检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Luo, Xinyu Musco, Christopher Widdershoven, Cas Department of Computer Science Purdue University Indiana United States Tandon School of Engineering New York University New York United States State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Beijing China

Finding the mode of a high dimensional probability distribution D is a fundamental algorithmic problem in statistics and data analysis. There has been particular interest in efficient methods for solving the problem when D is represented as a mixture model or kernel density estimate, although few algorithmic results with worst-case approximation and runtime guarantees are known. In this work, we significantly generalize a result of (Lee et al., 2021) on mode approximation for Gaussian mixture models. We develop randomized dimensionality reduction methods for mixtures involving a broader class of kernels, including the popular logistic, sigmoid, and generalized Gaussian kernels. As in Lee et al.’s work, our dimensionality reduction results yield quasi-polynomial algorithms for mode finding with multiplicative accuracy (1 − ϵ) for any ϵ > 0. Moreover, when combined with gradient descent, they yield efficient practical heuristics for the problem. In addition to our positive results, we prove a hardness result for box kernels, showing that there is no polynomial time algorithm for finding the mode of a kernel density estimate, unless P = NP. Obtaining similar hardness results for kernels used in practice (like Gaussian or logistic kernels) is an interesting future direction. Copyright © 2023, The Authors. All rights reserved.

关键词： Gradient methods

来源：评论

学校读者我要写书评

暂无评论

Adaptive Student Inference Network for Efficient Single Image Super-Resolution

Adaptive Student Inference Network for Efficient Single Imag...

引用

IEEE International Conference on Data Mining (ICDM)

作者： Kang Miao Zhao Zhang Jiahuan Ren Mingbo Zhao Haijun Zhang Richang Hong School of Computer Science and Information Engineering Hefei University of Technology Hefei China Key Laboratory of Knowledge Engineering with Big Data (Ministry of Education) & Intelligent Interconnected Systems Laboratory of Anhui Province Hefei University of Technology Hefei China Yunnan Key Laboratory of Software Engineering Kunming China School of Information Science and Technology Donghua University Shanghai China Department of Computer Science Harbin Institute of Technology (Shenzhen) Shenzhen China

Recent advances in single image super-resolution (SISR) have achieved remarkable performance through deep learning. However, the high computational cost hinders the deployment of SISR models on edge devices. Instead of proposing new SISR models, a new trend is emerging to improve network efficiency by reducing parameters, FLOPs, and inference time through slight modifications to the original models. However, recent methods usually focus on reducing only one of three metrics, i.e., FLOPs, parameters and inference time, which inevitably increases the other two metrics. In this paper, we propose a novel Adaptive Student Inference Network (ASIN) on popular SISR models, which aims at reducing FLOPs and inference time while maintaining the number of parameters and restoring clearer high-resolution images. Specifically, our ASIN divides a SISR model into three components (head, body and tail) and adopts various strategies for each part. For head and tail parts, to ensure the restored images contain more detailed information, a novel auxiliary Enhanced Teacher Network (ETNet) is designed, which is trained with the ground-truth images to obtain more prior knowledge to guide student network to extract more accurate textures using a new knowledge distillation method. For the body part, owing to the varying difficulties of the reconstructions in different regions, we propose an Adaptive Depth Predicted Module (ADPM) to dynamically shorten average depth of network to reduce the computational cost of overall network. Extensive experiments on two datasets demonstrate the effectiveness and state-of-the-art performance of our ASIN compared to its counterparts.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Challenges and Experiences of Iranian Developers with MLOps at Enterprise

Challenges and Experiences of Iranian Developers with MLOps ...

引用

Advances in Enterprise Architecture (NCAEA), National Conference on

作者： Mohammad Heydari Zahra Rezvani School of Industrial and Systems Engineering Tarbiat Modares University Tehran Iran School of Computer Science Institute for Research in Fundamental Sciences (IPM) Tehran Iran

Data is becoming more complex, and so are the approaches designed to process it. Enterprises have access to more data than ever, but many still struggle to glean the full potential of insights from what they have. This research explores the challenges and experiences of Iranian developers in implementing the MLOps paradigm within enterprise settings. MLOps, or Machine Learning Operations, is a discipline focused on automating the continuous delivery of machine learning models. In this study, we review the most popular MLOps tools used by leading technology enterprises. Additionally, we present the results of a questionnaire answered by over 110 Iranian Machine Learning experts and software Developers, shedding light on MLOps tools and the primary obstacles faced. The findings reveal that data quality problems, a lack of resources, and difficulties in model deployment are among the primary challenges faced by practitioners. Collaboration between ML, DevOps, Ops, and science teams is seen as a pivotal challenge in implementing MLOps effectively.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Survey on Modern Deep Neural Network for Traffic Prediction: Trends, Methods and Challenges (Extended Abstract)

A Survey on Modern Deep Neural Network for Traffic Predictio...

引用

International Conference on Data engineering

作者： David Alexander Tedjopurnomo Zhifeng Bao Baihua Zheng Farhana Murtaza Choudhury A. K. Qin School of Computing Technologies RMIT University Australia School of Computing and Information Systems Singapore Management University Singapore School of Computing and Information Systems The University of Melbourne Australia Department of Computer Science and Software Engineering Swinburne University of Technology Australia

In this modern era, traffic congestion has become a major source of negative economic and environmental impact for urban areas worldwide. One of the most efficient ways to mitigate this issue is through traffic prediction. This research field has evolved greatly ever since its inception in the late 70s. Recently, deep neural network models have gained popularity thanks to its predictive power, but despite this, literature surveys of such methods are rare; making it difficult to ascertain the progress of this research field. In this work, we address this issue by presenting an up-to-date survey of deep neural network for traffic prediction. We provide detailed explanations of popular deep neural network architectures used in the traffic flow prediction literatures, categorize and describe the literatures themselves, present an overview of the commonalities and differences among different works, and finally provide a discussion regarding the challenges and future directions for this field.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Optimizing Diabetes Prediction with Ensemble Learning with Voting and Cross-Validation: A Comprehensive Approach

Optimizing Diabetes Prediction with Ensemble Learning with V...

引用

2023 International Conference on Sustainable Communication Networks and Application, ICSCNA 2023

作者： Shaik, Moinuddin Ahmed Sundari, M. V. Rama Yadla, Jyothi Priyadarshini, V. Narasimha, V. Gadiyar, H. Manoj T. Maulana Azad National Urdu University Department of Csit Hyderabad India Gokaraju Rangaraju Institute of Engineering and Technology Department of Aimle Hyderabad India Sir C R Reddy College of Engineering Department of Cse Eluru India S R K R Engineering College Department of Cse Bhimavaram India Cmr College of Engineering and Technology Department of Cse Hyderabad India Canara Engineering College Bantwal Vtu Belagavi Department of Computer Science and Business Systems Mangalore India

ISBN: (纸本)9798350313987

Diabetes, an enduring metabolic illness distinguished by heightened amounts of glucose in the bloodstream, poses a substantial health obstacle on a global scale. The present study aims to address the pressing requirement for precise predictive instruments in the identification and control of diabetes. The present study investigated two basic classifications of diabetes, namely Type 1 and Type 2, which are characterized by unique etiological variables. The impetus behind this research stems from the increasing prevalence of diabetes and its significant implications for public health and healthcare expenditures. The prompt detection and timely intervention of medical conditions have the potential to enhance patient outcomes and alleviate the burdens placed on society. The stated research need pertains to the necessity for the development of robust and scalable machine learning algorithms specifically designed for the purpose of predicting diabetes. The present study utilizes a robust ensemble learning methodology by integrating Random Forest, Support Vector Machine, and XGBoost classifiers within a Voting Ensemble Classifier. The models are assessed using both 10-fold cross-validation (CV-10) and without it. During the evaluation process, important metrics such as accuracy, precision, recall, F1-score, and AUC are examined. The findings suggest that the inclusion of CV-10 improves the performance of the model, resulting in increased accuracy, precision, and a more favorable tradeoff between precision and recall. The proposed methodology seeks to incorporate a broader spectrum of diabetes problems, such as hypoglycaemia and diabetic retinopathy. The major aim of this work is to provide an academic contribution to the field of healthcare analytics through the presentation of a comprehensive technique for the prediction and management of diabetes. © 2023 IEEE.

关键词： Cross-validation Random Forest Support Vector Machine Voting Ensemble Classifier XGBoost classifiers

来源：评论

学校读者我要写书评

暂无评论

Comparing the Effect of Code Optimizations on Simulation Runtime Across Synchronous Cellular Automata Models of HIV

Comparing the Effect of Code Optimizations on Simulation Run...

引用

2021 Winter Simulation Conference, WSC 2021

作者： Li, Junjiang Giabbanelli, Philippe J. Koster, Till Miami University Department of Computer Science Software Engineering OH45056 United States Institute for Visual and Analytic Computing University of Rostock Rostock18059 Germany

ISBN: (纸本)9781665433112

Models developed by domain experts occasionally struggle to achieve a sufficient execution speed. Improving performances requires expertise in parallel and distributed simulations, hardware, or time to profile performances to identify bottlenecks. However, end-users in biological simulations of the Human Immunodeficiency Virus (HIV) have repeatedly demonstrated that these resources are either not available or not sought, resulting in models that are developed through user-friendly languages and platforms, then used on workstations. This situation becomes problematic when performances cannot cope with the salient characteristics of the phenomenon that is modeled, as is the case with cellular automata (CA) models of HIV. In this paper, we optimize the Python code of CA models of HIV to scale the number of cells handled by a simulation on a workstation commonly available to end-users. We demonstrate this scalability via five HIV CA models and compare these results to assess how modeling choices can impact runtime. © 2021 IEEE.

关键词： Viruses

来源：评论

学校读者我要写书评

暂无评论

Was Wuhan the early epicenter of the COVID-19 pandemic?——a critique

引用

National science Review 2023年第4期10卷 19-21页

作者： Yanan Cao Lingling Chen Hua Chen Yupeng Cun Xiaofeng Dai Hongli Du Feng Gao Fengbiao Guo Yalong Guo Pei Hao Shunmin He Shunping He XiongLei He Zheng Hu Boon-Peng Hoh Xin Jin Qian Jiang Qinghua Jiang Asifullah Khan Hong-Zhi Kong Jinchen Li Shuai Cheng Li Ying Li Qiang Lin Jianquan Liu Qi Liu Jian Lu Xuemei Lu Shujin Luo Qinghua Nie Zilong Qiu Tieliu Shi Xiaofeng Song Jianzhong Su Sheng-ce Tao Chaolong Wang Chuan-Chao Wang Guo-Dong Wang Jiguang Wang Qi Wu Shaoyuan Wu Shuhua Xu Yu Xue Wenjun Yang Zhaohui Yang Kai Ye Yuan-Nong Ye Li Yu Fangqing Zhao Yiqiang Zhao Weiwei Zhai Dandan Zhang Liye Zhang Houfeng Zheng Qi Zhou Tianqi Zhu Ya-ping Zhang Ruijin Hospital Shanghai Jiao Tong University College of Life Science and Technology Guangxi University Beijing Institute of Genomics Chinese Academy of Sciences Children's Hospital of Chongqing Medical University Wuxi School of Medicine Jiangnan University School of Biology and Biological Engineering South China University of Technology Department of Physics School of ScienceTianjin University School of Pharmaceutical Sciences Wuhan University Institute of Botany Chinese Academy of Sciences Institut Pasteur of Shanghai Chinese Academy of Sciences Institute of Biophysics Chinese Academy of Sciences Institute of Hydrobiology Chinese Academy of Sciences School of Life Sciences Sun Yat-sen University Institute of Synthetic Biology Shenzhen Institute of Advanced TechnologyChinese Academy of Sciences Faculty of Medicine and Health Sciences University College Sedaya International School of Medicine South China University of Technology Department of Medical Genetics Capital Institute of Pediatrics School of Life Science and Technology Harbin Institute of Technology Department of Biochemistry Abdul Wali Khan University Xiangya Hospital Central South University Department of Computer Science City University of Hong Kong College of Life Science and Technology Foshan University South China Sea Institute of Oceanology Chinese Academy of Sciences College of Ecology Lanzhou University School of Life Sciences and Technology Tongji University School of Life Sciences Peking University Kunming Institute of Zoology Chinese Academy of Sciences College of Animal Science South China Agricultural University Institute of Neuroscience Chinese Academy of Sciences School of Life Sciences East China Normal University Nanjing University of Aeronautics and Astronautics Wenzhou Institute University of Chinese Academy of Sciences Institute of Systems Biomedicine Shanghai Jiao Tong University Tongji Medical College Huazhong University of Science and Technology School of Life Sciences Xiamen University Division of Life Science and Department of

Recently, Worobey et al.(2022) published a report entitled ‘The Huanan Seafood Wholesale Market in Wuhan was the early epicenter of the COVID-19 pandemic' that succinctly summarizes their study [1]. A pre-print version of this study had earlier elicited a series of highprofile media coverages [2,3]. All these reports deliver a social-political message that the Huanan market is the epicenter of COVID-19.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Generalizable Decision Boundaries: Dualistic Meta-Learning for Open Set Domain Generalization

Generalizable Decision Boundaries: Dualistic Meta-Learning f...

引用

International Conference on computer Vision (ICCV)

作者： Xiran Wang Jian Zhang Lei Qi Yinghuan Shi Nanjing University State Key Laboratory for Novel Software Technology and National Institute of Healthcare Data Science Nanjing University China Southeast University School of Computer Science and Engineering Southeast University China

Domain generalization (DG) is proposed to deal with the issue of domain shift, which occurs when statistical differences exist between source and target domains. However, most current methods do not account for a common realistic scenario where the source and target domains have different classes. To overcome this deficiency, open set domain generalization (OSDG) then emerges as a more practical setting to recognize unseen classes in unseen domains. An intuitive approach is to use multiple one-vs-all classifiers to define decision boundaries for each class and reject the outliers as unknown. However, the significant class imbalance between positive and negative samples often causes the boundaries biased towards positive ones, resulting in misclassification for known samples in the unseen target domain. In this paper, we propose a novel meta-learning-based framework called dualistic MEta-learning with joint DomaIn-Class matching (MEDIC), which considers gradient matching towards inter-domain and inter-class splits simultaneously to find a generalizable boundary balanced for all tasks. Experimental results demonstrate that MEDIC not only outperforms previous methods in open set scenarios, but also maintains competitive close set generalization ability at the same time. Our code is available at https://***/zzwdx/MEDIC.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Simulation Environment for UAV Offshore Wind-Turbine Inspection

Simulation Environment for UAV Offshore Wind-Turbine Inspect...

引用

OCEANS

作者： A. Oliveira A. Dias T. Santos P. Rodrigues A. Martins E. Silva J. Almeida INESC TEC Institute for Systems and Computer Engineering Technology and Science Porto Portugal ISEP Instituto Superior de Engenharia do Porto Porto Portugal

Offshore wind farms are becoming the main alternative to fossil fuels and the future key to mitigating climate change by achieving energy sustainability. With favorable indicators in almost every environmental index, these structures operate under varying and dynamic environmental conditions, leading to efficiency losses and sudden failures. For these reasons, it’s fundamental to promote the development of autonomous solutions to monitor the health condition of the construction parts, preventing structural damage and accidents. This paper introduces a new simulation environment for testing and training autonomous inspection techniques under a more realistic offshore wind farm scenario. Combining the Gazebo simulator with ROS, this framework can include multi-robots with different sensors to operate in a customizable simulation environment regarding some external elements (fog, wind, buoyancy…). The paper also presents a use case composed of a 3D LiDAR-based technique for autonomous wind turbine inspection with UAV, including point cloud clustering, model estimation, and the preliminary results under this simulation framework using a mixed environment (offshore simulation with a real UAV platform).

关键词：

来源：评论

学校读者我要写书评

暂无评论

Enhancing EEG Artifact Removal Efficiency by Introducing Dense Skip Connections to IC-U-Net

Enhancing EEG Artifact Removal Efficiency by Introducing Den...

引用

Annual International Conference of the IEEE engineering in Medicine and Biology Society (EMBC)

作者： Kong-Yi Chang Yung-Chia Huang Chun-Hsiang Chuang Institute of Information Systems and Applications College of Electrical Engineering and Computer Science NTHU Taiwan Research Center for Education and Mind Sciences College of Education NTHU Taiwan Department of Engineering and System Science NTHU Taiwan

Electroencephalographic (EEG) data is considered contaminated with various types of artifacts. Deep learning has been successfully applied to developing EEG artifact removal techniques to increase the signal-to-noise ratio (SNR) and enhance brain-computer interface performance. Recently, our research team has proposed an end-to-end UNet-based EEG artifact removal technique, IC-U-Net, which can reconstruct signals against various artifacts. However, this model suffers from being prone to overfitting with a limited training dataset size and demanding a high computational cost. To address these issues, this study attempted to leverage the architecture of UNet++ to improve the practicability of IC-U-Net by introducing dense skip connections in the encoder-decoder architecture. Results showed that this proposed model obtained superior SNR to the original model with half the number of parameters. Also, this proposed model achieved comparable convergency using a quarter of the training data size.

关键词：

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：