教学课程资源库 更多>>

关键词： Engineering   Transportation   Computer science   Information technology   Political conventions   Datasets   Security management   Pedestrians  

摘要： Detecting pedestrian flow in different directions at at traffic-intersection has always been a challenging task. Challenges include different weather conditions, different crowd densities, occlusions, lack of available data, and so on. The emergence of deep learning and computer vision algorithms has shown promises to deal with these problems. Most of the recent works only focus on either detecting combined pedestrian flow or counting the total number of pedestrians. In this work, we have tried to detect not only combined pedestrian flow but also pedestrian flow indifferent directions. Our contributions are, 1) we are introducing a synthetic pedestrian dataset that we have created using a videogame and a real-world dataset we have collected from the street. Our dataset has small, medium and high-density pedestrians crossing a crossroad, captured from different camera height, 2) We have proposed a Pedestrian Flow Inference Model (PFIM) that is trained on the synthetic dataset first and then is tested extensively on our real-world dataset. While testing on real-world dataset, we have embraced domain adaptation to reduce the domain gap between synthetic data and real-world data. Our proposed Pedestrian Flow Inference Model (PFIM) can detect pedestrian density and flow regardless of the height of the camera in three different ways - from left to right direction, from right to left direction, and total. Combining all, It has successfully tackled the challenges mentioned above and achieved state-of the-art performances.

ISBN: (纸本)0252085663;9780252085666

关键词： Computer science   Artificial intelligence   Operations research  

摘要： Reinforcement learning (RL) is a generic paradigm for the modeling and optimization of sequential decision making. In the recent decade, progress in RL research has brought about breakthroughs in several applications, ranging from playing video games, mastering board games, to controlling simulated robots. To bring the potential benefits of RL to other domains, two elements are critical: (1) Efficient and general-purpose RL algorithms; (2) Formulations of the original applications into RL problems. These two points are the focus of this thesis. We start by developing more efficient RL algorithms. In Chapter 2, we propose Taylor Expansion Policy Optimization, a model-free algorithmic framework that unifies a number of important prior work as special cases. This unifying framework also allows us to develop a natural algorithmic extension to prior work, with empirical performance gains. In Chapter 3, we propose Monte-Carlo Tree Search as Regularized Policy Optimization, a model-based framework that draws close connections between policy optimization and Monte-Carlo tree search. Building on this insight, we propose Policy Optimization Zero (POZero), a novel algorithm which leverages the strengths of regularized policy search to achieve significant performance gains over MuZero. To showcase how RL can be applied to other domains where the original applications could benefit from learning systems, we study the acceleration of integer programming (IP) solvers with RL. Due to the ubiquity of IP solvers in industrial applications, such research holds the promise of significant real life impacts and practical values. In Chapter 4, we focus on a particular formulation of Reinforcement Learning for Integer Programming: Learning to Cut. By combining cutting plane methods with selection rules learned by RL, we observe that the RL-augmented cutting plane solver achieves significant performance gains over traditional heuristics. This serves as a proof-of-concept of how RL can be co

摘要： Autism Spectrum Disorder (ASD), a neurodevelopmental disorder, is characterized by significant impairments in social communication and interaction. Theories regarding the cognitive development of ASD individuals suggest that high-functioning individuals are adept at analyzing and constructing systems so they may find STEM occupations more appealing. Researchers found that young adults with ASD in STEM fields were more likely to concentrate be science or computer science majors in college than the general population. In 2001, Baron-Cohen et al. published the seminal paper on the Autism Spectrum Quotient (AQ), a self-administered questionnaire that measures autistic traits in adults. In the paper, they established that there is a significant difference between the AQ scores of high functioning autistic adults and adults of the normal population (35 vs. 16.4). Subsequent studies have confirmed the result. Interestingly, among Cambridge University students that Baron-Cohen et al. recruited, Math and Computer Science majors had the highest AQ scores (21.5 for Math majors, 21.1 for CS majors and 17.6 for all university students). In this thesis, we explored the findings about Computer Science majors further. Part of our motivation stems from research that shows high-functioning autistic young adults tend to pursue STEM majors. Moreover, among those who majored in STEM fields, Computer Science was the top *** recruited UW-Milwaukee students enrolled in Computer Science 250 (CS 250) the Introduction to Programming class, and Computer Science (CS) and Computer Engineering (CE) majors who have applied for graduation in Spring 2020 and Fall 2020. All students were asked to answer the AQ and their AQ scores were compared to their grades. Our results are as follows. The average AQ score among all the participants is 23.04, which was comparable to Baron-Cohen et. al.’s findings. The beginner CS students average AQ score is 22.8 while that of the graduating students is sligh

关键词： Computer science   Artificial intelligence  

关键词： Computer science   Artificial intelligence   Transportation  

摘要： This dissertation creates and examines algorithmic models for automated machine-to-machine negotiation in localized multi-agent systems at the edge of the Internet of Things. It provides an implementation of two such models for unsupervised resource allocation for the application domain of autonomous vehicle traffic as it pertains to lane changing and speed setting selection. The first part concerns negotiation via abstract argumentation. A general model for the arbitration of conflict based on abstract argumentation is outlined and then applied to a scenario where autonomous vehicles on a multi-lane highway use expert systems in consultation with private objectives to form arguments and use them to compete for lane positions. The conflict resolution component of the resulting argumentation framework is augmented with social voting to achieve a community supported conflict-free outcome. The presented model heralds a step toward independent negotiation through automated argumentation in distributed multi-agent systems. Many other cyber-physical environments embody stages for opposing positions that may benefit from this type of tool for collaboration. The second part deals with game-theoretic negotiation through mechanism design. It outlines a mechanism providing resource allocation for a fee and applies it to autonomous vehicle traffic. Vehicular agents apply for speed and lane assignments with sealed bids containing their private feasible action valuations determined within the context of their governing objective. A truth-inducing mechanism implementing an incentive-compatible strategyproof social choice functions achieves a socially optimal outcome. The model can be adapted to many application fields through the definition of a domain-appropriate operation to be used by the allocation function of the mechanism. Both presented prototypes conduct operations at the edge of the Internet of Things. They can be applied to agent networks in just about any domain where t

摘要： Many scholars are interested in the effective use of the interactive whiteboard (IWB) in teaching and learning. Traditional teaching approaches make it difficult for students to comprehend the process of cell division. Using a range of instructional devices, such as the interactive whiteboard, emphasize the dynamic nature of cell division (IWB). The purpose of the study is to explore how grade 10 life sciences teachers use interactive whiteboards when teaching meiosis. Using a cross-case analysis, this study analyzed the teaching and learning processes using IWBs in teaching meiosis at grade 10. A case study approach research aims at investigating the use of interactive whiteboards (IWBs) utilized in three different schools in South Africa so as to obtain first-hand information. Interviews and lesson observations were carried out with four different teachers. The Saldana coding model was used to analyze the responses, and narrative analysis was used. The IWBs have not been successfully used in the teaching of meiosis. Factors such as lack of teacher training on the use of IWBs and insufficient technological support have been cited as some of the reasons IWBs are not effective in teaching and learning. Recommendations such as, teacher training and setting aside a proper budget for IWB implementation were made. For future research, it would be great for researchers to investigate how to better recognize technology leaders in a school, including how personality, expertise, qualifications, and preferences all play a role in identifying a successful technology leader.

关键词： Computer science   Elementary education   Expansive framing   In-service teacher learning   Teacher professional development  

摘要： Elementary school teachers are being increasingly asked to teach computer science—something that most teacher certification programs do not prepare them for. In an attempt to study how elementary teachers learn to teach computer science, I analyzed the ways that teachers behaved during a professional development accompanying the implementation of a fifth-grade computer science curriculum. My findings suggest that teachers benefit from professional development that encourages collaboration and active participation in teachers through discussion and modeling. Furthermore, my findings suggest that teachers benefit from using curriculum that deliberately connects new concepts to content that they are already familiar and comfortable with—a model known as expansive framing. By encouraging active teacher participation in professional development and by using curriculum that relates to teachers’ existing content knowledge, we may be able to help elementary teachers prepare to teach computer science with more confidence and accuracy.

关键词： Predictive models   Correlation   Urban areas   Neural networks   Trajectory   Task analysis   Computer science   Deep learning   urban crowd flows   neural networks   long short-term memory   convolutional neural network  

摘要： Prediction of traffic crowd movement is one of the most important component in many applications ' domains ranging from urban management to transportation schedule. The key challenge of citywide crowd flows prediction is how to model spatial and dynamic temporal correlation. However, in recent years several studies have been done, but they lack the ability to effectively and simultaneously model spatial and temporal dependencies among traffic crowd flows. To address this issue, in this article a novel spatio-temporal deep hybrid neural network proposed termed STD-Net to forecast citywide crowd traffic flows. More specifically, STD-Net contains four major branches, i.e., closeness, period volume, weekly volume, and external branches, respectively. We design a residual neural network unit for each property to depict the spatio-temporal features of traffic flows. For various branches, STD-Net provides distinct weights and then combines the outputs of four branches together. Extensive experiments on two large-scale datasets from New York bike and Beijing taxi have demonstrated that STD-Net achieves competitive performances the existing state-of-the-art prediction baselines.

关键词： Bioinformatics   Biostatistics   Computer science   Nanoscience