs found
Sort by
Awaken Norbert Wiener
No full textA digital, net-based artwork that recreates the mathematician and cybernetics pioneer Norbert Wiener (1894–1964) as a so-called "death bot”. Commissioned by Cook as part of the research for “AI & The Paradox of Agency” international group exhibition at Bildmuseet, Umeå University, with support from the Jacob Wallenberg Fund and WASP-HS
Pose-Guided Focal Loss For Enhancing Vision Transformers In Continuous Sign Language Recognition
Get PDFContinuous Sign Language Recognition (CSLR) aims to convert sign language videos into a sequence of gloss annotations. A standard Connectionist Temporal Classification (CTC) based CSLR model comprises two primary modules: a visual encoder and a sequential decoder. This research evaluates the effectiveness of current deep learning architectures in CSLR by systematically replacing the visual encoder with CNN, 3D-CNN, and Vision Transformers (ViT), while also exploring different sequential decoders based on RNN and Transformer architectures. Findings indicate that a CNN+Transformer structure achieves the highest accuracy, whereas ViT-based models underperoms due to insufficient training data and noisy supervision. To mitigate this, the authors introduce Pose Guided Focal Loss (PGFLoss), a novel loss function that leverages pose heatmaps to refine the self-attention layers in ViT. Experiments demonstrate that PGFLoss substantially enhances ViT performance on Word Error Rate (WER), from 24.1 % to 21.1 % in PHOENIX14 dataset, closing the performance gap with leading methods. The code is available here
CUDA-X: unsupervised domain-adaptive vehicle-to-everything collaboration via knowledge transfer and alignment
Get PDFRecently emerged vehicle-to-everything (V2X) perception has revealed great potential to overcome the limitation of single-vehicle intelligence aided by vigorous interaction among on-road agents, while prior endeavors are practically developed on parameter-specific simulation or configuration-dynamic real-world setting, overlooking the transferability across various scenarios. In this article, we propose u––nsupervised d––omain-a–daptive vehicle-to-everything c–ollaboration framework dubbed CUDA-X, which is built on top of a de facto collective model with key-point information exchange and instance adaptation. Specifically, collaborative knowledge transfer (CKT) is responsible for domain-agnostic feature reconstruction from nearby car or infrastructure by spatial–channel pooling operation in an elementwise manner. To promote the candidate alignment, a brand-new bin-based location correction (BLC) provides an auxiliary supervision for cross-dataset box refinement via residual coordinate encoding (RCE), and category-aware pooling alignment (CPA) is further designed for pulling the category-specific instance closer between source and target samples. We benchmark CUDA-X against the counterparts on four prevalent cooperative perception datasets, i.e., OPV2V, V2X-Sim, V2V4Real, and DAIR-V2X: it establishes the new state-of-the-art vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) performances regardless of simulation or reality. We expect that this appealing attempt would provide an in-depth insight into domain generalization in the context of multiagent perception, and the code is publicly available soon
Adaptive and efficient authentication for VANETs: a cross-layer framework for resilient communication
No full textVehicular ad-hoc networks (VANETs) present distinct security challenges due to their highly dynamic topology, low-latency requirements, and susceptibility to wireless channel impairments, where traditional cryptographic techniques may become inefficient or insufficient for ensuring timely and robust authentication. This paper proposes a novel cross-layer authentication scheme that integrates crypto-based authentication for handshaking with tag-based physical (PHY)-layer message for re-authentication, enhancing the communication effectiveness of VANETs. Unlike traditional crypto-based approaches, the proposed scheme leverages PHY characteristics, embedding a watermarking tag into the transmitted signal, mitigating the effect of generating and transmitting a crypto-based signature with every traffic-related message. This way improves detection accuracy while maintaining low data bit error rates (BER), presenting the tradeoff between the data and the tag BERs. In addition to the informal security analysis, a formal analysis using BAN logic is conducted to rigorously evaluate the proposed scheme’s security guarantees. This analysis demonstrates the scheme’s resilience against passive and active adversaries, including message interception, modification, and replay attacks. Simulation results validate the scheme’s effectiveness, demonstrating reliable authentication and robust performance in low signal-to-noise ratios (SNR). A computation and communication comparisons show superior performance compared to traditional and competitive methods, making it well-suited for highly dynamic scenarios
Building bridges: establishing a multiple sclerosis rehabilitation research and clinical knowledge mobilization strategy
Get PDFBackground:
Evidence to guide multiple sclerosis (MS) rehabilitation and symptomatic care has grown, yet suboptimal access to care persists and uptake of evidence-based information is limited in practice. The movement of evidence into routine clinical care is not a spontaneous or linear process. Effective knowledge mobilization strategies may enhance equitable access to evidenced-based comprehensive MS care.
Methods:
To guide the development of a MS rehabilitation knowledge mobilization strategy with priorities and action items a Canadian summit was hosted to engage key stakeholders in identifying and discussing current MS rehabilitation and symptomatic care evidence and needs. This multifaceted summit included workshops, breakout groups, presentations, brainstorming, and consensus-building.
Results:
Forty-three key stakeholders participated. Varied disciplines, Canadian geographical regions, and content expertise were represented. This included early/mid/late-career researchers, healthcare providers, and people with MS. The summit process identified 18 key need statements. Participants individually rated the identified need statements on feasibility and importance, and the relationships in terms of timeliness and impact were discussed. The three top priorities were identified and focused on for action planning. Developing a best-practice guideline for MS rehabilitation was unanimously identified as the critical first step to improve access to care. Support for healthcare providers and establishing a network to support this knowledge mobilization work were the next two priorities. Priority topic areas for knowledge mobilization were fatigue, mobility, cognition, mood and emotion, and rehabilitation across the MS disease course.
Conclusion:
Knowledge mobilization priorities and key topic areas for MS rehabilitation have been identified using a collaborative process. The lessons learned from this summit will inform advocacy efforts for improved access to evidence-based comprehensive care and opportunities to support moving a sustainable MS rehabilitation knowledge mobilization agenda forward. Creating a formalized Canadian MS Rehab Knowledge Mobilization Network was an outcome of the summit, and our network will collaboratively support advancing and re-evaluating this agenda
Sex differences in outcomes after tenecteplase for minor stroke: a subanalysis of the TEMPO-2 trial
Get PDFIn this subanalysis of the TEMPO-2 (Tenecteplase Versus Standard of Care for Minor Ischaemic Stroke With Proven Occlusion) trial, a randomized clinical trial comparing tenecteplase and nonthrombolytic control in patients with minor stroke and symptomatic intracranial occlusion, we investigated sex differences in the efficacy and safety of tenecteplase. We compared outcomes after tenecteplase versus control, stratified by sex. We also compared outcomes in female versus male patients treated with tenecteplase. The primary outcome was a "responder" outcome, defined as return to baseline modified Rankin Scale score at 90 days. Secondary outcomes included the Lawton Instrumental Activities of Daily Living Scale, the EuroQol-5 Dimension, vessel recanalization, and adverse events. We used generalized linear modeling with a Poisson distribution adjusted for baseline differences to calculate adjusted risk ratios (aRR) and 95% CIs. There were 884 patients in the intention-to-treat analysis (48.9% tenecteplase, 41.5% female). Among female participants, the tenecteplase group was less likely to be a responder compared with control (63.8% tenecteplase, 73.9% control, aRR, 0.87 [95% CI, 0.76-1.00]). Among male participants, the responder outcome was similar between groups (77.5% tenecteplase, 75.4% control, 1.03 [95% CI, 0.94-1.13]). Female participants randomized to tenecteplase were less likely to be responders than male counterparts (63.8% female, 77.5% male, 0.85 [95% CI, 0.75-0.96]). Early recanalization was more frequent after tenecteplase than control in both sexes. Tenecteplase was not associated with better clinical outcomes over nonthrombolytic control in female or male patients with minor ischemic stroke, despite more frequent recanalization. Fewer women treated with tenecteplase returned to baseline function compared with men
Living upon networks: a heterogeneous graph neural embedding integrating waterway and street systems for urban form understanding
Get PDFCities are supported by multiple, interacting networks, most prominently streets, which channel movement and economic exchange, and, in many contexts, waterways, which regulate flows of goods, people, and environmental amenities. Conventional quantitative studies of urban form have tended to privilege streets alone, limiting their ability to capture the full spatial logic of the urban fabric. This paper introduces a Heterogeneous Graph Autoen-coder (HeterGAE) that jointly embeds street and waterway systems, providing a unified, graph-based representation of urban form. Using Singapore as a case study, we train HeterGAE embeddings and employ them in two downstream tasks: predicting daytime and night-time land-surface temperature (LST) and estimating resale prices of public housing. Relative to a baseline model that encodes streets only, the dual-network embeddings improve predictive accuracy by about 20% for both tasks, confirming that natural and built infrastructures make complementary contributions to urban socio-environmental processes. By capturing the interaction between street junctions and waterway nodes within a single latent space, the proposed approach provides a flexible template for GeoAI-assisted urban analytics in diverse settings. The results underscore the value of integrating heterogeneous urban networks in evidence-based planning and highlight the potential of graph-neural techniques for developing more nuanced and sustainable urban strategies
Dominant ionic currents in rabbit ventricular action potential dynamics
Get PDFMathematical models of cardiac cell electrical activity include numerous parameters,
making calibration to experimental data and individual-specific modeling challenging.
This study applies Sobol sensitivity analysis, a global variance-decomposition method,
to identify the most influential parameters in the Shannon model of rabbit ventricular
myocyte action potential (AP). The analysis highlights the background chloride current (IClb) as the dominant determinant of AP variability. Additionally, the inward rectifier
potassium current (IK1), fast/slow delayed rectifier potassium currents (IKr, IKs), sodiumcalcium exchanger current (INaCa), the slow component of the transient outward potassium current (Itos), and L-type calcium current (ICaL) significantly affect AP biomarkers,
including duration, plateau potential, and resting potential. Exploiting these results, a hierarchical reduction of the model is performed and demonstrates that retaining only six key
parameters can capture sufficiently well individual biomarkers, with a coefficient of determination exceeding 0.9 for selected cases. These findings improve the utility of the Shannon model for personalized simulations, aiding applications like digital twins and drug
response predictions in biomedical research
Data sources:
Access Repository Dashboard
Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇