检索结果-内蒙古大学图书馆

IEEE Symposium on Nuclear science (NSS/MIC)

作者： Wenbo Huang Greta S. P. Mok Department of Electrical and Computer Engineering Faculty of Science and Technology and the Center for Cognitive and Brain Sciences Institute of Collaborative Innovation University of Macau Taipa China

Parkinson’s disease (PD) can be early diagnosed using radionuclide imaging. Digital phantoms and Monte Carlo (MC) simulations serve as important research tools for PD imaging physics research. Here we designed and developed a new digital brain phantom population for different stages of PD. We selected 206 T1 MRI brain images and 160 DaTscan SPECT reconstructed images for normal subjects and patients with different PD stages from Parkinson’s Progression Markers Initiative (PPMI) datasets. We used the model-based segmentation tool FIRST to segment putamen and caudate region and used Brain Extraction Tool (BET) to extract brain region from the MRI images, generating striatum and brain maps, respectively. Striatum maps were then rigidly registered to a referenced striatum model using Elastix, and the corresponding brain maps were transformed using the same parameters. Different activities were set on the maps based on the striatal binding ratio (SBR) in PPMI and the SPECT Visual interpretation assessment scheme. New phantoms were generated by randomly pairing the striatum and brain maps. Corresponding CT maps were generated by non-rigidly registered the XCAT-attenuation map to MRI images corresponding to the brain maps and transformed using the same parameters. We then used the Simind MC code to generate 120 realistic noisy projections with full physics modelling for each new brain phantom, based on a clinical SPECT scanner mounted with low-energy high-resolution parallel-hole collimators. OS-EM algorithm (8 iterations, 4 subsets) was used to reconstruct projections with and without attenuation correction (AC) respectively. SBR was evaluated on reconstructed images. Our results show that we could generate realistic PD SPECT data with the newly developed digital brain phantom population combined with a MC code, and AC is important for PD SPECT based on both visual assessment and SBR analysis. Our phantoms could be applied for further PD radionuclide imaging study.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Image Captioning with Reinforcement Learning 2

Image Captioning with Reinforcement Learning

引用

2nd IEEE International Conference on computer Vision and Machine Intelligence, CVMI 2023

作者： Verma, Anand Agarwal, Saurabh Arya, K.V. Petrlik, Ivan Esparza, Roberto Rodriguez, Ciro ABV-IIITM Gwalior Department of Information Technology Madhya Pradesh 474015 India ABV-Indian Institute of Information Technology and Management Multimedia and Information Security Research Group Department of Computer Science and Engineering Gwalior474015 India National University Federico Villarreal Faculty of Industrial and Systems Engineering Lima Peru National University Mayor de San Marcos Faculty of Systems Engineering and Informatics Peru

ISBN: (纸本)9798350305142

Image captioning involves generating a natural language description that accurately represents the content and context of an image. To achieve this, image captioning utilises various machine learning techniques and fields, such as computer vision and natural language processing. In the field of image captioning, a lot of advances have been made with encoder-decoder models and reinforcement learning algorithms. However, there are still problems of imbalance between testing and training, as reinforcement learning only handles single comparator metrics such as CIDEr, SPICE, and BLEU and could not perform better in multiple metrics at once. Which is why a lack of diversity can be seen in generated captions. This idea proposes a general technique for collaborative updating that can bridge the gap between evaluation measures and test metrics to produce captions that are more human-like. To increase the precision of image captions, the approach involves using a compiled reward system that considers multiple evaluation metrics to compare the generated sentence with the provided sentences. We will evaluate the model's performance and the reward updating process on standard datasets like MS COCO. © 2023 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Epidemic spreading and digital contact tracing: Effects of heterogeneous mixing and quarantine failures

引用

Physical Review E 2022年第4期105卷 044313-044313页

作者： Abbas K. Rizi Ali Faqeeh Arash Badie-Modiri Mikko Kivelä Department of Computer Science School of Science Aalto University FI-00076 Finland Mathematics Applications Consortium for Science & Industry University of Limerick Limerick V94 T9PX Ireland Luddy School of Informatics Computing and Engineering Indiana University Bloomington Indiana 47408 USA

Contact tracing via digital tracking applications installed on mobile phones is an important tool for controlling epidemic spreading. Its effectivity can be quantified by modifying the standard methodology for analyzing percolation and connectivity of contact networks. We apply this framework to networks with varying degree distributions, numbers of application users, and probabilities of quarantine failures. Further, we study structured populations with homophily and heterophily and the possibility of degree-targeted application distribution. Our results are based on a combination of explicit simulations and mean-field analysis. They indicate that there can be major differences in the epidemic size and epidemic probabilities which are equivalent in the normal susceptible-infectious-recovered (SIR) processes. Further, degree heterogeneity is seen to be especially important for the epidemic threshold but not as much for the epidemic size. The probability that tracing leads to quarantines is not as important as the application adoption rate. Finally, both strong homophily and especially heterophily with regard to application adoption can be detrimental. Overall, epidemic dynamics are very sensitive to all of the parameter values we tested out, which makes the problem of estimating the effect of digital contact tracing an inherently multidimensional problem.

关键词： Complex systems Continuous phase transition Epidemic Network phase transitions Percolation phase transition Stochastic processes Barabasi-Albert network Networks & random structures Poisson degree distribution networks Scale free & inhomogeneous networks Social networks Critical phenomena Epidemic spreading Percolation theory

来源：评论

学校读者我要写书评

暂无评论

Evaluating student levelling based on machine learning model’s performance

引用

Discover Internet of Things 2022年第1期2卷 3页

作者： Ghareeb, Shatha Hussain, Abir Jaafar Al-Jumeily, Dhiya Khan, Wasiq Al-Jumeily, Rawaa Baker, Thar Al Shammaa, Ahmed Khalaf, Mohammed Faculty of Engineering and Technology Liverpool John Moores University Byrom Street LiverpoolL3 3AF United Kingdom College of Computing and Informatics University of Sharjah Sharjah27272 United Arab Emirates College of Engineering University of Sharjah Sharjah27272 United Arab Emirates Department of Computer Science Al-Maarif University College Ramadi31001 Iraq

In this paper, a novel application of machine learning algorithms is presented for student levelling. In multicultural countries such as UAE, there are various education curriculums where the sector of private schools and quality assurance is supervising various private schools for many nationalities. As there are various education curriculums in United Arab Emirates, specifically Abu Dhabi, to meet expats’ needs, there are different requirements for registration and success. In addition, there are different age groups for starting education in each curriculum. Every curriculum follows different education methods such as assessment techniques, reassessment rules, and exam boards. Currently, students who transfer to other curriculums are not correctly placed to their appropriate year group as a result of the start and end dates of each academic year as well as due to their date of birth, in which students who are either younger or older for that year group can create gaps in their learning and performance. In addition, pupils’ academic journeys are not stored which create a gap for the schools to track their learning process. In this paper, we propose a computational framework applicable in multicultural countries such as United Arab Emirates in which multi-education systems are implemented. Machine Learning are used to provide the appropriate student’ level aiding schools to provide a smooth transition when assigning students to their year groups and provide levelling and differentiation information of pupils for a smooth transition between one education curriculums to another, in which retrieval of their progress is possible. For classification and discriminant analysis of pupils levelling, three machine learning classifiers are utilised including random forest classifier, Artificial Neural Network, and combined classifiers. The simulation results indicated that the proposed machine learning classifiers generated effective performance in terms of accuracy. © The Au

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

Super-Resolution of License Plate Images Using Attention Modules and Sub-Pixel Convolution Layers

arXiv

引用

arXiv 2023年

作者： Nascimento, Valfride Laroca, Rayson de Lambert, Jorge A. Schwartz, William Robson Menotti, David Department of Informatics Federal University of Paraná Curitiba Brazil Regional Superintendence at Bahia Brazilian Federal Police Salvador Brazil Department of Computer Science Federal University of Minas Gerais Belo Horizonte Brazil

Recent years have seen significant developments in the field of License Plate Recognition (LPR) through the integration of deep learning techniques and the increasing availability of training data. Nevertheless, reconstructing license plates (LPs) from low-resolution (LR) surveillance footage remains challenging. To address this issue, we introduce a Single-Image Super-Resolution (SISR) approach that integrates attention and transformer modules to enhance the detection of structural and textural features in LR images. Our approach incorporates sub-pixel convolution layers (also known as PixelShuffle) and a loss function that uses an Optical Character Recognition (OCR) model for feature extraction. We trained the proposed architecture on synthetic images created by applying heavy Gaussian noise to high-resolution LP images from two public datasets, followed by bicubic downsampling. As a result, the generated images have a Structural Similarity Index Measure (SSIM) of less than 0.10. Our results show that our approach for reconstructing these low-resolution synthesized images outperforms existing ones in both quantitative and qualitative measures. Our code is publicly available at https://***/valfride/lpr-rsr-ext/. Copyright © 2023, The Authors. All rights reserved.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

Extended Abstract: An Attention-Guided Multistream Feature Fusion Network for Early Localization of Risky Traffic Agents in Driving Videos

Extended Abstract: An Attention-Guided Multistream Feature F...

引用

IEEE Symposium on Intelligent Vehicle

作者： Muhammad Monjurul Karim Zhaozheng Yin Ruwen Qin Department of Civil & Environmental Engineering University of Washington Seattle WA USA Department of Biomedical Informatics Department of Computer Science AI Institute Stony Brook University Stony Brook NY USA Department of Civil Engineering Stony Brook University Stony Brook NY USA

ISBN: (数字)9798350348811

ISBN: (纸本)9798350348828

Detecting dangerous traffic agents in videos captured by a dashboard camera (dashcam) mounted on vehicles is essential to ensure safe navigation in complex driving environments. Crash-related videos are corner cases in driving-related big data, and pre-crash processes are transient and complex. Besides, risky and non-risky traffic agents can be similar in their appearance. These make the localization of risky traffic agents in driving videos particularly challenging. In addressing the challenges, this paper proposes an attention-guided multistream feature fusion network (AM-Net) to localize dangerous traffic agents from dashcam videos ahead of potential accidents. Two Gated Recurrent Unit (GRU) networks use object bounding box and optical flow features extracted from consecutive video frames to capture spatio-temporal cues for distinguishing risky traffic agents. An attention module, coupled with the GRUs, learns to identify traffic agents that are relevant to a crash. Fusing the two streams of global and object-level features, AM-Net predicts the riskiness scores of traffic agents in the video. This paper also introduces a new benchmark dataset called Risky Object Localization (ROL), which contains spatial, temporal, and categorical annotations of the crash, object, and scene-level attributes. The proposed AM-Net achieves a promising performance of 85.59% AUC on the ROL dataset. Additionally, the AM-Net outperforms the current state-of-the-art for video anomaly detection by 3.5% AUC on the public DoTA dataset. A thorough ablation study further reveals AM-Net’s merits by assessing the contributions of its functional constituents.

关键词： Location awareness Navigation Intelligent vehicles Logic gates Feature extraction Transient analysis Streams

来源：评论

学校读者我要写书评

暂无评论

Structured world belief for reinforcement learning in POMDP

arXiv

引用

arXiv 2021年

作者： Singh, Gautam Peri, Skand Kim, Junghyun Kim, Hyunseok Ahn, Sungjin Department of Computer Science Rutgers University Electronics and Telecommunications Research Institute Rutgers Center for Cognitive Science

Object-centric world models provide structured representation of the scene and can be an important backbone in reinforcement learning and planning. However, existing approaches suffer in partially-observable environments due to the lack of belief states. In this paper, we propose Structured World Belief, a model for learning and inference of object-centric belief states. Inferred by Sequential Monte Carlo (SMC), our belief states provide multiple object-centric scene hypotheses. To synergize the benefits of SMC particles with object representations, we also propose a new object-centric dynamics model that considers the inductive bias of object permanence. This enables tracking of object states even when they are invisible for a long time. To further facilitate object tracking in this regime, we allow our model to attend flexibly to any spatial location in the image which was restricted in previous models. In experiments, we show that object-centric belief provides a more accurate and robust performance for filtering and generation. Furthermore, we show the efficacy of structured world belief in improving the performance of reinforcement learning, planning and supervised reasoning. https://***/view/ structuredworldbelief. Copyright © 2021, The Authors. All rights reserved.

关键词： Monte Carlo methods

来源：评论

学校读者我要写书评

暂无评论

Dynamic Output State Classification for Quantum computers

Dynamic Output State Classification for Quantum Computers

引用

Quantum Software Engineering (Q-SE), International Workshop on

作者： Hector D. Menendez Luciano Bello David Clark Department of Informatics King's College London London United Kingdom Zurich Research Laboratory IBM Quantum Zurich Switzerland Department of Computer Science University College London London United Kingdom

Quantum computers promise a potentially disruptive approach to improving computation in fields such as physics, chemistry, cryptography, optimisation, and machine learning. However, testing quantum computations for faults is currently impractical because of the existence of noise and errors associated with the output. Executing in a quantum system a circuit with only a few valid output states can generate a significant number of implausible states that have zero probability in an ideal computation. Among other sources of noise, readout errors come from the difficulty of discriminating a measurement between 0 and 1 for the different qubits. These issues are affected by readout drift, requiring regular recalibration of the process. In this paper, we provide a novel technique for post-computation analysis of the output probability distributions that permits better discrimination of kerneled data, delaying the need for recalibration. We achieve this by altering the linear discrimination of the final output states by way of a dynamic state selection process that combines Gaussian mixture models with a probability threshold. As an initial assessment of the technique we examine its effect on three to five qubits GHZ states. Our results on almost every one of nine IBM quantum computers show that the number of implausible states is reduced significantly and that the resulting probability distribution is closer to the expected one.

关键词：

来源：评论

学校读者我要写书评

暂无评论

RGB Color Model Aware Computational Color Naming and Its Application to Data Augmentation

RGB Color Model Aware Computational Color Naming and Its App...

引用

2022 IEEE International Conference on Big Data, Big Data 2022

作者： Yan, Zipei Xu, Linchuan Suzuki, Atsushi Wang, Jing Cao, Jiannong Huang, Jun The Hong Kong Polytechnic University Department of Computing Kowloon Hung Hom Hong Kong King's College London Department of Informatics London United Kingdom University of Greenwich Department of Computing and Mathematical Sciences London United Kingdom Anhui University of Technology School of Computer Science and Technology Ma'anshan China

ISBN: (纸本)9781665480451

Computational color naming (CCN) aims to learn a mapping from pixels into semantic color names, e.g., red, green and blue. CCN has wide applications including color vision deficiency assistance and color image retrieval. Existing research on CCN mainly studies pixels collected under laboratory settings or studies images collected from the web. However, laboratory pixels are very limited such that the learned mapping may not generalize well on unseen pixels, and the mapping discovered from images is usually data-specific. In this paper, we aim to learn a universal mapping by studying pixels collected from the web. To this end, we formulate a novel classification problem that incorporates both the pixels and the RGB color model. The RGB color model is beneficial for learning the mapping because it characterizes the production of colors, e.g., the addition of red and green produces yellow. However, the characterization is rather qualitative. To solve this problem, we propose ColorMLP, which is a multilayer perceptron (MLP) embedded with graph attention networks (GATs). Here, the GATs are designed to capture color relations that we construct by referring to the RGB color model. In this way, the parameters of the MLP can be regularized to comply with the RGB model. We conduct comprehensive experiments to demonstrate the superiority of ColorMLP to alternative *** expand the application of CCN, we design a novel data augmentation method named partial color jitter (PCJ), which performs color jitter (CJ) on a subset of pixels belonging to the same color of an image. In this way, PCJ partially changes the color properties of images, thereby significantly increasing images' diversity. We conduct extensive experiments on CIFAR10/100 and ImageNet datasets, showing that PCJ can consistently improve the classification performance. Our data and software can be found at https://https://***/yanzipei/CCN_and_ItsApp. © 2022 IEEE.

关键词： Mapping

来源：评论

学校读者我要写书评

暂无评论

可信联邦学习进化优化算法综述

引用

Engineering 2024年第3期34卷 23-42页

作者： Qiqi Liu Yuping Yan Yaochu Jin Xilu Wang Peter Ligeti Guo Yu Xueming Yan School of Engineering Westlake UniversityHangzhou 310030China Faculty of Informatics Department of Computer AlgebraEötvös Loránd UniversityBudapest 1053Hungary Faculty of Technology Bielefeld UniversityBielefeld 33619Germany Institute of Intelligent Manufacturing Nanjing Tech UniversityNanjing 211816China School of Information Science and Technology Guangdong University of Foreign StudiesGuangzhou 510006China

With the development of edge devices and cloud computing,the question of how to accomplish machine learning and optimization tasks in a privacy-preserving and secure way has attracted increased attention over the past *** a privacy-preserving distributed machine learning method,federated learning(FL)has become popular in the last few ***,the data privacy issue also occurs when solving optimization problems,which has received little attention so *** survey paper is concerned with privacy-preserving optimization,with a focus on privacy-preserving data-driven evolutionary *** aims to provide a roadmap from secure privacy-preserving learning to secure privacy-preserving optimization by summarizing security mechanisms and privacy-preserving approaches that can be employed in machine learning and *** provide a formal definition of security and privacy in learning,followed by a comprehensive review of FL schemes and cryptographic privacy-preserving ***,we present ideas on the emerging area of privacy-preserving optimization,ranging from privacy-preserving distributed optimization to privacy-preserving evolutionary optimization and privacy-preserving Bayesian optimization(BO).We further provide a thorough security analysis of BO and evolutionary optimization methods from the perspective of inferring attacks and active *** the basis of the above,an in-depth discussion is given to analyze what FL and distributed optimization strategies can be used for the design of federated optimization and what additional requirements are needed for achieving these ***,we conclude the survey by outlining open questions and remaining challenges in federated data-driven *** hope this survey can provide insights into the relationship between FL and federated optimization and will promote research interest in secure federated optimization.

关键词： Federated learning Privacy-preservation Security Evolutionary optimization Data-driven optimization Bayesian optimization

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：