检索结果-南通市图书馆

Combat data shift in few-shot learning with knowledge graph

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Frontiers of Computer Science 2023年第1期17卷 101-111页

作者： Yongchun ZHU Fuzhen ZHUANG Xiangliang ZHANG Zhiyuan QI zhiping shi Juan CAO Qing HE Key Lab of Intelligent Information Processing of Chinese Academy of Sciences(CAS) Institute of Computing TechnologyCASBeijing 100190China University of Chinese Academy of Sciences Beijing 100049China Institute of Artificial Intelligence Beihang UniversityBeijing 100191China Xiamen Institute of Data Intelligence Xiamen 361021China Computer Science and Engineering University of Notre DameIN 46556USA University of California San Diego La JollaCA 92093USA College of Information Engineering&Academy for Multidisciplinary Studies Capital Normal UniversityBeijing 100089China

Many few-shot learning approaches have been designed under the meta-learning framework, which learns from a variety of learning tasks and generalizes to new tasks. These meta-learning approaches achieve the expected performance in the scenario where all samples are drawn from the same distributions (i.i.d. observations). However, in real-world applications, few-shot learning paradigm often suffers from data shift, i.e., samples in different tasks, even in the same task, could be drawn from various data distributions. Most existing few-shot learning approaches are not designed with the consideration of data shift, and thus show downgraded performance when data distribution shifts. However, it is non-trivial to address the data shift problem in few-shot learning, due to the limited number of labeled samples in each task. Targeting at addressing this problem, we propose a novel metric-based meta-learning framework to extract task-specific representations and task-shared representations with the help of knowledge graph. The data shift within/between tasks can thus be combated by the combination of task-shared and task-specific representations. The proposed model is evaluated on popular benchmarks and two constructed new challenging datasets. The evaluation results demonstrate its remarkable performance.

关键词： few-shot data shift knowledge graph

Observer-based Passive Control for Neutral Type Descriptor Markovian Jump Systems 36

在线全文

学校读者我要写书评

暂无评论

Observer-based Passive Control for Neutral Type Descriptor M...

Blue emitting CsPbBr3 perovskite quantum dot inks obtained from sustained release tablets

第36届中国控制会议

作者： Chunjiao Wang Yuechao Ma zhiping shi Institute of Electrical Engineering Yanshan University College of Science Yanshan University State Grid LiaoNing Electric Power Co.Ltd Branch Materials

ISBN: (纸本)9781538629185

The problem of delay-dependent observer-based robustly passive control is addressed for neutral type descriptor Markov jump systems with time-varying delays. The parameter uncertainties of the system are assumed unknown and norm-bounded. Using Lyapunov-Krasovskii functional method and weighted matrix, a delay-dependent robustly passive condition for closed-loop system is given. Then, the sufficient conditions for the existence of designed observer are presented in terms of linear matrix inequalities（LMI）, and the parameters of controller are provided with the algorithm to solve the problem. Finally, a numerical example is presented to illustrate the validity of the approach.

关键词： neutral type descriptor systems Markov jump delay-dependent passive control

来源： cnki会议评论

在线全文

cnki会议

学校读者我要写书评

暂无评论

Nano Research 2019年第12期12卷 3129-3134页

作者： Hewei Yang Yaqing Feng Zhiyu Tu Kuo Su Xiaozhi Fan Bingjie Liu zhiping shi Yuzhe Zhang Chenyang Zhao Bao Zhan School of Chemical Engineering and Technology Tianjin UniversityTianjin 300350China Collaborative Innovation Center of Chemical Science and Engineering(Tianjin) Tianjin 300072China School of Life Sciences Tianjin UniversityTianjin 300072China

Blue emitting perovskite ink obtained from cesium lead halide quantum dots bearing chlorine(CsPbClxBr3-x,0 详细信息

Blue emitting perovskite ink obtained from cesium lead halide quantum dots bearing chlorine(CsPbClxBr3-x,0shift of emission wavelength to blue ***,by polymerizing acrylamide under UV-light irradiation to form polymer gel networks in dimethyl sulfoxide(DMSO)with CsPbBr3 precursors and passivating agents trapped,wesuccessfully prepared novel sustained release tablets with different shapes and *** to the limitation of the polymer networks onsolve nt releasi ng,the resulting CsPbBr3 qua ntum dots have the average size of 1.1±0.2 *** the basis of the excelle nt quantum confin eme nteffect and optimized surface passivation,the obtained PQD ink can emit high quality blue light for more than 6 *** work elucidates anew and convenient technique to prepare blue emission perovskite quantum dots ink with high stability and photoluminescence qua ntumyield and provides a great potential technology for the preparation of perovskite optoelectronic devices.

关键词： sustained release tablet perovskite quantum dot blue emitting perovskite ink polymer gel surface passivation

HIHO Decoding Algorithms of Turbo Product Codes for High Rate Optical Communications

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

China Communications 2012年第6期9卷 114-123页

作者： shi zhiping Zhou Liang Fan Xiang Yang Yang National Key Laboratory of Science and Technology on Communications University of Electronic Science andTechnology of ChinaChengdu 611731P.R.China

This paper presents a new Hard-Input Hard-Output (HIHO) iterative decoding algorithm for Turbo Product Codes (TPC), and especially describes the BCH-TPC codes aiming to alleviate error propagation and lower error floor. This algorithm mainly emp hasizes a decision mechanism for bit-flips, which thoroughly evaluates four different aspects of the decoding process, properly weighs and combines their respective reliability measures, and then employs the combined measure to make a judgment with regard to whether any particular bit should be flipped or not. Simulations result in a very steep Bit Error Rate (BER) curve indicating that a high-level net coding gain can be expected with a reasonable complexity. The simplicity and effectiveness of this HIHO decoding algorithm makes it a p romising candidate for the application in future high-speed fiber optical communications.

关键词： turbo product codes HIHO decodingalgorithm reliability measurement threshold selec-tion op tical communication

在线全文

学校读者我要写书评

暂无评论

基于自组织动态神经网络的图像分割

计算机研究与发展 2009年第1期46卷 23-30页

作者：史春奇施智平刘曦史忠植中国科学院计算技术研究所智能信息处理实验室北京100190 中国科学院研究生院北京100049

图像分割是图像处理和模式识别的重要课题,而图像特征空间聚类是图像分割的一种重要方法,认为图像的特征是图像中待分割物体表面所特有而且恒定的特征,并将图像的特征映射到某种几何空间,称为特征空间,并且假定图像中不同的待分割物体... 详细信息

图像分割是图像处理和模式识别的重要课题,而图像特征空间聚类是图像分割的一种重要方法,认为图像的特征是图像中待分割物体表面所特有而且恒定的特征,并将图像的特征映射到某种几何空间,称为特征空间,并且假定图像中不同的待分割物体在该特征空间中呈现为不同的聚集.提出了自组织动态网络(SODNN)聚类算法,并且利用该算法对图像特征空间聚类.该算法实现了神经网络结构的快速生长和动态调节,具有自动适应数据内在分布特征和聚类结果更为准确稳定的特点.利用SODNN算法对图像颜色空间进行聚类的同时综合了图像的位置信息来实现图像分割.实验表明分割结果与人工分割结果具有较好的一致性.

关键词：竞争学习自组织特征空间聚类图像分割图像特征

Analysis of Abnormal Seat Belt Shoulder Force Under the Front Collision Conditions

在线全文

学校读者我要写书评

暂无评论

Analysis of Abnormal Seat Belt Shoulder Force Under the Fron...

综合语义特征和视觉特征的二阶段纹理图像检索

2023中国汽车工程学会年会暨展览会

作者： shi zhiping Li Jiawen Chen Hongji Dongfeng-Nissan Passenger Vehicle Co. Ltd.

The phenomenon of abnormal seat belt load limit force under front collision conditions was *** on the seat belt restraint system model and FTA,the shoulder force in each stage under front collision conditions was analyzed,with a focus on analyzing various possible factors of abnormal shoulder force during the force limiting stage and FTA model for abnormal shoulder force was *** on a specific practical case,analyze the impact of each factor in the FTA model on abnormal shoulder force and eventually form a force limiting correlation table for the root cause.A comprehensive analysis of the abnormal seat belt shoulder force from the constraint system level is summarized,which can quickly and comprehensively identify the root cause for the abnormal shoulder force and has a certain reference significance in the industry.

关键词：

来源： cnki会议评论

在线全文

cnki会议

学校读者我要写书评

暂无评论

计算机辅助设计与图形学学报 2008年第4期20卷 499-505页

作者：李清勇施智平史忠植北京交通大学计算机与信息技术学院北京100044 中国科学院计算技术研究所智能信息处理重点实验室北京100190

提出一种融合模糊语义概念和精细视觉特征的纹理图像检索方法.首先根据语言表达式和模糊语义概念对整个图像库进行快速有效的粗搜索,得到具有"软"边界的语义检索结果;然后根据视觉特征在语义检索结果中(而不是整个图像库)进行精细的检索... 详细信息

提出一种融合模糊语义概念和精细视觉特征的纹理图像检索方法.首先根据语言表达式和模糊语义概念对整个图像库进行快速有效的粗搜索,得到具有"软"边界的语义检索结果;然后根据视觉特征在语义检索结果中(而不是整个图像库)进行精细的检索.该方法很好地结合了基于内容的图像检索和基于语义的图像检索两者的优势,使得用户既可以根据语义概念对图像库进行快速浏览和检索,也能根据查询用例图像的视觉特征进行精细的匹配;另外,由粗到细的二阶段策略也明显地提高了其检索性能.在Brodatz纹理库中的实验结果表明,通过调整合适的语义检索边界,该方法的检索性能明显优于基于视觉特征的图像检索方法.

关键词：基于内容的图像检索语义特征视觉特征语言变量

在线全文

学校读者我要写书评

暂无评论

集成视觉特征和语义信息的相关反馈方法

计算机辅助设计与图形学学报 2007年第9期19卷 1138-1142页

作者：施智平李清勇史俊史忠植中国科学院计算技术研究所智能信息处理重点实验室北京100080 北京交通大学计算机与信息技术学院北京100044

为了有效地利用图像检索系统的语义分类信息和视觉特征,提出一种基于Bayes的集成视觉特征和语义信息的相关反馈检索方法.首先,将图像库的数据经语义监督的视觉特征聚类算法划分为小的聚类,每个聚类内数据的视觉特征相似并且语义类别相同... 详细信息

为了有效地利用图像检索系统的语义分类信息和视觉特征,提出一种基于Bayes的集成视觉特征和语义信息的相关反馈检索方法.首先,将图像库的数据经语义监督的视觉特征聚类算法划分为小的聚类,每个聚类内数据的视觉特征相似并且语义类别相同;然后以聚类为单位标注正负反馈的实例,这显著区别于以单个图像为单位的相关反馈过程;最后分别以基于视觉特征的Bayes分类器和基于语义的Bayes分类器修正相似距离.在图像库上的实验表明,只用较少的反馈次数就可以达到较高的检索准确率.

关键词：相关反馈图像检索聚类 Bayes

一种CBR与RBR相结合的快速预案生成系统

在线全文

学校读者我要写书评

暂无评论

计算机研究与发展 2007年第4期44卷 660-666页

作者：罗杰文施智平何清史忠植中国科学院计算技术研究所智能信息处理重点实验室

将范例推理(case basedreasoning,CBR)与规则推理(rule basedreasoning,RBR)两种人工智能技术相结合,实现一种快速预案生成系统.它有效地解决了单纯RBR系统在预案生成过程中的时间延迟缺陷和知识库难以获取的瓶颈.通过CBR工具,能够把以... 详细信息

将范例推理(case basedreasoning,CBR)与规则推理(rule basedreasoning,RBR)两种人工智能技术相结合,实现一种快速预案生成系统.它有效地解决了单纯RBR系统在预案生成过程中的时间延迟缺陷和知识库难以获取的瓶颈.通过CBR工具,能够把以前发生的紧急事件和解决方案生成预案.一旦新的事件发生,首先从预案库中进行案例的相似性检索,如果没有检索到预案或者检索到的预案匹配度很低,再采用RBR系统对紧急事件进行规则推理,然后把推理结果重新存入预案库.实验数据表明,这种方法对单纯RBR系统在时间响应上进行了有效的优化.另外,因为案例的获取比专家系统推理规则的获取容易得多,它同时解决了RBR系统推理规则难以获取的瓶颈.根据这种思想,实现了CBR与RBR结合的快速预案生成系统.目前,它已经应用到抗洪抢险的预案生成和城市应急联动的决策支持上,效果表明它在预案生成速度以及实际可操作性上都具有明显优势.

关键词：规则推理范例推理紧急事件决策支持相似度主体