教学课程资源库 更多>>

关键词： Environmental science   Public health   Computer science  

摘要： 3D Printing is a form of Additive manufacturing and it has become more common due to technological advancements causing the prices to decrease. 3D printers were very expensive and not applicable to the general public, but as the technology advanced the price of 3-D printers decreased, making it a popular form of manufacturing to a wide variety of consumers. However, it is unclear whether users of 3D are aware of the risks and the need for control measures. This study aimed to evaluate if environmental concentration of particles are influenced by short (1 hour) and long (4 hours) print times and the second aim is to quantify the difference in environmental concentration in rooms of two different sizes (28ft x 16ft x 10ft and 9.75ft x 9.5ft x 8 ft). Environmental sampling measured the concentration of particulate matter in various size ranges (respirable and inhalable) using a portable aerosol spectrometer. The study consisted of printing a half-face respirator with a print time of 254 minutes and a valve with a print of 63 minutes using a single fused deposition modelling 3-D printer with a PLA filament. Results showed that there were no significant main effects for either location or print length, with no significant interaction between them for PM10, PM2.5, and PM1. The most significant emissions were during the long-term print in the graduate office with a mean of 7.17 ug/m3 for PM10, 2.10 ug/m3 for PM2.5, and 1.49 ug/m3 for PM1 and during the short-term print in the fabrication lab with 17.72 ug/m3 for PM10, 6.57 ug/m3 for PM2.5, and 2.87 ug/m3 for PM1. Though the maximum concentrations did not exceed the NAAQS outdoor air quality standards, longer print time and multiple printers could lead to a higher concentration. Good dilution ventilation and local exhaust ventilation are simple measures that should be utilized to protect against printer emissions. Indeed, the results of this study indicate they are sufficient to control PM emissions from a single FDM type 3

关键词： Computer science   Computer engineering   Urban planning  

摘要： Human movement in urban areas is a complex phenomenon to analyze and understand. The geographical spread of human movement, described by locations between which individuals tend to travel, varies from city to city. The geographical spread of such locations within a city would also vary with time. This thesis explores how to formally analyze and understand human mobility using large sets of real-world data from ride-sharing services for more than a dozen cities and to derive succinct characterizations for large urban areas which account for both geographical and temporal changes. A wide range of machine learning problems require immense amounts of data. To overcome this issue for human mobility, we propose a framework which includes a stochastic graph model, and adversarial networks to generate synthetic human mobility data which conforms with the geographical and temporal characteristics observed in real data. Deriving interesting insights from large sets of data and applying those to real-world applications can be challenging. Here we also highlight how formal characterizations can be applied to applications like ride pooling and vehicle placement. We also derive performance bounds for online algorithms using city level characterizations. Finally, we provide an open-source framework to understand properties of human movement, generate synthetic human mobility data, and apply it to different what-if scenarios for real-world applications. Using such a framework would help public entities and researchers alike to thoroughly understand a complex and highly relevant phenomenon. Throughout, we link our problem formulations with other domains like social network graphs and online algorithms along with extensive empirical evaluations to increase our understanding of not just human mobility but established techniques in the linked domains.

关键词： Computer science   Artificial intelligence  

摘要： The occurrence of fog has adverse impacts to human activities, aviation and water transportation operations. These events can cause postponement and cancellation of flights and accidents between ships and vessels leading to economic costs. The design of accurate models is required to forecast the low visibility events caused by fog. However, the prediction of fog remains a challenge due to the rare occurrence of these events. In this study, deep learning networks (DNN) were proposed using the output from Numerical Weather Prediction (NWP) to predict 6 hr, 12 hr, and 24 hour lead time low visibility levels in the Corpus Christi Area. These models based on the autoencoder architecture were applied as a post-processing of deterministic NWP model and sea surface temperature (SST) output. The autoencoder was utilized to reduce the dimension of the input features to select a higher order of representation from raw data. By converting data from high dimensional space into a lower dimension, autoencoder models preserve the meaningful properties of original features in an unsupervised learning fashion. A logistic regression was also added to solve the classification problem of visibility level. Additionally, the undersampling and oversampling were also examined to solve the class imbalance problem caused by the less positive cases (fog cases). A 11 year database of NWP and SST was used to develop, train, validate, and test the proposed models to predict the occurrence of fog. The target of the models was categorized into three overlapping classes, including ≤ 1600 m, ≤ 3200 m, and ≤ 6400 m. The prediction skill of these models was evaluated by relative operating characteristic curves and seven different skill scores. The results indicate that the DNN models are able to generate good discrimination for all lead times and visibility categories. The DNN models consistently outperform the operational model used by National Weather Service with respect to five skill scores (HSS,

关键词： Electrical engineering   Computer engineering   Computer science   Algorithms   Software   Datasets   Artificial intelligence   Neural networks  

摘要： Artificial intelligence is one of the leading technologies that mimics the problem solving and decision making capabilities of the human brain. Machine learning algorithms, especially deep learning algorithms, are leading the way in terms of performance and robustness. They are used for various purposes, mainly for computer vision, speech recognition, and object detection. The algorithms are usually tested inaccuracy, and they utilize full floating-point precision (32 bits). The hardware would require a high amount of power and area to accommodate many parameters with full precision. In this exploratory work, the convolution autoencoder is quantized for the working of an event base camera. The model is designed so that the autoencoder can work on-chip, which would sufficiently decrease the latency in processing. Different quantization methods are used to quantize and binarize the weights and activations of this neural network model to be portable and power efficient. The sparsity term is added to make the model as robust and energy-efficient as possible. The network model was able to recoup the lost accuracy due to binarizing the weights and activation's to quantize the layers of the encoder selectively. This method of recouping the accuracy gives enough flexibility to introduce the network on the chip to get real-time processing from systems like event-based cameras. Lately, computer vision, especially object detection have made strides in their object detection accuracy. The algorithms can sufficiently detect and predict the objects in real-time. However, end-to-end detection of the algorithm is challenging due to the large parameter need and processing requirements. A change in the Non Maximum Suppression algorithm in SSD(Single Shot Detector)-Mobilenet-V1 resulted in less computational complexity without change in the quality of output metric. The Mean Average Precision(mAP) calculated suggests that this method can be implemented in the post-processing of other

关键词： Computer science   Hydraulic engineering   Mathematics   Nuclear engineering  

摘要： The development of high-performance computing hardware and software has stimulated the increased use of reliable predictive modeling codes in the nuclear industry for confirmatory power plant safety analysis for regulatory purposes. The rod ejection accident is one of the design basis accidents and play a major role in the licensing process of pressurized water reactors (PWR’s). The use of high-performance computing allows the simulation of normal and accidental plant conditions with high confidence. Using the thermal hydraulic code CTF and 3D neutron Kinetics code PARCS, input models are developed to simulate cycle 23 of Koeberg Nuclear Power Plant (KNPP). The codes are used in both stand alone and coupled modes to verify the inputs and cross sections with the reference data of KNPP safety reports. The stand-alone PARCS code and the coupled CTF-PARCS system were used in performing core cycle depletion simulation. The rod ejection accident was simulated for beginning of cycle (BOC), hot zero power (HZP) and end of cycle (EOC), HZP states with the worst ejected rod locations determined for the rods with largest worth at each state.

关键词： Computer science  

关键词： Computer science   Artificial intelligence  

摘要： The dissertation builds on my current research to demonstrate the connection between affect and learning through machine learning and qualitative analysis of interactions where players use a complex systems game. The project is threefold: First, I developed a thinking and learning intervention, the agent-based modeling simulation Ant Adaptation. I showed that the intervention can shift people's schemas from a process schema to an emergent schema during 10-minute museum interactions. Second, to track that conceptual change common in agent-based modeling interventions, I developed a novel form of concept mapping, constructivist dialogue mapping (CDM), which is particularly useful as a learning analytic. Through CDM, I analyzed participants’ spoken elaborations in small subsets to study how people develop their understanding of a system or museum exhibit over time. Third, through video analysis, I developed a method of affect detection to identify how participants are engaged across 17 facial action units. I map those affective-states to moments of learning using machine learning methods. Because the data source is video, the method has outsized potential for scale to predict unseen data. In short, in my work I have applied advanced methods to the design and evaluation of educational interventions.

关键词： Aerospace engineering   Robotics   Computer science  

摘要： Teams of collaborative robots or rovers are expected to play a major role in future aerospace applications, including exploration of the lunar surface. Efficient, reliable methods to determine the position, velocity, and timing status of the rovers are required to facilitate this work and reduce risk of interference. This thesis presents two distributed Kalman Filter-based estimation algorithms for fully-connected networks using time-of-flight measurements. The network is composed of static beacons at known locations and mobile rovers whose positions and velocities are to be determined. Beacons and rovers are collectively called agents. Each agent’s filter tracks the position, velocity, and time (PVT) states of all of the rovers in addition to the time states of the beacons. Knowledge of all position and velocity states enables rovers to accomplish collaborative tasks without interfering with each other; tracking time states supports the use of time-of-flight relative range measurements. This work addresses the limitations in the state of the art by introducing two decentralized algorithms where each agent can estimate the time-varying position, velocity, and clock bias/bias rate of itself and all other agents in a distributed fashion. In the first method, agents exchange full PVT state estimates, which are conservatively fused with local estimates using covariance intersection. In many applications it is desirable to reduce the amount of data transmission required for navigation purposes to free up space for other purposes. To that end, the second method reduces communication packet size by only sending sufficiently surprising measurements as determined by an event-triggering threshold. Simulation results using a low-quality clock with measurements precise on the order of centimeters show that the covariance intersection method achieves a 2D position RMSE error of 0.5 m. The event-triggering method significantly reduces required communication, but results in larger

关键词： Environmental science   Computer science   Artificial intelligence  

摘要： This dissertation will combine new tools and methodologies to answer pressing questions regarding inundation area and hurricane events in complex, heterogeneous changing environments. In addition to remote sensing approaches, citizen science and machine learning are both emerging fields that harness advancing technology to answer environmental management and disaster response questions. Freshwater lakes supply a large amount of inland water resources to sustain local and regional developments. However, some lake systems depend upon great fluctuation in water surface area. Poyang lake, the largest freshwater lake in China, undergoes dramatic seasonal and interannual variations. Timely monitoring of Poyang lake surface provides essential information on variation of water occurrence for its ecosystem conservation. Application of histogram-based image segmentation in radar imagery has been widely used to detect water surface of lakes. Still, it is challenging to select the optimal threshold. Here, we analyze the advantages and disadvantages of a segmentation algorithm, the Otsu Method, from both mathematical and application perspectives. We implement the Otsu Method and provide reusable scripts to automatically select a threshold for surface water extraction using Sentinel-1 synthetic aperture radar (SAR) imagery on Google Earth Engine, a cloud-based platform that accelerates processing of Sentinel-1 data and auto-threshold computation. The optimal thresholds for each January from 2017 to 2020 are -14.88, -16.93, -16.96 and -16.87 respectively, and the overall accuracy achieves 92% after rectification. Furthermore, our study contributes to the update of temporal and spatial variation of Poyang lake, confirming that its surface water area fluctuated annually and tended to shrink both in the center and boundary of the lake on each January from 2017 to 2020. Natural disasters cause significant damage, casualties and economical losses. Twitter has been used to support promp

关键词： Computer science   Medical imaging  

摘要： Multiple Sclerosis (MS) is a demyelinating disease of the central nervous system which affects nearly 1 million people in the US. Magnetic resonance imaging (MRI), as the most sensitive non-invasive technique, is used to visualize the disease course. In this dissertation, I present techniques to improve automated quantification of MS injuries based on MRI. Currently, focal white matter (WM) lesions and cortical gray matter (cGM) injuries are considered important biomarkers for diagnosing MS. Among these, WM lesion quantification has attracted more attention and there are many proposed methods focusing on WM lesion segmentation. However, due to the difficulty of this task, there is no fully-satisfactory method available to the community. To improve the performance of MS lesion segmentation, I develop a deep neural network and propose a data representation named 2.5D stacked slices. This approach achieved state-of-the-art performance for MS lesion segmentation and is currently listed at the top of one MS lesion segmentation challenge leaderboard. On the other hand, evaluation of cortical thinning, as one aspect of cGM injury quantification, can be improved by lesion inpainting. Towards this goal, I introduced the ``edge prior'' to guide the lesion inpainting process. With the edge prior, the deep networks are able to inpaint the images and better preserve the boundaries between tissues. Further, to make the proposed methods practical for clinical use, I explored the robustness of these deep learning models. One aspect is the models' robustness to adversarial attacks. Another aspect is the models' performance change on unseen domains (e.g., different clinic sites). To generalize the models to unseen datasets, I use many categories of data augmentations during the training process, assuming these augmentations can fill the gap between domains. Also, to address MRI sequence variations across sites, I apply modality dropout and let the models learn to make reasonable pred