32584 research outputs found
Sort by
Exploring Electric Bicycle Safety Performance Data and Policy Options for California
No full textThis study was conducted as directed by California Senate Bill 381 (2023), which called for research to help policymakers develop effective laws and policy to support the twin goals of expanding electric bicycle use and protecting the safety of electric bicycle riders and other road users. The three major strands of findings presented in this report are (1) a review of how California and other states (and countries) regulate electric bicycle use, (2) a review of the electric bicycle safety literature, including original analysis of primary data on crashes, injuries, and deaths, and (3) strategies that the state could adopt to promote the safe use of electric bicycles. The strategies discussed include revising the way the California Vehicle Code defines and regulates electric bicycles, opportunities for improving electric bicycle safety data quality and analysis, building safe infrastructure for electric bicycling, and public education on electric bicycle rules of the road and safe riding practices
Advocating for the Field of Occupational Therapy Among High School Students
No full textThe field of occupational therapy (OT) is a healthcare field that is impactful on the lives of many individuals however there is currently a lack of awareness of the field of OT due to a lack of knowledge of and interest in OT among healthcare providers and the public (Richards and Valleé, 2020). The Aim of Research. The purpose of this project is to advocate for the field of OT and assess if an introductory presentation and hands-on lab about OT’s scope can improve the interest in OT among high school students in both rural and urban environments. Methods. The Study included an introductory workshop on the field of OT that was presented to the three high schools of Bret Harte High School (BHHS), Calaveras High School (CHS), and Pioneer High School (PHS). This workshop consisted of a 30-minute presentation followed by a hands-on lab. The hands-on lab did vary in length based on the school and introduced the students to commonly used OT assistive devices (AD). To assess the outcome of the workshop, the students completed a pre- and post-test survey that assessed their interest in and understanding of OT. Results. At the beginning of the workshop, the students had low interest in and knowledge of OT. Following the completion of the workshop, BHHS and PHS had a greater interest in and knowledge of OT with their mean interest score rising 4.79 points (p \u3c 0.05) and 4.28 (p \u3c 0.05), respectively. BHHS and PHS mean knowledge scores rose by 5.79 (p \u3c 0.001) and 4.39 (p \u3c 0.001), respectively. CHS showed a different outcome as the students’ interest in OT did rise by 2.91, but it was not statistically significant as p \u3e 0.05. The mean knowledge score for Calaveras did rise significantly with the mean score rising by 7 (p \u3c 0.001). Conclusion. High School students have a low awareness of and interest in the field of OT, however an introductory OT workshop can effectively increase their knowledge and interest in OT. Based on this, it is assumed that this project was successful at advocating for the field of OT and had shown that it can be a valuable advocacy tool for increasing awareness of OT, to scaffold students’ career interests, and is a plausible option for advocacy
Evaluating Conditional Diffusion-Based Data Augmentation in InsightFace Based Facial Recognition Systems
No full textIn the realm of facial recognition biometric systems, there are many challenges correlated with the robustness of the user templates in the system. Low sample counts in user image templates can weaken biometric recognition systems, reducing accuracy and robustness. Using generative AI, these challenges can be mitigated through the creation of realistic synthetic user images to introduce additional samples to user templates. Previous research focused on using generative adversarial networks (GAN) and variational autoencoders (VAE) as their generative models of choice. This work utilizes a conditional diffusion model, a newer technology in the realm of generative AI. The model was trained on the PubFig83 dataset, a dataset of 13838 cropped images of 83 public figures. The model the generated images by feeding in noised images from the dataset alongside an identity embedding to guide the model to generate images containing features of specified figures. A pre-trained facial feature extractor and embedding generator model pack from InsightFace was utilized. Cosine similarity alongside machine learning approaches like SVCs and KNNs utilized these embeddings for training and testing. These approaches were provided with various datasets of increasing percentages of synthetic recreated images, ranging from 0 to 100% in increments of 10%. Additional low sample experiments were also explored. The primary objective was to see if diffusion models could increase user template robustness through data augmentation. The results show promise that diffusion model synthetic recreations contain the necessary identity features for facial recognition
Assessing Pre-processing, Data Augmentation, and Traditional Edge Detection on Arabian Gulf Corals
No full textCoral reefs in the Arabian Gulf are a thermally tolerant reef system compared to others globally, making them critical models for studying climate resilience in marine ecosystems [1],[2],[3],[4]. As climate change introduces stressors such as rising sea surface temperatures and ocean acidification there is growing interest in developing tools to monitor coral health over time [5],[6],[7],[8]. Deep learning techniques such as Mask R-CNN offer an automated way to detect coral in photoquadrat images commonly used in ecological surveys [9],[10],[11],[12],[13],[14],[15],[28]. This project report covers the use of Coral Vision as a Deep Learning Computer Vision tool to support Arabian Gulf reef monitoring. Using Coral Visions’ RGB histogram equalization capacities has shown to be an effective method for preprocessing images resulting in more accurate segmentation. Moreover, RGB histogram equalization has shown to be a valid method for data augmentation to supplement small data sets and resulting in more accurate and robust models for detecting corals. Finally, this report will highlight the use of Open CV packages that can be used in a post processing capacity to feature tag masked corals and assess bleaching events in the Arabian Gulf’s coral reef system
Resume Bullet Point Enhancement Using Multi-Model Fine-Tuning: A Comparative Study
No full textApplicants often use generalized bullet points in resumes, which do not emphasize the skills and accomplishments relevant to a particular job listing, and thus reduce their chances of being selected by recruiters and Applicant Tracking Systems. This paper presents a way of automation that can be used to complement the existing resume point bullets by adding job-specific keywords, quantifiable measurements, and action verbs specific to the software-engineering job. We train and compare three large language models: Llama 3 8B, Llama 2 7B, and Gemma 7B using effective training regimes. The quality and relevance are estimated through a complex set of measures (ROUGE, BLEU, METEOR). Therefore, this work will provide job applicants with effective tools to build stronger, personalized resumes that are more in line with job requirements
Semantic and Structural Fusion for Code Smell Detection Using CodeBERT and Random Forest
No full textThis project explores automatic code smell detection in Java code using transformer based embeddings and machine learning classifiers. I utilize Code-BERT, a pre-trained transformer model, to extract semantic features from code snippets and evaluate the efficacy of Random Forest and Neural Network classifiers. The study uses the SmellyCode++ dataset, which has 107,554 Java code examples with four types of code smells: Long Method, God Class, Feature Envy, and Data Class. My methodology comprises extracting 768-dimensional embeddings with CodeBERT, training two separate classifiers, and assessing their performance on a balanced subset of 5,000 samples. Presented the metrics of accuracy, precision, recall, F1-score, and confusion matrices, and verified generalization through the application of real world Java examples. Class imbalance is addressed through stratified sampling and the implementation of a balanced training division. Both models performed similarly with an accuracy of around 78The results show that CodeBERT embeddings are good at finding semantic patterns in code structure. This makes it possible to find code smells and gives us a standard that I can use for future multi-label extensions and large-scale deployment. Beyond metrics, I recorded preprocessing decisions, hyperparameter configurations, and inference procedures to ensure complete reproducibility, and I address potential threats to validity. I delineated integration pathways into IDEs and CI systems, explored opportunities for explainability through token-level saliency, and described remediation workflows, positioning this work as a scalable foundation for multi-language, repository-level analysis
Realities of Nutrition
Get PDFWhat Is Nutrition Reality? Is it in the advertising claims that a food is natural or supplies instant energy? Is it in the myriad of dietary supplements? Is it in the diet plans for instant weight loss and glowing health? The only way to find out is to learn the basic principles of nutrition science, so you can be your own nutrition expert. Here is the true introduction to nutrition that you will read with pleasure and real understanding. It will free you from dependence on popular sources of information - often misinformation - so you can distinguish illusion from the realities of nutrition.https://scholarworks.sjsu.edu/oer/1003/thumbnail.jp
Patient Perception of Nurse Practitioner Care, Nurse Practitioner Scope of Practice and How That Influences Provider Choice
Get PDFThis project investigates patients\u27 perceptions of nurse practitioner (NP) care versus physician (MD) care, the factors influencing these perceptions, and patients\u27 understanding of NP scope of practice. A cross-sectional design was used, gathering quantitative data through an anonymous one-time survey distributed via Qualtrics, open for three months. Questions were based on The Watson Caritas Patient Score (WCPS) and the Hospital Consumer Assessment of Healthcare Providers and Systems Survey (HCAHPS). Data were analyzed using Intellectus Statistics software™, with descriptive statistics summarizing the findings. A two-tailed paired samples t-test revealed statistically significant differences in favor of patients\u27 perceptions of NP care versus MD care (p=.009). However, the multiple linear regression model did not find significant predictors for provider choice. Another key finding was patients’ limited understanding of NP scope of practice. These results suggest that while patient perceptions of NP care are more favorable than MD care, other factors beyond the variables assessed may influence provider selection. This project provides valuable insights into how patients perceive different care providers and how these perceptions influence their provider choice
Hydrogen Fuel Cell Application for Port Drayage Truck: Integrated Transportation and Energy Modeling
Get PDFThis report investigates the viability of hydrogen Fuel Cell Electric Vehicles (FCEVs) for port drayage applications, focusing on both energy modeling and economic feasibility. Port drayage—trucking operations that move goods a short distance to and from ports—is a key part of freight logistics and the critical movement of goods in American communities. This project developed a microscopic energy consumption model using second-by-second activity data from 38 Class 8 diesel drayage trucks operating in Southern California. The model emulated FCEV operations and determined an average hydrogen fuel consumption of 0.15 kg/mile across 749 trips. This data informed a levelized cost of hydrogen (LCOH) analysis under various production methods, station capacities, and fleet adoption rates. In other words, the analysis helps show conditions under which hydrogen FCEVs could become a cost-competitive and sustainable option for port drayage fleets. The project used two economic modeling approaches: a general parametric study and a comprehensive analysis using established spreadsheet tools (H2A-Lite and HDSAM-4.5). Results from the parametric study showed a concave relationship between hydrogen station capacity utilization and LCOH. Higher fleet conversion rates (e.g., 25%) significantly reduced LCOH (as low as $1.4/kg for blue hydrogen) but strained infrastructure, highlighting the need for strategic station deployment. Conversely, underutilized stations led to elevated hydrogen costs, especially with green hydrogen. This means that as station utilization increases, LCOH initially drops quickly—because more trucks using the station spreads out fixed costs, making each kilogram of hydrogen cheaper, but after a certain point, the rate of cost savings slows down—and, eventually, adding more demand may even introduce new costs (such as the need for upgrades). The comprehensive analysis reinforced these findings and showed that while grey hydrogen remains the cheapest to produce, green hydrogen—particularly from hybrid solar-wind PEM electrolysis—becomes competitive at scale. Delivery methods also impacted cost: liquid hydrogen delivery proved more cost-efficient at low utilization, while gaseous delivery was better suited for high-demand scenarios. The study concludes that achieving cost-effective hydrogen adoption for port drayage trucks hinges on optimizing station utilization, scaling infrastructure, and supporting green hydrogen technologies through policy and investment. Given the critical role of port drayage in transportation networks, this study helps lay the path forward toward decarbonizing a high-impact sector through hydrogen technology
The Normalization of Clandestine Graves in Mexico
Get PDFClandestine graves are increasing at an alarming rate in Mexico, as cartels fight for dominance over the wealthiest territories. Those who fall victim to cartels are brutally tortured into submission or are murdered and disposed of. These graves are the preferred method of body disposal by the cartels. By disposing of various victims in hidden locations, the cartels seek to erase the existence of these individuals. Victims are reduced to only bone fragments and ash. Friends and family of those who have gone missing have created a large search organization, Madres Buscadoras, to combat the helplessness of authorities. Madres Buscadoras search remote locations in hopes of finding the remains of their missing loved ones in clandestine graves. Citizens are forced to live in a violent and death-normalized environment, unable to receive aid from corrupt law enforcement, making it a humanitarian crisis. Aside from corruption, authorities do not have enough forensic anthropologists to help identify remains. In rare cases, forensic anthropology is used to analyze the remains. Forensic anthropologists can determine species, death signs, ancestry, age of death, sex, height, individual characteristics, and cause of death, which benefits investigations and brings families closure