14878 research outputs found
Sort by
Exploring the impact of explainable heterogeneous fusion performance for target tracking
No full textThe prolific use of deep learning models in the information age has reached a point where it is almost ubiquitous with sensor fusion. The combination of available data and the ever-increasing processing power of available hardware has made many believe that data science is just a matter of throwing obscene amounts data into a deep learning algorithm. Rather than considering factors like the quality or available quantity of the data or trying to break down the complexity of the problem the model is designed to solve, the naive use of brute force training is the chosen approach. This misconception, coupled with the inherent blackbox nature of deep learning algorithms, makes that the importance of explainability and transparency is integral to the success of machine learning, particularly with respect to using data we inherently lack an understanding of. This is especially the case for Passive Radiofrequency (P-RF) data, which despite radar having existed since the 19th century, there is a lack of literature regarding the usage of available I/Q data for detection purposes. While P-RF data has many potential benefits for detection, with our research group’s previous research and tentative patent requiring considerably less physical hardware to implement (using commercially available software defined radio dipole antennas), the ability to utilize the modality is primarily dependent on the application of AI. In this thesis, we showcase research to the detection of vehicle targets via multimodal fusion in the ESCAPE dataset. The research presented includes a comparison of over thirty multimodal fusion models for the shared objective of detection and differentiation of vehicle targets and examine the sensor data’s impact on the model’s performance. The integration of acoustic, electro optical, passive radiofrequency, and seismic data is implemented over three scenarios of data to determine the impact of different modalities using explainable AI methods. By comparing the local and global impact of each modality, as well as the F1 Score of the trained model, we can draw conclusions regarding how the modality was utilized with the benefit of the context of the training data and scenario. Computational costs of each model are considered in the context of FLoating Point Operations (FLOPs), and used to make determinations as to which modalities provided the most value to the fusion process
Healthcare utilization for balance problems in community-dwelling adults in the united states of america
No full textThis dissertation addresses healthcare utilization for balance problems in community-dwelling adults in the United States of America. Balance problems are one of the major risk factors for falls. Falls are the leading cause of mortality and chronic disability in elderly adults. The first project (review article) presented in this dissertation explored factors associated with healthcare utilization for balance problems as guided by the Andersen Healthcare Utilization Model. This model is frequently used to examine the factors leading to the use of health services. Age, sex, race/ethnicity, BMI, and comorbidities were identified as some of the factors that pre-disposed the underutilization of healthcare services for balance problems. Socioeconomic status, health insurance, and access to primary care could enable or disable healthcare utilization. The severity of balance problems, perceived illness, and its impact on daily activities were the factors that could affect the need for care. The second study used real-world nationally representative data from the National Health and Nutrition Examination Survey (NHANES) to investigate the associations suggested by the literature review. A total of 1834 adults who self-reported having balance problems in the past 12 months were included in this study. Outcome measure was whether the individual ever saw a health professional for balance problems. Only 32.13 of the individuals who reported having balance problems sought healthcare services for balance problems. Older age, lack of health insurance, not seeing a healthcare provider in the past year, and not experiencing any fall(s) in the past year had a significant association with reduced healthcare utilization for balance problems. These findings can help identify populations at increased risk of underutilization. The third study investigated the congruency between self-reported balance information and performance-based balance measures (Romberg Test of Standing Balance on Firm and Compliant Support Surfaces, RTSBFCSS), along with exploring the predictors of congruency between these balance measures. A nationally representative sample of 4939 community-dwelling adults (≥40 years) for whom self-reported balance status responses and performance-based balance examination results were available was used for this study. Of the 4939 study participants, 36.9 had evidence of balance problems on RTSBFCSS. About 7 in 10 adults with performance-based balance deficits reported no balance problems on the self-reported question. Sole reliance on self-reported information for balance screening may be inadequate. Results can help identify populations more likely to have discrepancies between balance measures
Severity grading and early detection of alzheimer’s disease through transfer learning
No full textAlzheimer’s disease (AD) is a neurological disorder that predominantly affectsindividuals aged 65 and older. It is one of the primary causes of dementia, and it contributes significantly and progressively to impairing and destroying brain cells. Recently, efforts to mitigate the impact of AD have focused with particular emphasis on early detection through computer aided diagnosis (CAD) tools. This study aims to develop deep learning models for the early detection and classification of AD cases into four categories: non-demented, moderate-demented, mild-demented, and very mild demented. Using Transfer Learning technique (TL), several models were implemented including AlexNet, ResNet-50, GoogleNet (InceptionV3), and SqueezeNet, by leveraging magnetic resonance images (MRI) and applying image augmentation techniques. A total of 12,800 images across the four classifications that were preprocessed to ensure balance and meet the specific requirements of each model. The dataset was split into 80 for training and 20 for testing. AlexNet achieved an average accuracy of 98.05, GoogleNet (InceptionV3) reached 97.80, ResNet-50 attained 91.11, and SqueezeNet 86.37. The use of transfer learning method addresses data limitations, allowing effective model training without the need for building from scratch, thereby enhancing the potential for early and accurate diagnosis of Alzheimer’s disease [1]
An exploratory study of early childhood teachers’ self-regulatory well-being and support of children’s self-regulatory development
No full textA path analysis of 92 preservice early childhood (EC) teachers’ psychosocial characteristics and how these related to their emotion self-regulation and self-reported practices for children’s self-regulatory development had significant positive results. EC teachers’ intrapersonal mindfulness, strongly, and their compassion for self and others, moderately, predicted their emotion self-regulation, which, in turn, moderately predicted their developmentally adaptive support of children’s self-regulation. EC teachers’ emotional self-regulatory well-being also strongly predicted their psychological and self-determinative well-being, and the latter strongly mediated their support of children’s self-regulation. EC teachers’ self-compassion strongly correlated with their intrapersonal mindfulness, particularly between the respective components of self-kindness and mindful-nonreactivity. Self-compassion also strongly to moderately predicted psychological and self-determinative well-being, respectively. Other significant results were that EC teachers who had more secure attachment also had more psychological well-being, self-determinative well-being, and self-compassion, especially. They also had more compassion for others, intrapersonal mindfulness, emotional self-regulatory well-being, and supportive practices for children’s self-regulatory growth. Moreover, EC teachers’ self compassion, strongly, and compassion for others, moderately, positively predicted the secure dimension of their attachment style, and negatively predicted their fearful and preoccupied dimensions. Given that compassion for self and others may be more actionable and, hence, malleable than attachment style, supporting the former not only as personal but pedagogical skills may benefit attachment, as well as emotion self-regulation and support for child self-regulation. The most consistent finding was the positive relationship of self-compassion to intrapersonal mindfulness and other psychosocial characteristics. Intentional kindness and empathy towards oneself and others mattered for self-regulatory well-being and support of children’s self-regulatory development. Policy implications included a) developing educative systems at all levels that support the whole person and self-regulatory well-being of adults and children, b) deepening whole teacher pedagogy in preservice teacher education and inservice professional learning through relational processes to synergize knowledge, belief, and practice systems, and c) normalizing self-care as integral to the helping professions, not only to prevent compassion fatigue but flourish as individuals and caregivers
