检索结果-内蒙古大学图书馆

arXiv 2021年

作者： Kong, Junsheng Li, Weizhao Liu, Zeyi Liao, Ben Qiu, Jiezhong Hsieh, Chang-Yu Cai, Yi Zhang, Shengyu School of Software Engineering South China University of Technology China University of Cambridge United Kingdom Tencent Quantum Lab China Tsinghua University China The Key Laboratory of Big Data and Intelligent Robot South China University of Technology Ministry of Education

The notion of word embedding plays a fundamental role in natural language processing (NLP). However, pre-training word embedding for very large-scale vocabulary is computationally challenging for most existing methods. In this work, we show that with merely a small fraction of contexts (Q-contexts) which are typical in the whole corpus (and their mutual information with words), one can construct high-quality word embedding with negligible errors. Mutual information between contexts and words can be encoded canonically as a sampling state, thus, Q-contexts can be fast constructed. Furthermore, we present an efficient and effective WEQ method, which is capable of extracting word embedding directly from these typical contexts. In practical scenarios, our algorithm runs 11∼ 13 times faster than well-established methods. By comparing with well-known methods such as matrix factorization, word2vec, GloVe and fasttext, we demonstrate that our method achieves comparable performance on a variety of downstream NLP tasks, and in the meanwhile maintains run-time and resource advantages over all these baselines. © 2021, CC BY.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

PhotoRedshift-MML: a multimodal machine learning method for estimating photometric redshifts of quasars

arXiv

引用

arXiv 2022年

作者： Hong, Shuxin Zou, Zhiqiang Luo, A-Li Kong, Xiao Yang, Wenyu Chen, Yanli College of Computer Nanjing University of Posts and Telecommunications Nanjing210023 China Jiangsu Key Laboratory of Big Data Security and Intelligent Processing Nanjing210023 China CAS Key Laboratory of Optical Astronomy National Astronomical Observatories Beijing100101 China School of Astronomy and Space Science University of Chinese Academy of Sciences Beijing100049 China

We propose a Multimodal Machine Learning method for estimating the Photometric Redshifts of quasars (PhotoRedshift-MML for short), which has long been the subject of many investigations. Our method includes two main models, i.e. the feature transformation model by multimodal representation learning, and the photometric redshift estimation model by multimodal transfer learning. The prediction accuracy of the photometric redshift was significantly improved owing to the large amount of information offered by the generated spectral features learned from photometric data via the MML. A total of 415,930 quasars from Sloan Digital Sky Survey (SDSS) data Release 17, with redshifts between 1 and 5, were screened for our experiments. We used |Δz| = |(zphot− zspec)/(1 + zspec)| to evaluate the redshift prediction and demonstrated a 4.04% increase in accuracy. With the help of the generated spectral features, the proportion of data with |Δz| Copyright © 2022, The Authors. All rights reserved.

关键词： Surveys

来源：评论

学校读者我要写书评

暂无评论

Origin of minimal conductivity in Dirac materials: Momentum-dependent self-energy function from long-ranged disorder scattering

引用

Physical Review B 2023年第17期108卷 174205-174205页

作者： Weiwei Chen Yedi Shen Bo Fu Qinwei Shi W. Zhu Key Laboratory of Intelligent Manufacturing Quality Big Data Tracing and Analysis of Zhejiang Province College of Science China Jiliang University Hangzhou 310018 China Hefei National Laboratory for Physical Sciences at the Microscale and Synergetic Innovation Center of Quantum Information and Quantum Physics University of Science and Technology of China Hefei Anhui 230026 China School of Sciences Great Bay University Dongguan 523000 Guangdong Province China Department of Physics The University of Hong Kong Pokfulam Road Hong Kong China School of Science Westlake University Hangzhou 310024 China

We present a unified understanding of the experimentally observed minimal dc conductivity in Dirac materials. First, based on the linear response theory, we reveal that the momentum-dependent self-energy function induces a tunable minimal conductivity in Dirac electrons, the magnitude of which is directly related to the coefficient of the momentum-dependent part in the self-energy function. Taking the long-ranged Gaussian and Coulomb potentials as examples, the momentum-dependent self-energy function is perturbatively derived using the Born approximation and supplemented by the self-consistent Born approximation and renormalization group analysis. Moreover, we further validate our theory via numerical simulations using the large-scale Lanczos algorithm. The explicit momentum dependence of the self-energy on the intensity, concentration, and range of potential are critically addressed. Therefore, our theory provides a reasonable interpretation of the sample-dependent minimal conductivity observed in experiments.

关键词： Dirac fermions Quantum transport Disordered systems Diagrammatic methods

来源：评论

学校读者我要写书评

暂无评论

Suppressing Biased Samples for Robust VQA

引用

IEEE TRANSACTIONS ON MULTIMEDIA 2021年 24卷 3405-3415页

作者： Ouyang, Ninglin Huang, Qingbao Li, Pijian Cai, Yi Liu, Bin Leung, Ho-fung Li, Qing Guangxi Univ Sch Elect Engn Nanning 530005 Peoples R China Guangxi Univ Sch Elect Engn Inst Artificial Intelligence Guangxi Key Lab Multimedia Commun & Network Techn Nanning 530004 Peoples R China South China Univ Technol Sch Software Engn Guangzhou 510006 Peoples R China MOE China Key Lab Big Data & Intelligent Robot SCUT Guangzhou 510006 Peoples R China Chinese Univ Hong Kong Dept Comp Sci & Engn Hong Kong 999077 Peoples R China Hong Kong Polytech Univ Dept Comp Hong Kong 999077 Peoples R China

Most existing visual question answering (VQA) models strongly rely on language bias to answer questions, i.e., they always tend to fit question-answer pairs on the train split and perform poorly on the test spilt when the answer distributions are different. This behavior makes them hard to be applied in real scenarios. To reduce the language biases, previous studies mainly integrate modules to overcome language priors (ensemble-based methods) or generate additional training data to balance dataset biases (data-balanced methods). However, all the existing ensemble-based methods drop their accuracies on the VQA v2 dataset, while data-balanced methods may introduce new biases and cannot guarantee the quality of the generated data. In this paper, we propose a model-agnostic training scheme called Suppressing Biased Samples (SBS) to overcome language priors. SBS consists of two collaborative parts, i.e., a data Classifier Module to divide the dataset into biased samples and unbiased samples by utilizing the similarity in the semantic space, and a Bias Penalty Module to suppress the biased samples to weaken their influence. As a new way of balancing data to address language bias, SBS overcomes the shortcomings of previous data-balanced methods. Experimental results show that our method can be merged into other bias-reduction methods and achieves a new state-of-the-art performance on the commonly used VQA-CP v2 dataset.

关键词： Training Visualization Training data Image color analysis Sports Knowledge discovery Annotations Visual Question Answering Language Priors Language Bias Robust VQA

来源：评论

学校读者我要写书评

暂无评论

Short-term Traffic Flow Prediction by Graph Deep Learning with Spatial Temporal Modeling

Short-term Traffic Flow Prediction by Graph Deep Learning wi...

引用

Electronic Information Technology and Smart Agriculture (ICEITSA), International Conference on

作者： Lilian Yuan Weilin Fang Hongbo Xiao Jianhua Xiao Yuanquan Shi Yujun Yang School of Computer Science and Engineering Huaihua University Huaihua Hunan P. R. China Key Laboratory of Wuling-Mountain Health Big Data Intelligent Processing and Application in Hunan Province Universities Huaihua Hunan P. R. China Key Laboratory of Intelligent Control Technology for Wuling-Mountain Ecological Agriculture in Hunan Province Huaihua Hunan P. R. China

ISBN: (纸本)9781665464024

An extensive amount of research applications on traffic flow prediction models taking into account spatio-temporal modeling have been undertaken in recent years. Relative to traditional spatio-temporal modeling approaches, good application results have been achieved in predicting traffic flow. However, in real life, traffic flow proves to be time- and space-dependent, which poses difficulties for accurate modeling of traffic flow prediction. To address these problems, we propose the ST-GNN-GRU model, which mainly contains S-GNN and T-GRU modules. The model uses a combination of spatio-temporal modeling and deep learning, which can not only cope with the spatio-temporal modeling problem, but also improve the traffic flow prediction accuracy. The experiments are conducted using the METR-LA dataset, and the outcomes show that the proposed method performs more effectively than other baseline approaches.

关键词： Deep learning Smart agriculture Predictive models Prediction algorithms Information technology Testing

来源：评论

学校读者我要写书评

暂无评论

TomoSAR-ALISTA: Efficient TomoSAR imaging via deep unfolded network

TomoSAR-ALISTA: Efficient TomoSAR imaging via deep unfolded ...

引用

International Conference on Radar Systems (RADAR 2022)

作者： M. Wang Z. Zhang Y. Wang S. Gao X. Qiu Key Laboratory of Technology in Geo-spatial Information Processing and Application System Chinese Academy of Sciences Beijing 100190 People's Republic of China Key Laboratory of Intelligent Aerospace Big Data Application Technology Suzhou 215123 People's Republic of China Electrical and Computer Engineering Department George Mason University Fairfax VA 22030 USA

ISBN: (数字)9781839537776

Synthetic aperture radar (SAR) tomography (TomoSAR) has attracted remarkable interest for its ability in achieving threedimensional reconstruction along the elevation direction from multiple observations. In recent years, compressed sensing (CS) technique has been introduced into TomoSAR considering for its super-resolution ability with limited samples. Whereas, the CS-based methods suffer from several drawbacks, including weak noise resistance, high computational complexity and complex parameter fine-tuning. Among the different CS algorithms, iterative soft-thresholding algorithm (ISTA) is widely used as a robust reconstruction approach, however, the parameters in the ISTA are manually chosen, which usually requires a time-consuming fine-tuning process to achieve the best performance. Aiming at efficient TomoSAR imaging, a novel sparse unfolding network named analytic learned ISTA (ALISTA) is proposed towards the TomoSAR imaging problem in this paper, and the key parameters of ISTA are learned from training data via deep learning to avoid complex parameter fine-tuning and significantly relieves the training burden. In addition, experiments verify that it is feasible to use traditional CS algorithms as training labels, which provides a tangible supervised training method to achieve better 3D reconstruction performance even in the absence of labeled data in real applications.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Design and Optimization of Evaluation Metrics in Object Detection and Tracking for Low-Altitude Aerial Video 1st

Design and Optimization of Evaluation Metrics in Object Dete...

引用

1st International Conference on big data and Security, ICBDS 2019

作者： Wang, Li Shu, Xin Zhang, Wei Chen, Yunfang Nanjing University of Posts and Telecommunications NanjingJiangsu210023 China Jiangsu Key Laboratory of Big Data Security and Intelligent Processing Nanjing University of Posts and Telecommunications NanjingJiangsu210023 China

ISBN: (纸本)9789811575297

The combination of Unmanned Aerial Vehicle (UAV) technology and computer vision has become popular in a wide range of applications, such as surveillance and reconnaissance, while popular evaluation measures are sometimes not applicable for specific tasks. In order to evaluate visual object detection and tracking algorithms of low-altitude aerial video properly, we first summarize the evaluation basis of computer vision tasks, including ground truth, prediction-to-ground truth assignment strategy and distance measures between prediction and ground truth. Then, we analyze the advantages and disadvantages of visual object detection and tracking performance measures, including average precision (AP), F-measure, and accuracy. Finally, for the low-altitude (nearly 100 m) surveillance mission of small unmanned aerial vehicles, we discuss the threshold optimization method of popular measures and the design strategy of application measures. Our work provides a reference in the aspect of performance measures design for researchers of UAV vision. © 2020, Springer Nature Singapore Pte Ltd.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

A Recommendation for Flight Extra Fuel Based on Random Forest 2020

A Recommendation for Flight Extra Fuel Based on Random Fores...

引用

Proceedings of the 2020 9th International Conference on Computing and Pattern Recognition

作者： Yiming Wang Linkai Luo Qifeng Zhou Junsong Lin Zixuan Wu Department of Automation Xiamen University Xiamen P.R. China Xiamen Key Laboratory of Big Data Intelligent Analysis and Decision Xiamen P.R. China Department of Automation Xiamen University Xiamen P.R. China Xiamen Key Laboratory of Big Data Intelligent Analysis and Decision Xiamen P.R. China Department of Automation Xiamen University Xiamen P.R. China Xiamen Key Laboratory of Big Data Intelligent Analysis and Decision Xiamen P.R. China Xiamen Airlines Xiamen P.R.China

ISBN: (纸本)9781450387835

It is critical for airlines to develop an effective fuel-saving strategy because the fuel cost accounts for a large proportion of the operating costs in air transport industry. This paper aims to optimize the flight extra fuel predicted by dispatchers via data mining techniques on the consideration of both economy and safety. We first establish a random forest (RF) model from history data to predict the consumption of trip fuel. Then to ensure safety index, we add a margin to the output of the model as the modified prediction value. Finally, the experimental results on real flight data show that the proposed method can make a more accurate prediction for the extra fuel adding amount under the condition of ensuring safety. It indicates that the proposed method is beneficial for airlines in economic cost saving.

关键词： Fuel consumption prediction extra fuel optimization random forest

来源：评论

学校读者我要写书评

暂无评论

Highly sensitive ethylene glycol gas sensor based on SnO 2 /ZnO heterojunction composites

引用

Vacuum 2025年 240卷

作者： JunTang Dong Te Duan TingTing Shao NingJuan Li JiaKe Guo Fuchun Zhang School of Physics and Electronic Information Yan'an University Yan'an 716000 China Shaanxi Key Laboratory of Intelligent Processing for Big Energy Data Yan'an University Yan'an 716000 China

Ethylene glycol (EG), widely used in industry, poses severe health and environmental risks even at low concentrations, necessitating the development of highly sensitive gas sensors. Although ZnO and SnO 2 are promising semiconductor metal oxides for gas detection, their single-component limitations hinder optimal performance. This work focuses on SnO 2 -decorated ZnO nanosheets, where a heterojunction structure is constructed to overcome individual material drawbacks. The aim is to enhance EG detection performance through compositional tuning and interface engineering. SnO 2 /ZnO composites with varying SnO 2 contents were synthesized via a one-step hydrothermal method and systematically evaluated. The SnO 2 /ZnO-1 sample (1:1 M ratio) exhibited the best performance, achieving a response of 169.89 to 100 ppm EG at 240 °C, a detection limit as low as 158 ppb, high linearity (R 2 = 0.99593), and strong repeatability and selectivity. Structural characterizations confirmed the successful formation of an n–n heterojunction and increased surface area and mesoporosity, enhancing gas adsorption and electron transfer. XPS revealed a higher proportion of adsorbed oxygen (30.21 %) in the composite versus pure ZnO (27.28 %), contributing to better reactivity. This research demonstrates that tailoring heterostructure interfaces and material morphology significantly improves gas sensing performance, offering a promising strategy for reliable and real-time EG detection in environmental and safety monitoring.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Community Detection Based on DeepWalk in Large Scale Networks 1st

Community Detection Based on DeepWalk in Large Scale Network...

引用

1st International Conference on big data and Security, ICBDS 2019

作者： Chen, Yunfang Wang, Li Qi, Dehao Zhang, Wei Nanjing University of Posts and Telecommunications NanjingJiangsu210023 China Jiangsu Key Laboratory of Big Data Security and Intelligent Processing Nanjing University of Posts and Telecommunications NanjingJiangsu210023 China

ISBN: (纸本)9789811575297

The large scale and complex structure of real networks bring enormous challenges to traditional community detection methods. In order to detect community structure in large scale networks more accurately and efficiently, we propose a community detection algorithm based on network embedding representation method. This algorithm first selects the DeepWalk network embedding method to represent the vertices in the network graph as low-dimensional vector representations which reflect the topology in the network. Then this algorithm treats each vertex as a sample, and each dimension of the vector representation as a feature. Finally, the sample data are input into the variational Gaussian mixture clustering model to obtain the clusters in the network. Experimental results on the DBLP dataset show that the model method of this paper can more effectively discover the communities in large scale networks. Through further analysis of the excavated community structure, the organizational characteristics within the community are better revealed. © 2020, Springer Nature Singapore Pte Ltd.

关键词： Network embeddings

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：