检索结果-内蒙古大学图书馆

arXiv 2018年

作者： Wang, Shuai Huang, Zili Qian, Yanmin Yu, Kai Key Laboratory of Shanghai Education Commission for Intelligent Interaction Cognitive Engineering SpeechLab Department of Computer Science and Engineering Brain Science and Technology Research Center Shanghai Jiao Tong University Shanghai China

Linear Discriminant Analysis (LDA) has been used as a standard post-processing procedure in many state-of-the-art speaker recognition tasks. Through maximizing the inter-speaker difference and minimizing the intra-speaker variation, LDA projects i-vectors to a lower-dimensional and more discriminative subspace. In this paper, we propose a neural network based compensation scheme(termed as deep discriminant analysis, DDA) for i-vector based speaker recognition, which shares the spirit with LDA. Optimized against softmax loss and center loss at the same time, the proposed method learns a more compact and discriminative embedding space. Compared with the Gaussian distribution assumption of data and the learnt linear projection in LDA, the proposed method doesn't pose any assumptions on data and can learn a non-linear projection function. Experiments are carried out on a short-duration text-independent dataset based on the SRE Corpus, noticeable performance improvement can be observed against the normal LDA or PLDA methods. Copyright © 2018, The Authors. All rights reserved.

关键词： Discriminant analysis

来源：评论

学校读者我要写书评

暂无评论

DDGNet-YOLO: A Target Detection Network For Dangerous Driving Gestures

DDGNet-YOLO: A Target Detection Network For Dangerous Drivin...

引用

Chinese Automation Congress (CAC)

作者： Yefan Zhou Zhao Lv Yongtai Zhou Anhui Province Key Laboratory of Multimodal Cognitive Computation School of Computer Science and Technology Anhui University Hefei China Institute of Physical Science and Information Technology Hefei China Zhejiang Key Laboratory for Brain-Machine Collaborative Intelligence Hangzhou Dianzi University Hangzhou China

ISBN: (数字)9781728176871

ISBN: (纸本)9781728176888

As we all know, some bad driving behaviors such as using mobile phone, eating, drinking etc., may lead to traffic accidents. In order to reduce these potential dangers, we proposed a target detection network to detect the abovementioned dangerous driving gestures. We have improved the YOLOv4 target detection network and designed the DDGNet- YOLO target detection network, so as to reduce the detection time of dangerous driving gesture. Our algorithm have been performed on the CVRR-HANDS 3D dataset. The mAP@0.75 of dangerous driving gesture is 76.4%, and the average real-time processing speed of the algorithm was 143 frames per second.

关键词： Three-dimensional displays Automation Lighting Object detection Jitter Real-time systems Mobile handsets

来源：评论

学校读者我要写书评

暂无评论

Joint I-Vector with End-to-End System for Short Duration Text-Independent Speaker Verification

Joint I-Vector with End-to-End System for Short Duration Tex...

引用

IEEE International Conference on Acoustics, Speech and Signal Processing

作者： Zili Huang Shuai Wang Yanmin Qian Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering SpeechLab Department of Computer Science and Engineering Brain Science and Technology Research Center Shanghai Jiao Tong University Shanghai China

ISBN: (纸本)9781538646595

Factor analysis based i-vector has been the state-of-the-art method for speaker verification. Recently, researchers propose to build DNN based end-to-end speaker verification systems and achieve comparable performance with i -vector. Since these two methods possess their own property and differ from each other significantly, we explore a framework to integrate these two paradigms together to utilize their complementarity. More specifically, in this paper we develop and compare four methodologies to integrate traditional i -vector into end-to-end systems, including score fusion, embeddings concatenation, transformed concatenation and joint learning. All these approaches achieve significant gains. Moreover, the hard trial selection is performed on the end-to-end architecture which further improves the performance. Experimental results on a text-independent short-duration dataset generated from SRE 2010 reveal that the newly proposed method reduces the EER by relative 31.0% and 28.2% compared to the i-vector and end-to-end baselines respectively.

关键词： speaker verification end-to-end i-vector triplet loss hard trial selection speaker recognition end-to-end end system Loudspeakers Joints energy efficiency ratio Short circuits

来源：评论

学校读者我要写书评

暂无评论

Deep brain stimulation for psychiatric disorders: roleof imaging in identifying/confirming DBS targets,predicting, and optimizing outcome and unravellingmechanisms of action

引用

Psychoradiology 2021年第3期1卷 118-151页

作者： Dejan Georgiev Harith Akram Marjan Jahanshahi Department of Clinical and Movement Neurosciences UCL Queen Square Institute of NeurologyQueen SquareLondonWC1N 3BGUK Department of Neurology University Medical Centre LjubljanaZaloska cesta 21000 LjubljanaSlovenia Artificial Intelligence Laboratory Faculty of Computer and Information ScienceUniversity of LjubljanaVecna pot 1131000 LjubljanaSlovenia The Clinical Hospital of Chengdu Brain Science Institute MOE Key Laboratory for NeuroinformationUniversity of Electronic Science and Technology of ChinaChengdu611731China

Following the established application of deep brain stimulation(DBS)in the treatment of movement disorders,new non-neurological indications have emerged,such as for obsessive–compulsive disorders,major depressive disorder,dementia,Gilles de la Tourette Syndrome,anorexia nervosa,and *** DBS is a network mod-ulation surgical treatment,the development of DBS for both neurological and psychiatric disorders has been partly driven by advances in neuroimaging,which has helped explain the brain networks *** in magnetic resonance imaging connectivity and electrophysiology have led to the development of the con-cept of modulating widely distributed,complex brain ***,the increasing number of targets for treating psychiatric disorders have indicated that there may be a convergence of the effect of stimulating dif-ferent targets for the same disorder,and the effect of stimulating the same target for different *** aim of this paper is to review the imaging studies of DBS for psychiatric ***,and particularly connectivity analysis,offers exceptional opportunities to better understand and even predict the clinical out-comes of DBS,especially where there is a lack of objective biomarkers that are essential to properly guide DBS pre-and *** future,imaging might also prove useful to individualize DBS ***,one of the most important aspects of imaging in DBS is that it allows us to better understand the brain through observing the changes of the functional connectome under neuromodulation,which may in turn help explain the mechanisms of action of DBS that remain elusive.

关键词： deep brain stimulation psychiatric disorders major depressive disorder obsessive–compulsive disorder dementia

来源：评论

学校读者我要写书评

暂无评论

A Deep Learning Method for Colon Polyp Segmentation with Better Universality

A Deep Learning Method for Colon Polyp Segmentation with Bet...

引用

Medical Artificial Intelligence (MedAI), IEEE International Conference on

作者： Yuanyuan Wang Zhaohong Deng Jiangsu Key Laboratory of Media Design and Software Technology School of Artificial Intelligence and Computer Science Jiangnan University Wuxi China Key Laboratory of Computational Neuroscience and Brain-Inspired Intelligence (LCNBI) and ZJLab. School of Artificial Intelligence and Computer Science Jiangnan University Wuxi China

Colon polyp segmentation is an important work in the adjuvant treatment of colon diseases. Traditional colon polyp segmentation methods are mainly designed according to the underlying characteristics of polyps, but the size of polyp tissues is usually different. Thus the segmentation accuracy of such methods is low. In recent years, deep learning colon polyp segmentation methods developed based on vision Transformer have higher segmentation accuracy than traditional methods on some colon polyp datasets. However, this method is more specific to a certain dataset and is not universal. In order to address the above challenges, this paper proposes a general colon polyp segmentation method-MixFormNet, which can effectively improve the segmentation accuracy of multiple datasets. MixFormNet is a encoder-decoder structure. MixFormNet utilizes a Mix Transformer with multi-scale output as an encoder. Then this paper designs a new decoder structure for multi-scale output. The decoder designed in this paper processes the characteristics of the different scale feature maps output from the encoder separately, with the aim of maximizing the use of the feature information contained in the deep and shallow feature maps. Specifically, the low-level output of the encoder has a large resolution but contains weak semantic information. Therefore, the convolutional neural network is constructed in the decoder and applied to the low-level output in order to obtain richer local feature information. A large number of experiments have shown that MixFormNet effectively improves the segmentation accuracy of multiple colon polyp datasets. In particular, the mDice of the CVC-ClinicDB dataset reached 0.94. It is proved that MixFormNet has good universality and effectiveness in improving the segmentation accuracy of colon polyp datasets.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Learning hierarchical translation spans

Learning hierarchical translation spans

引用

2014 Conference on Empirical Methods in Natural Language Processing, EMNLP 2014

作者： Zhang, Jingyi Utiyama, Masao Sumita, Eiichro Zhao, Hai Department of Computer Science and Engineering Center for Brain-Like Computing and Machine Intelligence Shanghai Jiao Tong Unviersity Shanghai200240 China Key Laboratory Shanghai Education Commission Intelligent Interaction Cognitive Engineering Shanghai Jiao Tong Unviersity Shanghai200240 China National Institute of Information and Communications Technology 3-5Hikaridai Keihanna Science City Kyoto619-0289 Japan

ISBN: (纸本)9781937284961

We propose a simple and effective approach to learn translation spans for the hierarchical phrase-based translation model. Our model evaluates if a source span should be covered by translation rules during decoding, which is integrated into the translation system as soft constraints. Compared to syntactic constraints, our model is directly acquired from an aligned parallel corpus and does not require parsers. Rich source side contextual features and advanced machine learning methods were utilized for this learning task. The proposed approach was evaluated on NTCIR-9 Chinese-English and Japanese-English translation tasks and showed significant improvement over the baseline system. © 2014 Association for Computational Linguistics.

关键词： Syntactics

来源：评论

学校读者我要写书评

暂无评论

{\cal U}Boost: Boosting with the Universum

引用

IEEE Transactions on Pattern Analysis and Machine Intelligence 2012年第4期34卷 825-832页

作者： Chunhua Shen Peng Wang Fumin Shen Hanzi Wang Australian Center for Visual Technologies and School of Computer Science University of Adelaide North Terrace SA Australia School of Automation Science and Electrical Engineering Beihang University Beijing China School of Computer Science and Technology Nanjing University of Science and Technology of PLA Nanjing China Cognitive Science Department School of Information Science and Technology and Fujian Key Laboratory of the Brain-like Intelligent Systems Xiamen University Xiamen Fujian China School of Computer Science University of Adelaide SA Australia

It has been shown that the Universum data, which do not belong to either class of the classification problem of interest, may contain useful prior domain knowledge for training a classifier [1], [2]. In this work, we design a novel boosting algorithm that takes advantage of the available Universum data, hence the name UBoost. UBoost is a boosting implementation of Vapnik's alternative capacity concept to the large margin approach. In addition to the standard regularization term, UBoost also controls the learned model's capacity by maximizing the number of observed contradictions. Our experiments demonstrate that UBoost can deliver improved classification accuracy over standard boosting algorithms that use labeled data alone.

关键词： Boosting Training Educational institutions Optimization Algorithm design and analysis Histograms Accuracy

来源：评论

学校读者我要写书评

暂无评论

Theoretical Analyses of Evolutionary Algorithms on Time-Linkage OneMax with General Weights

arXiv

引用

arXiv 2023年

作者： Zheng, Weijie Yao, Xin School of Computer Science and Technology International Research Institute for Artificial Intelligence Harbin Institute of Technology Shenzhen China Guangdong Provincial Key Laboratory of Brain-inspired Intelligent Computation Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen China Cercia School of Computer Science University of Birmingham Birmingham United Kingdom

Evolutionary computation has shown its superiority in dynamic optimization, but for the (dynamic) time-linkage problems, some theoretical studies have revealed the possible weakness of evolutionary computation. Since the theoretically analyzed time-linkage problem only considers the influence of an extremely strong negative time-linkage effect, it remains unclear whether the weakness also appears in problems with more general time-linkage effects. Besides, understanding in depth the relationship between time-linkage effect and algorithmic features is important to build up our knowledge of what algorithmic features are good at what kinds of problems. In this paper, we analyze the general time-linkage effect and consider the time-linkage OneMax with general weights whose absolute values reflect the strength and whose sign reflects the positive or negative influence. We prove that except for some small and positive time-linkage effects (that is, for weights 0 and 1), randomized local search (RLS) and (1 + 1) EA cannot converge to the global optimum with a positive probability. More precisely, for the negative time-linkage effect (for negative weights), both algorithms cannot efficiently reach the global optimum and the probability of failing to converge to the global optimum is at least 1-o(1). For the not so small positive time-linkage effect (positive weights greater than 1), such a probability is at most c + o(1) where c is a constant strictly less than 1. Copyright © 2023, The Authors. All rights reserved.

关键词： Evolutionary algorithms

来源：评论

学校读者我要写书评

暂无评论

Data-Driven Preference Sampling for Pareto Front Learning

arXiv

引用

arXiv 2024年

作者： Ye, Rongguang Chen, Lei Liao, Weiduo Zhang, Jinyuan Ishibuchi, Hisao Guangdong Provincial Key Laboratory of Brain-inspired Intelligent Computation Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen518055 China School of Mathematics and Statistics Guangdong University of Technology Guangzhou51000 China

Pareto front learning is a technique that introduces preference vectors in a neural network to approximate the Pareto front. Previous Pareto front learning methods have demonstrated high performance in approximating simple Pareto fronts. These methods often sample preference vectors from a fixed Dirichlet distribution. However, no fixed sampling distribution can be adapted to diverse Pareto fronts. Efficiently sampling preference vectors and accurately estimating the Pareto front is a challenge. To address this challenge, we propose a data-driven preference vector sampling framework for Pareto front learning. We utilize the posterior information of the objective functions to adjust the parameters of the sampling distribution flexibly. In this manner, the proposed method can sample preference vectors from the location of the Pareto front with a high probability. Moreover, we design the distribution of the preference vector as a mixture of Dirichlet distributions to improve the performance of the model in disconnected Pareto fronts. Extensive experiments validate the superiority of the proposed method compared with state-of-the-art *** Codes 68T05 Copyright © 2024, The Authors. All rights reserved.

关键词： Vectors

来源：评论

学校读者我要写书评

暂无评论

Random Occlusion Recovery with Noise Channel for Person Re-identification 16th

Random Occlusion Recovery with Noise Channel for Person Re-i...

引用

16th International Conference on Intelligent Computing, ICIC 2020

作者： Zhang, Kun Wu, Di Yuan, Changan Qin, Xiao Wu, Hongjie Zhao, Xingming Zhang, Lijun Du, Yuchuan Wang, Hanli Institute of Machine Learning and Systems Biology School of Electronics and Information Engineering Tongji University Shanghai China Guangxi Academy of Science Nanning530025 China School of Computer and Information Engineering Nanning Normal University Nanning530299 China School of Computer Science and Technology Soochow University Suzhou215006 China School of Electronic and Information Engineering Suzhou University of Science and Technology Suzhou215009 China Fudan University Shanghai200433 China Key Laboratory of Computational Neuroscience and Brain-Inspired Intelligence Ministry of Education Shanghai China Collaborative Innovation Center of Intelligent New Energy Vehicle and School of Automotive Studies Tongji University Shanghai201804 China The Key Laboratory of Road and Traffic Engineering of the Ministry of Education Department of Transportation Engineering Tongji University Shanghai201804 China Department of Computer Science and Technology the Key Laboratory of Embedded System and Service Computing and Shanghai Institute of Intelligent Science and Technology Tongji University Shanghai200092 China

ISBN: (纸本)9783030607982

Person re-identification, as the basic task of a multi-camera surveillance system, plays an important role in a variety of surveillance applications. However, the current mainstream person re-identification model based on deep learning requires a lot of labeled data, which takes a lot of time and manpower. In this study, we proposed a person re-identification method based on random occlusion recovery with noise channel. We add random occlusion blocks to the original image, use the GAN model for repair, and use the repaired image to expand the original training set. After that, the generated image is adjusted through the noise channel. Finally, we use the enhanced data set to train the baseline model. Our model achieves the state-of-the-art on Market-1501 dataset, proving that the method is effective. © 2020, Springer Nature Switzerland AG.

关键词： Generative adversarial networks

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：