教学课程资源库 更多>>

关键词： Storms   Probability   Software packages   Weather forecasting   Success   Neighborhoods   Seasons   Atmospheric sciences   Computer science   Mathematics   Meteorology  

摘要： Forecasting severe convection hazards such as tornadoes and damaging straight-line winds in environments with limited atmospheric instability and high amounts of vertical wind shear has proven to be an arduous task with life-threatening consequences in the event of an erroneous forecast. In the southeastern United States, these “High-Shear, Low-CAPE” (HSLC) environments tend to occur in the cool season and frequently occur at night, times when the general public’s situational awareness of potential severe hazards is low. For these reasons, it is essential that forecasters be given as many effective tools as possible to try and successfully predict HSLC severe convection. This study examines the performance of one forecast metric which has shown success at predicting severe convection hazards in environments with large atmospheric instability, updraft helicity (UH), in HSLC environments. In this study, High Resolution Ensemble Forecast (HREF) 0-3 km ensemble maximum UH guidance is verified against Storm Prediction Center (SPC) Local Storm Reports (LSRs) over 144 six-hour periods (“events”) where HSLC severe convection was observed in the southeastern United States between 15 October and 15 April over the span of three cool seasons, 2017-18, 2018-19, and 2019-20. Neighborhood verification metrics such as the Fractions Skill Score (FSS) were employed to allow for some leniency for spatial discrepancies between the UH forecasts and the observations. Traditional verification methods including contingency table statistics were also used to determine more information about the nature of forecast errors. In addition to observing the overall performace of UH forecasts, this study also seeks to determine the relative performance between various UH thresholds in HSLC environments. On average, all verification metrics suggest that UH guidance does not perform as well as in high-CAPE environments, with mean FSS and contingency table results placing forecast performace closer to

关键词： Cameras   Deep learning   Video teleconferencing   Social interaction   Memory   Biometrics   Power   Artificial intelligence   Sensors   Neural networks   Algorithms   Surveillance   Virtual reality   Christmas   Computer science  

摘要： Automatic analysis and interpretation of human motion from visual data has been one of the most significant computer vision challenges since 1970. In recent years, deep learning has fueled the rapid advancement of computer vision topics. In particular, human motion analysis has drawn substantial attention due to its practical importance in many applications in a variety of domain including social behavior studies, medical assistance, robotics, sport analytics, and more. Human motion is one of the key parts of human social behavior and a rich source of information. We move our whole body involving head, shoulders, hands, trunk, legs, and limbs combined with facial expressions flavored with our individualized style to transmit social signals. A number of studies have suggested the existence of unique motion signatures of individuals by analyzing data obtained from KinectTM devices, and Electromyography (EMG) electrodes attached to muscles. Meaning that when we move and communicate, we tend to use our characteristic style of motion. These distinct motion patterns are attributed to behavioral and anatomical di↵erences between individuals as well as their di↵erent muscle activation strategies. This research aims at establishing a fully-automated framework to push the envelope of understanding information hidden in human motions from visual inputs and its potential applications on a set of fundamental tasks including classification, identification, and user authentication. For this purpose, we propose a number of deep learning approaches and try to tackle the problem from a data-driven perspective and figure out to what extend we would be able to model human motion signatures and see if it is possible to authenticate or identify people based on their movement pattern. Our results demonstrate an accuracy of 94.04% for human authentication and 92.62% for human identification among 10 subjects confirming that human motion conveys information regarding their identity and can

关键词： Stars & galaxies   Telescopes   Collaboration   Algorithms   Neutron stars   Pulsars   Astronomy   Astrophysics   Computer science   Optics   Physics  

摘要： Fast radio bursts (FRBs) are millisecond-duration radio transients of unknown origin. As these bursts are detectable at cosmological distances, they can be used to study the ionized plasma in the intergalactic medium. In order to harness the potential of FRBs as cosmological probes, it is important to develop a detailed understanding of the FRB population. This thesis aims to further our current understanding of two distinct aspects of the population, namely, source environments and activity levels, using the results of the CHIME/FRB project. The Canadian Hydrogen Intensity Mapping Experiment (CHIME) is a transit telescope operating in the frequency range of 400-800 MHz. The CHIME/FRB project searches for FRBs in real time over the instantaneous field of view of CHIME (~250 sq. deg.). As the telescope observes the full northern sky every day, the FRB detection pipeline can both discover new sources and monitor the ones which are known to repeat. The first CHIME/FRB catalog consists of 474 non-repeating sources and 61 bursts from 18 repeating sources. These bursts are well suited for population studies as they have been detected with a uniform selection function. We present algorithms to determine the on-sky exposure and sensitivity variations for the CHIME/FRB system. Using these algorithms, we generate full-sky maps of the total exposure during periods of nominal sensitivity. The estimated exposure can be used to characterize activity levels for known FRB sources. We demonstrate this using a study which uses the exposure information to validate the discovery of a periodic modulation in the activity of the repeating FRB source, FRB 20180916B. We then present a multi-band observation campaign of FRB 20180916B with the CHIME/FRB system, the Green Bank telescope (GBT) and the Low Frequency Array (LOFAR). We report on the detection of seven bursts in the frequency range of 300-400 MHz with the GBT. Knowing that the circumburst environment is optically thin to free-free

关键词： Computer science   Artificial intelligence  

摘要： Shared mobility systems regularly suffer from an imbalance of vehicle supply within the system, leading to users being unable to receive service. If such imbalance problems are not mitigated some users will not be serviced. There is an increasing interest in the use of reinforcement learning (RL) techniques for improving the resource supply balance and service level of systems. The goal of these techniques is to produce an effective user incentivization policy scheme to encourage users of a shared mobility system to slightly alter their travel behavior in exchange for a small monetary incentive. These slight changes in user behavior are intended to over time increase the service level of the shared mobility system and improve user experience. In this thesis, two important questions are explored: (1) What state-action representation should be used to produce an effective user incentive scheme for a shared mobility system? (2) How effective are reinforcement learning-based solutions on the rebalancing problem under varying levels of resource supply, user demand, and budget? Our extensive empirical results based on data-driven simulation show that:1. A state space with predicted user behavior coupled with a simple action mechanism produces an effective incentive scheme under varying environment scenarios.2. The reinforcement learning-based incentive mechanisms perform at varying degrees of effectiveness under different environmental scenarios in terms of service level.

关键词： Computer science   Ophthalmology   Biomedical engineering   Patients   Quality of life   Eye movements   Simulation   Telemedicine   Macular degeneration   Diabetic retinopathy   Metric system   Calibration   Medical research   Virtual reality  

关键词： Computer science   Geographic information systems   Artificial intelligence  

摘要： Trajectories are spatio-temporal data that represent traces of moving objects, such as humans, migrating animals, vehicles, and tropical cyclones. In addition to the geo-location information, a trajectory data often contain other (non-spatial) features describing the states of the moving objects. The time-varying geo-location and state information would collectively characterize a trajectory dataset, which can be harnessed to understand the dynamics of the moving objects. This thesis focuses on the development of efficient and accurate machine learning algorithms for forecasting the future trajectory path and state of a moving object. Although many methods have been developed in recent years, there are still numerous challenges that have not been sufficiently addressed by existing methods, which hamper their effectiveness when applied to critical applications such as hurricane prediction. These challenges include their difficulties in terms of handling concept drifts, error propagation in long-term forecasts, missing values, and nonlinearities in the data. In this thesis, I present a family of online learning algorithms to address these challenges. Online learning is an effective approach as it can efficiently fit new observations while adapting to concept drifts present in the data. First, I proposed an online learning framework called OMuLeT for long-term forecasting of the trajectory paths of moving objects. OMuLeT employs an online learning with restart strategy to incrementally update the weights of its predictive model as new observation data become available. It can also handle missing values in the data using a novel weight renormalization strategy. Second, I introduced the OOR framework to predict the future state of the moving object. Since the state can be represented by ordinal values, OOR employs a novel ordinal loss function to train its model. In addition, the framework was extended to OOQR to accommodate a quantile loss function to improve its predic

关键词： Design optimization   Layouts   Magnetism   Washers & dryers   Genetic algorithms   Artificial intelligence   Computer science  

摘要： In recent years, an interesting trend of the design of electrical motors has been the development of rare-earth-free motors or such of reduced rare-earth content with high efficiency and low cost as rare-earth material usually suffers from high cost and unstable supply. Potential candidates include switched reluctance motors, synchronous reluctance motors and interior permanent magnet synchronous motors with ferrite magnets. In this thesis, research is focused on interior permanent magnet synchronous motors with ferrite magnets. Two types of motors were selected and designed for application in a drum drive for washing machines: the spoke-type synchronous motor and the permanent-magnet assisted synchronous reluctance motor. The spoke-type synchronous motor can achieve higher flux density in the air gap due to the flux concentrating effects of its structure. The permanent-magnet assisted synchronous reluctance motor usually render a higher saliency ratio, which means larger difference between the inductances in d- and q-axes leading to an increase of reluctance torque. An automatic optimization procedure based on parametric finite element model and different optimization algorithms was established in order to perform a fair comparison of both types of motors. The evaluations of performance were based on two typical operating points: one with high torque and low speed and the other one with low torque and high speed. Various constraints have been taken into account by adding penalty functions into the objective function, classified into priority levels labeled with different penalty coefficients. Three optimization algorithms have been considered: a genetic algorithm, the Nelder-Mead method and the quasi-Newton method, representing a stochastic algorithm, a non-derivative based deterministic and a deterministic algorithm with derivative information, respectively. Different configurations formed by combining the named algorithms have been tested and compared. Finally, t

关键词： Computer science   Information technology   Artificial intelligence  

摘要： Deep Learning allows for great advancements in computer vision research and development. An area that is garnering attention is single object tracking and multi-object tracking. Object tracking continues to progress vastly in terms of detection and building re-identification features, but more effort needs to be dedicated to data association. In this thesis, the goal is to use a graph neural network to combine the information from both the bounding box interaction as well as the appearance feature information in a single association chain. This work is designed to explore the usage of graph neural networks and their message passing abilities during tracking to come up with stronger data associations. This thesis combines all steps from detection through association using state of the art methods along with novel re-identification applications. The metrics used to determine success are Multi-Object Tracking Accuracy (MOTA), Multi-Object Tracking Precision (MOTP), ID Switching (IDs), Mostly Tracked, and Mostly Lost. Within this work, the combination of multiple appearance feature vectors to create a stronger single feature vector is explored to improve accuracy. Different types of data augmentations such as random erase and random noise are explored and their results are examined for effectiveness during tracking. A unique application of triplet loss is also implemented to improve overall network performance as well. Throughout testing, baseline models have been improved upon and each successive improvement is added to the final model output. Each of the improvements results in the sacrifice of some performance metrics but the overall benefits outweigh the costs. The datasets used during this thesis are the UAVDT Benchmark and the MOT Challenge Dataset. These datasets cover aerial-based vehicle tracking and pedestrian tracking. The UAVDT Benchmark and MOT Challenge dataset feature crowded scenery as well as substantial object overlap. This thesis demonstrates the incr

关键词： Physics   Computational physics   Computer science   Mathematics   Medical imaging   Oncology   Pharmaceutical sciences   Thermodynamics  

摘要： Radiofrequency ablation (RFA) is the preferred choice when minimally invasive methods are recommended for liver cancer treatment. Being categorized under local hyperthermia, RFA destroys tumours without surgery and instead uses high temperatures to cook the cancerous cells. This method works effectively when the target area is kept within an exact temperature range for a certain duration. However, this goal is not always easy to achieve as accurate measurement of the temperature inside a tumour is difficult. In addition, in particular for liver tumours, the outcome of RFA treatment is challenged by the presence of hepatic blood vessels dissipating heat and changing the size and shape of the ablation zone. Nonetheless, prediction of the ablation zone is feasible with the help of mathematical modelling and computer simulation. Such simulations must be fast and accurate to be practical in clinical procedures. We first overview the mathematical models of RFA of hepatic tumours and the multi-physical aspects of these models. In order to reduce the simulation time, the multi-physics model must be implemented on high performance parallel computers. We explain the limitations of the finite element method (FEM) for solving multi-physics problems on parallel computers and then present the finite element Gaussian belief propagation (FGaBP) method as an alternative to FEM for parallel multi-physics simulations. In a second step, we focus on parallel implementation of FGaBP for the solution of the multi-physics problems emerging in RFA models. Both weak and strong coupling approaches are considered. For the strong coupling, the Newton-Raphson (NR) method is implemented in parallel by calculating local Jacobian matrices for each finite element instead of explicitly forming a global Jacobian matrix. This provides an NR algorithm amenable to different parallel computing architectures. The algorithm is tested on shared-memory and distributed-memory architectures and compared to high

关键词： interactive approach   IoT computer science teaching   interactive teaching   computer science  

摘要： The traditional approach to teaching computer system and IOT involves the presentation of a large amount of theoretical and mathematical data and materials to students. The common approach is a theoretical explanation of the principles, on the whiteboard and if possible, with the use of graphical animations. The main issue in this area is raising interactivity and getting collaboration between students on project assignments. In this paper, authors are proposing more interactive approach that could be beneficial to teachers of related subjects. Presented are the results of using software and simulating tools (in combination with Arduino hardware development solution) for teaching collaboratively and interactively courses related to IoT. The paper presents ideas for implementation of practical and experimental project assignments, student engagement and collaboration. The paper also presents the results of authors after applying this approach to one generation of students. The application access with practical assignments gave markedly better results, measured through student engagement, classroom attendance and distribution of student results. In this manner, the students had a better understanding of the basics of the microcontrollers, sensors, IoT and were willing to engage in programming of microcontrollers at the lower level. Students have also shown better results in closely related courses such as the courses of operating systems. In addition, some students continued to use IoT platforms for future projects (in college or practice).