检索结果-内蒙古大学图书馆

Unsupervised and Semi-Supervised Domain Adaptation Networks Based on Instance-Level Contrastive Prototype Matching for Motor Imagery Classification

SSRN

引用

SSRN 2023年

作者： Zhang, Dongxue Li, Huiying Xie, Jingmeng College of Computer Science and Technology Jilin Province Changchun China Key Laboratory of Symbol Computation and Knowledge Engineering Jilin University Changchun130012 China Xi’an Jiaotong University College of Electronic information Shanxi Province Xi’an China

Due to the convenience and safety of EEG signals, it is widely used in motor imagery (MI) brain-computer interface (BCI). However, there exist distribution differences between different sessions of the same subject. Models trained with data from previous sessions (source domain) often perform poorly on the current sessions (target domain). Domain adaptation techniques have met this challenge to great effect. However, most of these existing studies only consider coarse-grained methods with global information. In this paper, we propose a method for instance-level comparative prototype (ICP) analysis. The approach incorporates global domain alignment and fine-grained local subdomain alignment to simultaneously account for both global and local differences. Depending on the amount of labeled data available in the target domain, the method is implemented in both unsupervised and semi-supervised versions. Specifically, at the domain level, we employ the maximum mean difference (MMD) loss to globally constrain the feature space, achieving comprehensive alignment. In the context of class-level operations, we propose two memory banks designed to house class prototypes within each domain. We meticulously regulate feature embeddings through the application of two contrastive losses based on prototypes. Among them, source contrastive loss is used to spatially organize source features according to categories, thereby mediating inter- and intra-class relationships. At the same time, we propose an entropy-aware interactive contrastive loss to dynamically and personalized apply class information across domains by evaluating the confidence level of target data. To validate our approach, extensive experiments were conducted on a highly regarded public EEG dataset, namely Dataset IIa of the BCI Competition IV, as well as a large-scale EEG dataset called GigaDB. The experiments yielded average classification accuracies of 86.03% and 84.22% respectively. These results demonstrate that o

关键词： Electroencephalography

来源：评论

学校读者我要写书评

暂无评论

Prediction of multiple sequence alignment based on an improved particle swarm optimization

引用

Journal of Bionanoscience 2013年第1期7卷 92-96页

作者： Gao, Yu-Xi Wang, Yan-Min Pei, Zhi-Li Department of Computer Science Jilin Business and Technology College China Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education College of Computer Science and Technology Jilin University China College of Computer Science and Technology Inner Mongolia University for the Nationalities Tongliao 028000 China

In this paper we use an improved Particle Swarm Optimization algorithm to solve Multiple Sequence Alignment (MSA). MSA is a key problem in bioinformatics. The thesis starts with the theory of Particle swarm optimization algorithm and the design an improved algorithm for Multiple Sequence Alignment. The simulated experiments show that the improved algorithm surpasses existing algorithms in performance. Copyright © 2013 American Scientific Publishers.

关键词： Bioinformatics

来源：评论

学校读者我要写书评

暂无评论

Ontology based automatic attributes extracting and queries translating for deep web

Journal of Software

引用

Journal of Software 2010年第7期5卷 713-720页

作者： Liang, Hao Ren, Fei Zuo, Wanli He, Fengling College of Computer Science and Technology Jilin University Changchun 130012 China Department of Information Changchun Taxation College Changchun 130117 China China Development Bank Center of Operations Beijing 100037 China Key Laboratory of Symbol Computation and Knowledge Engineering of the Ministry of Education Changchun 130012 China

Search engines and web crawlers can not access the Deep Web directly. The workable way to access the hidden database is through query interfaces. Automatic extracting attributes from query interfaces and translating queries is a solvable way for addressing the current limitations in accessing Deep Web. However, the query interface provides semantic constraints, some attributes are co-occurred and the others are exclusive sometimes. To generate a valid query, we have to reconcile the key attributes and semantic relation between them. We design a framework to automatically extract attributes from query interfaces taking full advantage of instances information and enrich the attribute sets embedded in the semantic query interface by Ontology technique. Each attribute is extended into a candidate attribute expressed by a hierarchy tree and describes the semantic relation of the attributes. We carry out our experiments in the real-world domain and results showed the validation of query translation framework. © 2010 ACADEMY PUBLISHER.

关键词： Ontology

来源：评论

学校读者我要写书评

暂无评论

BAENET: A Brain Age Estimation Network with 3D Skipping and Outlier Constraint Loss

BAENET: A Brain Age Estimation Network with 3D Skipping and ...

引用

IEEE International Symposium on Biomedical Imaging

作者： Taiping Qu Yangming Yue Qirui Zhang Cheng Wang Zhiqiang Zhang Guangming Lu Wei Du Xiuli Li Key Laboratory of Symbol Computation and Knowledge Engineering of the Ministry of Education College of Computer Science and Technology Jilin University Changchun China Deepwise AI Lab Beijing China Jinling Hospital Medical School of Nanjing University Nanjing Jiangsu China

ISBN: (数字)9781538693308

ISBN: (纸本)9781538693315

The potential pattern changes in brain micro-structure can be used for the brain development assessment in children and adolescents by MRI scans. In this paper, we propose a highly accurate and efficient end-to-end brain age estimation network (BAENET) on T1-weighted MRI images. On the network, 3D skipping and outlier constraint loss are designed to accommodate deeper network and increase the robustness. Besides, we incorporate the neuroimaging domain knowledge into stratified sampling for better generalization ability for datasets with different age distributions, and gender learning for more gender-specific features during modeling. We verify the effectiveness of the proposed method on the public ABIDE2 and ADHD200 benchmark, consisting of 382 and 378 normal children scans respectively. Our BAENET achieves MAE of 1.11 and 1.16, significantly outperforming the best reported methods by 5.1 % and 9.4%.

关键词： Three-dimensional displays Brain modeling Estimation Solid modeling Feature extraction Magnetic resonance imaging Task analysis

来源：评论

学校读者我要写书评

暂无评论

Pancreatic cancer biomarker detection using recursive feature elimination based on Support Vector Machine and large margin distribution machine

Pancreatic cancer biomarker detection using recursive featur...

引用

International Conference on Systems and Informatics (ICSAI)

作者： Yidan Lv Yan Wang Yongfei Tan Wei Du Keke Liu Hao Wang Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education Jilin University Changchun China Basic Medicine School Jilin University Changchun China Department of Hepatopancreatobiliary Surgery and Second Affiliated Hospital Harbin Medical University Harbin China

Pancreatic cancer acts as one of the leading causes of cancer-connective deaths. Its five-year overall survival rate being reported is about 7.7% from 2006 to 2012 by the National Cancer Institute. One of the main causes for its poor prognosis is because its non-typical symptoms make early diagnosis very challenging. Therefore, a predominant strategy for early accurate detection and prognostication on pancreatic cancer is vital to the whole course of comprehensive therapy. In this research, we proposed a method which combined Recursive Feature Elimination (RFE) method based on Support Vector Machine (SVM) and Large Margin Distribution Machine (LDM) to identify potential biomarkers for pancreatic cancer. In our experiments, we have strengthened the process of RFE to achieve better performance. The dataset GSE15471 we adopted are from GEO database with 39 pairs of pancreatic ductal carcinoma and adjacent control pancreatic tissues. Through experiments, a panel of twelve genes was identified as biomarkers in pancreatic cancer with 91.28% classification accuracy. The universality of the candidate genes was examined on another dataset GSE28735 and the classification accuracy was higher than 80%. In addition, by using the SVM, LDM and BP classifiers, we compared the ordered feature sets generated by our proposed method with T-test, SVM-RFE and LDM-RFE, and the results indicated our proposed method obtained higher average classification accuracy.

关键词： Cancer Support vector machines Feature extraction Tumors Training data Testing

来源：评论

学校读者我要写书评

暂无评论

A New Seamless Multi-resolution Simplification Method for the Terrain Model

A New Seamless Multi-resolution Simplification Method for th...

引用

Web Intelligence and Intelligent Agent Technology Workshops (WI-IAT)

作者： Huijie Zhang Jigui Sun Changsong Qi Nan Lv College of Computer Northeast Normal University Changchun China College of Computer Science & Technology Jilin University Changchun China College of Computer Science & Technology Key Laboratory of Symbol Computation and Knowledge Engineering of the Ministry of Education Jilin University China

This paper presents a fast method to construct the simplified terrain model with the multi-resolution. In this method, we adopt the normal quad-tree hierarchy to subdivide the original terrain surface into the multi- resolution levels. For the sake of implementing the compressive storage and the efficient index for the elevation data, a connotative hierarchy is proposed and its corresponding index strategies are deduced. On the basis of the connotative hierarchy, we focus on resolving the crack problem of the simplified model. We firstly create the space filling curve during the simplification process to accelerate the search for the potential cracks. Afterwards, the different approaches are utilized to process the cracks in terms of the dissimilar terrain features. For this reason, we put forward an evaluation function that can selfadaptively identify the terrain feature according to the normal vector angle of the adjacent nodes in the connotative hierarchy. Consequently, there are not redundant triangles in our seamless multi-resolution terrain model. The proposed approaches are experimented on the real data and the results show that our method is relatively efficient and robust. Besides, the seamless simplified model has less number of triangles than the other common algorithm.

关键词： Multiresolution analysis Educational institutions Intelligent agent Computer science Surface cracks Filling Conferences Acceleration Robustness Sun

来源：评论

学校读者我要写书评

暂无评论

Adaptive and iterative gene selection based on least squares support vector regression

引用

Journal of Information and computational Science 2006年第3期3卷 443-451页

作者： Jiang, Jingqing Wu, Chunguo Song, Chuyi Liang, Yanchun Key Laboratory of Symbol Computation and Knowledge Engineering College of Computer Science and Technology Jilin University Changchun 130012 China College of Mathematics and Computer Science Inner Mongolia University for Nationalities Tongliao 028043 China Key Laboratory of Information Science and Engineering of Railway Ministry The Key Laboratory of Advanced Information Science and Network Technology of Beijing Beijing Jiaotong University Beijing 100044 China

Cancer classification and identification are major areas in medical research. DNA microarrays could provide useful information for cancer classification at the gene expression level. The number of genes in a microarray is always several thousands while the number of training samples always several dozens. In such case most of the machine learning models suffer from the overfitting and it is necessary to select a handful of most informative genes. An adaptive and iterative gene selection algorithm based on least squares support vector machines is proposed in this paper. The algorithm adopts sequential forward selection search scheme. The number of selected genes can be determined adaptively. The total number of genes processed by the proposed algorithm is smaller than that processed by other algorithms using support vector machines. Results of numerical experiments show that the proposed algorithm trains fast and achieves comparable performance on two well-known benchmark problems.

关键词： Genes

来源：评论

学校读者我要写书评

暂无评论

Prediction of ncRNA based on least squares support vector machine

引用

Journal of Bionanoscience 2013年第1期7卷 121-125页

作者： Xing, Chong Wan, Liming Wang, Jiaxin Liang, Yanchun College of Electrical and Information Engineering Guanghua College of Changchun University Changchun 130033 China College of Computer Science and Technology Jilin University Key Laboratory of Symbol Computation and Knowledge Engineering of Ministry of Education Changchun 130012 China Research Institute on General Development and Evaluation of Equipment EAAF of PLA Beijing 100076 China

In this paper, on the basis of principal component analysis, we use least squares support vector machine (LS-SVM) to predict tRNA. Appearance frequencies of single nucleotide, 2-nucleotides, (G-C)% and (A-T)% were chosen as characteristic inputs. Results from tests showed that the prediction accuracy was 90.51% on prokaryotic tRNA dataset. The results indicate that the proposed method is adoptable for prokaryotic ncRNA prediction. Copyright © 2013 American Scientific Publishers.

关键词： RNA

来源：评论

学校读者我要写书评

暂无评论

Improved cross-media relevance model for automatic image annotation

引用

ICIC Express Letters, Part B: Applications 2012年第5期3卷 1231-1237页

作者： Song, Haiyu Li, Xiongfei Wang, Pengjie Bao, Cuizhu Liu, Yongkui Xu, Lu Institute of Computer Graphics and Image Processing Dalian Nationalities University No. 18 Liaohe Road Economic and Technological Development Zone-Dalian 116600 China Key Laboratory of Symbol Computation and Knowledge Engineering Ministry of Education Jilin University No. 2699 Qianjin Avenue Changchun 130012 China

Automatic image annotation has been an active research topic in the last decade due to its potentially large impact on image retrieval, object recognition and image understanding. Many approaches have been proposed for image annotation in recent years. The most influential annotation model is relevance model, whose representative algorithm is Cross-Media Relevance Model (CMRM). In this paper, to overcome the shortcomings of the original CMRM, we propose improved CMRM. In the improved CMRM, the annotation word distribution is estimated using binary probability, the significance of blob is discriminately depicted using TF-IDF (term frequency-inverse document frequency) schemes, and the similarity measurement of blob is based on blob and continuousvalued feature vector. The experimental results on the standard Corel5K dataset show that our proposed approach outperforms the original CMRM by about 111% and 73% in terms of F1 and N+ measures respectively. The improved CMRM can get better performance in specific domains such as cultural heritage and special necessities digitization for ethnic minorities. © 2012 ICIC International.

关键词： Image retrieval

来源：评论

学校读者我要写书评

暂无评论

Prewavelets with finite decomposition and finite reconstruction

引用

Journal of Information and computational Science 2007年第1期4卷 35-41页

作者： Guan, Yujing Li, Ying Jiang, Zhengyi Institute of Mathematics Jilin University Changchun 130021 China State Key Laboratory of Intelligent Technology and Systems Department of Computer Science and Technology Tsinghua University Beijing 100084 China College of Computer Science and Technology Jilin University Changchun 130021 China Key Laboratory of Symbol Computation and Knowledge Engineering Changchun 130021 China

The necessary and sufficient condition for the existence of prewavelets with finite decomposition and finite reconstruction is presented. Especially when r = 1, we show that {φj(x-k) | 1 &le j &le r, k ∈ s} is shift-orthogonal if there exist prewavelets with finite decomposition and finite reconstruction, i.e., non-trivial prewavelets with finite decomposition and finite reconstruction does not exist. As an example, when r = 2 we construct a prewavelet with finite decomposition and finite reconstruction where the scale function is not shift-orthogonal, which demonstrate that there exist vector prewavelets with finite decomposition and finite reconstruction.

关键词： Wavelet transforms

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：