32584 research outputs found
Sort by
Quantification of Platelet-Fibrin Interactions and Expansion Microscopy of Platelet Receptor Distribution
No full textThe formation of blood clots, also known as the coagulation cascade, is driven by cascading enzymatic interactions that stop blood from leaving vessels at sites of injury while maintaining hemostasis. The platelets and fibrin present within a blood clot can be affected by mechanical and biochemical interactions that may influence overall clot strength and stiffness. A portion of these interactions includes the cross-linking of fibrin fibers by Factor XIII-A, whose activity increases the clot resistance to lysis. Additionally, the distribution of receptors upon platelet surfaces is critical to understand how platelets up and downregulate specific functions between their activated state. To investigate platelet-fibrin interactions mediated through FXIII-A presented on activated platelet surfaces, we utilized confocal microscopy and image analysis to investigate differences between platelet suspensions treated with T101, to inhibit FXIII, and the addition of extracellular FXIII-A to mimic natural physiological conditions. Additionally, the use of novel super resolution expansion microscopy was established as a method of investigating platelet receptor distributions, namely GP Ib whose function leads to platelet adhesion. Imaging revealed that the inhibition of FXIII reduces the likelihood of platelet-fibrin interactions. Analysis of individual expanded resting and activated platelets revealed that GP Ib redistributes across the platelet membrane and can exist in various patterns in activated platelets. These results indicate that FXIII is important for platelet mechano-sensing and optimized protocol for visualization of receptor distribution upon the platelet surface at the nanoscale
Using Moss as an Inexpensive Air Pollution Bioindicator for Environmental Equity in Santa Clara Country, California
Get PDFToxins pervade urban areas and pose long-term health risks to humans. I examine moss as an affordable bioindicator and assess whether moss monitor results reflect other air quality (AQ) indicators, environmental factors, and social vulnerability. Specifically, I investigate (1) whether moss is an effective atmospheric pollution indicator and (2) whether low-income or high-lead-risk neighborhoods experience greater exposure to air pollution. I set out Hypnum cupressiforme moss for two months of exposure in 28 Santa Clara County locations and then measured chromium, zinc, nickel, and the loss of chlorophyll fluorescence in the moss tissue. I gathered real-time particulate matter (PM2.5) measurements from the open data source Purple Air and used CalEnviroScreen4.0 as a source for environmental and data. To achieve my two objectives, I use Pearson’s correlations, ArcGIS bivariate analyses, and principal component analyses for univariate, spatial, and multivariate correlations among AQ measurements and social factors, and ANOVA and Kruskal-Wallis to assess the air pollution exposed by neighborhood income and lead risk form old housing. My results confirm that changes in moss chlorophyll fluorescence patterns reflect ambient PM2.5. Moss in lower-income neighborhoods exhibited a greater loss of chlorophyll fluorescence than in higher-income neighborhoods, but patterns of metals in moss tissue are less correlated with environmental pollution. This research underscores concerns about public health for low-income residents of Santa Clara Valley in the long term and suggests that more effective policies are needed to protect residents living in these polluted areas
A Novel Reinforcement Learning Method for Efficient Cross-Training Between Real and Simulated Robots
Get PDFTraining reinforcement learning policy in a simulated environment can be a go-to choice for many research topics, as a simulated environment provides flexibility and costs less than building a physical environment for training. However, policy trained in a simulated environment often fails to transfer to the real environment for problems that have a more complex dynamics. As the solution, Sim2Real was proposed and it categorizes techniques that improve the performance of the transfer from simulation to reality. Sim2Real is an area of study that focuses on utilizing simulation data to train models that can apply to real world environment. It can greatly boost the range of reinforcement learning application by lowering the bar of training effective policy, furthermore it makes tasks that are dangerous or difficult to experiment in the real world accessible. This work will propose an approach of utilizing ensemble variance from the physical and simulated ensemble to bootstrap training with limited real world data, and it will be examined and validated by a dedicated cyber-physical system. Data collected by physical robots can enhance samples collected from simulated environment with a ensemble learning approach. Enhanced Simulated data can also provide feedback to real robots’ learning process. This learning cycle can reduce the cost of training without sacrificing the model’s robustness
The Impacts of Recreational Infrastructure on Mammal Use of Riparian Zones in Northern California Regional Parks
Get PDFPublic lands have been important for the protection of riparian habitats; however, many public lands are also used for recreation, which can reduce mammal habitat use. Understanding how mammals are affected by recreation in riparian areas can help guide management decisions. The goal of this research is to understand the effect of recreation on mammal use of riparian habitats in two popular East Bay Regional Parks, Charles Lee Tilden and Wildcat Canyon, and to map the extent of that impact. Cameras were deployed in the riparian areas of a high, moderate, and low development site (Alvarado, Tilden, and Wildcat Canyon) for two seasons to observe mammal activity. GIS was used to map the distribution and impacts of recreational infrastructure. Human activity was greatest at Alvarado and then Tilden and was very low at Wildcat Canyon. Mule deer and bobcats were seen less frequently at sites with more recreation. Observations mule deer higher at night in busy parks and negatively correlated with people. The observation frequencies of coyotes and urban adapted species were highest at Tilden and Alvarado, respectively. The maps showed that infrastructure was concentrated around creeks, resulting in 21.48% of the riparian zone impacted by non-trail features and 70.02% was impacted by trail. To protect sensitive species, managers should reduce trail density and refrain from expanding infrastructure into undeveloped areas like Wildcat Canyon
Spartan Daily, November 14, 2024
Get PDFVolume 163, Issue 35https://scholarworks.sjsu.edu/spartan_daily_2024/1079/thumbnail.jp
AI Generated Text Detection & Source Identification
No full textThe detection of AI-generated text by the application of advanced machine learning techniques not only presents a promising approach toward distinguishing human-written content from machine-generated text, but also identifies the source model used for the generation of the text. This helps address the growing concerns about authenticity and accountability in digital communication. The difference between human-generated and AI-generated text lies in the core of several applications, from news media to academic integrity.It also helps in ensuring the transparency and trust in content-driven environments. However, existing models fall short in accurately detecting AI-generated text and identifying the specific AI source due to the complex nature of AI-generated text. To address this, it is essential to leverage advanced machine learning models and embedding techniques that can capture subtle linguistic and contextual patterns of the AI generated text. In this study, text classification was performed to develop classification models that distinguish AI-generated content from human-written text and further identify the specific AI model used, offering a multilayered approach to detection. The results reveal significant improvements in the detection accuracy and source identification, supporting applications in content moderation
LlamaTalk: Empowering Conversations with Retrieval-Augmented Generation
No full textThis research report talks about the implementation and a comparative study of Llama 7B model’s fine-tuning technique and Retrieval Augmented Generation (RAG) capabilities in the context of creating a reliable AI therapist. This study focuses on training these models using diverse datasets consisting of doctor-patient conversations predominantly addressing general health issues. Using a technique like fine-tuning within the Llama 7B model, the project focuses on training the model with a diverse dataset comprising doctor-patient interactions primarily addressing general health concerns. Additionally, carefully organized mental health dataset from HOPE dataset, ensuring the bot\u27s responsiveness to mental health inquiries. Through integration with a vector database like ChromaDB and RAG techniques, the model generates contextually relevant responses, specifically tailored to address mental health concerns. This interesting approach aims to bridge the gap in automated mental health support. The report highlights the methodology, challenges, and outcomes of the project, offering insights into the potential of AI based solutions in enhancing mental healthcare accessibility and suppor
The 40th Annual TEI-SJSU High Tech Tax Institute Conference on Nov. 4-5, 2024: AI in the Corporate Tax Department - Let’s See It!
Get PDFThe 40th Annual TEI-SJSU High Tech Tax Institute Conference on Nov 4-5, 2024: The Future of the Corporate Tax Department
Get PDFEffect of fermentation, malting and ultrasonication on sorghum, mopane worm and Moringa oleifera: improvement in their nutritional, techno-functional and health promoting properties
Get PDFBackground: Food processing offers various benefits that contribute to food nutrition, food security and convenience. This study investigated the effect of three different processes (fermentation, malting and ultrasonication) on the nutritional, techno-functional and health-promoting properties of sorghum, mopane worm and Moringa oleifera. Methods: The fermented and malted flours were prepared at 35°C for 48 h, and for ultrasonication, samples were subjected to 10 min at 4°C with amplitudes of 40–70 Hz. The biochemical, nutritional quality and techno-functional properties of the obtained flours were analysed using standard procedures. Results: Fermentation resulted in significantly lower pH and higher titratable acidity in sorghum and mopane worm (4.32 and 4.76; 0.24 and 0.69% lactic acid, respectively), and malting resulted in higher total phenolic content and total flavonoid content in sorghum (3.23 mg GAE/g and 3.05 mg QE/g). Ultrasonication resulted in higher protein and fibre in raw sorghum flour (13.38 and 4.53%) and mopane worm (56.24 and 11.74%) while raw moringa had the highest protein (30.68%). Biomodification by fermentation in sorghum led to higher water and oil holding capacity and increased dispersibility in the ultrasonicated samples. Ultrasonication of mopane worms led to higher water holding capacity, oil holding capacity and dispersibility. Lightness was found to be significantly higher in the fermented samples in sorghum and mopane worm. Raw moringa had the greatest lightness compared to the ultrasonicated moringa. Moringa had the most redness and browning index among all samples. Conclusion: In this study, all the investigated processes were found to have caused variations in flours’ biochemical, nutritional and techno-functional properties. Ultrasonication process was noteworthy to be the most efficient to preserve the nutritional value in sorghum, mopane worm and M. oleifera flours