检索结果-内蒙古大学图书馆

Towards Compact Broad Learning System by Combined Sparse Regularization

INTERNATIONAL JOURNAL OF INFORMATION TECHNOLOGY & DECISION MAKING 2022年第1期21卷 169-194页

作者： Miao, Jianyu Yang, Tiejun Jin, Jun-Wei Sun, Lijun Niu, Lingfeng Shi, Yong Minist Educ Key Lab Grain Informat Proc & Control HAUT Zhengzhou 450001 Peoples R China Henan Key Lab Grain Photoelect Detect & Control H Zhengzhou 450001 Peoples R China Henan Univ Technol Coll Artificial Intelligence & Big Data Zhengzhou 450001 Peoples R China Henan Univ Technol Coll Informat Sci & Engn Zhengzhou 450001 Peoples R China Chinese Acad Sci Key Lab Big Data Min & Knowledge Management Beijing 100190 Peoples R China Univ Chinese Acad Sci Sch Econ & Management Beijing 100190 Peoples R China Southwest Minzu Univ Sch Elect & Informat Engn Chengdu 610041 Peoples R China Univ Nebraska Omaha Coll Informat Sci & Technol Omaha NE 68182 USA

Broad Learning System (BLS) has been proven to be one of the most important techniques for classification and regression in machine learning and data mining. BLS directly collects all the features from feature and enhancement nodes as input of the output layer, which neglects vast amounts of redundant information. It usually leads to be inefficient and overfitting. To resolve this issue, we propose sparse regularization-based compact broad learning system (CBLS) framework, which can simultaneously remove redundant nodes and weights. To be more specific, we use group sparse regularization based on l(2,1) norm to promote the competition between different nodes and then remove redundant nodes, and a class of nonconvex sparsity regularization to promote the competition between different weights and then remove redundant weights. To optimize the resulting problem of the proposed CBLS, we exploit an efficient alternative optimization algorithm based on proximal gradient method together with computational complexity. Finally, extensive experiments on the classification task are conducted on public benchmark datasets to verify the effectiveness and superiority of the proposed CBLS.

关键词： Broad learning system (BLS) group sparsity nonconvex regularization proximal gradient method

来源：评论

学校读者我要写书评

暂无评论

Improved Diversity-Promoting Collaborative Metric Learning for Recommendation

引用

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 2024年第12期46卷 9004-9022页

作者： Bao, Shilong Xu, Qianqian Yang, Zhiyong He, Yuan Cao, Xiaochun Huang, Qingming Chinese Acad Sci Inst Informat Engn State Key Lab Informat Secur SKLOIS Beijing 100093 Peoples R China Univ Chinese Acad Sci Sch Cyber Secur Beijing 100049 Peoples R China Chinese Acad Sci Inst Comp Technol Key Lab Intelligent Informat Proc Beijing 100190 Peoples R China Univ Chinese Acad Sci Sch Comp Sci & Technol Beijing 101408 Peoples R China Alibaba Grp Secur Dept Hangzhou 311121 Peoples R China Sun Yat Sen Univ Sch Cyber Sci & Technol Shenzhen Campus Shenzhen 518107 Peoples R China Univ Chinese Acad Sci Sch Comp Sci & Technol Key Lab Big Data Min & Knowledge Management BDKM Beijing 101408 Peoples R China Peng Cheng Lab Shenzhen 518055 Peoples R China

Collaborative Metric Learning (CML) has recently emerged as a popular method in recommendation systems (RS), closing the gap between metric learning and collaborative filtering. Following the convention of RS, existing practices exploit unique user representation in their model design. This paper focuses on a challenging scenario where a user has multiple categories of interests. Under this setting, the unique user representation might induce preference bias, especially when the item category distribution is imbalanced. To address this issue, we propose a novel method called Diversity-Promoting Collaborative Metric Learning (DPCML), with the hope of considering the commonly ignored minority interest of the user. The key idea behind DPCML is to introduce a set of multiple representations for each user in the system where users' preference toward an item is aggregated by taking the minimum item-user distance among their embedding set. Specifically, we instantiate two effective assignment strategies to explore a proper quantity of vectors for each user. Meanwhile, a Diversity Control Regularization Scheme (DCRS) is developed to accommodate the multi-vector representation strategy better. Theoretically, we show that DPCML could induce a smaller generalization error than traditional CML. Furthermore, we notice that CML-based approaches usually require negative sampling to reduce the heavy computational burden caused by the pairwise objective therein. In this paper, we reveal the fundamental limitation of the widely adopted hard-aware sampling from the One-Way Partial AUC (OPAUC) perspective and then develop an effective sampling alternative for the CML-based paradigm. Finally, comprehensive experiments over a range of benchmark datasets speak to the efficacy of DPCML.

关键词： Collaborative metric learning (CML) machine learning partial AUC optimization recommendation system

来源：评论

学校读者我要写书评

暂无评论

Mega research infrastructure as a driver for high-quality development and innovation: Promoting scientific cooperation and interdisciplinarity☆

PROJECT LEADERSHIP AND SOCIETY

引用

PROJECT LEADERSHIP AND SOCIETY 2025年 6卷

作者： Xing, Ying Wu, Yuejia Xiao, Xiao Wang, Daojuan Zhang, Ling-ling Univ Chinese Acad Sci Sino Danish Coll Beijing 100190 Peoples R China Univ Chinese Acad Sci Sch Humanities Beijing 100049 Peoples R China Univ Chinese Acad Sci Sch Econ & Management Beijing 100190 Peoples R China Aalborg Univ Business Sch DK-9200 Aalborg Denmark Chinese Acad Sci Key Lab Big Data Min & Knowledge Management Beijing 100190 Peoples R China MOE Social Sci Lab Digital Econ Forecasts & Policy Beijing 100190 Peoples R China

Mega research infrastructures, as megaprojects, play a crucial role in promoting high-quality development, particularly in relation to innovation. They contribute to the development of disciplinary layouts and scientific collaborations. Additionally, during the operational phase of a infrastructure, maximizing efficiency and ensuring sustainable development become key aspects of project management. The operation of mega research infrastructures involves increasing levels of interdisciplinarity, interagency collaboration, and cross-regional cooperation. Therefore, it is essential to further clarify the characteristics of scientific collaboration within these facilities and understand the current status and evolving trends of disciplinary and collaborative networks. This work serves as a fundamental and necessary endeavor. In this study, we employ social network analysis methods to investigate the Shanghai Synchrotron Radiation Facility as a case study. We analyze the network structure and evolution of its collaborative and disciplinary networks from 2009 to 2022. The findings reveal the enormous potential and value of mega research infrastructures in promoting scientific research cooperation and development. They emphasize the need for increased attention and support for instrument scientists by facility managers, as well as the importance of expanding collaboration areas and formats to enrich the range of disciplines involved. The contribution of this study lies in its systematic analysis of the evolutionary process of collaborative and disciplinary networks within mega research infrastructures. It reveals the scientific value of these facilities in driving scientific research cooperation and interdisciplinary integration, providing a scientific basis for project management and policy formulation in the operational phase of mega research infrastructures. The research will align the strategic development direction of mega research infrastructures with the requirement

关键词： Social network analysis Collaborative network Disciplinary network Mega research infrastructure Innovation High-quality development

来源：评论

学校读者我要写书评

暂无评论

Two-Stage Polishing Network for Camouflaged Object Detection 11th

Two-Stage Polishing Network for Camouflaged Object Detectio...

引用

11th International Conference on Image and Graphics, ICIG 2021

作者： Jiang, Xuan Wu, Zhe Zhang, Yajie Su, Li Huang, Qingming University of Chinese Academy of Sciences Beijing China Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of Sciences Beijing China Peng Cheng Laboratory Shenzhen China

ISBN: (纸本)9783030873547

Due to the needs of medical and military fields, Camouflaged Object Detection (COD) becomes one of important branches of object detection. It has gradually gained people’s attention in recent years. How to correctly locate camouflaged objects accurately segment them are the main problems in this field. The COD task is far different from SOD because of the more complex background with similar colors and textures. Therefore it is more challenging. At present, there are still few existing methods, and lacks targeted method for the edge detection problem. In this paper, we propose a novel Two-Stage Polishing Network (TSPNet). This network consists of Front Feature Fusion Module (FFFM) and Polishing Module (PM). FFFM adopts Cross-modal Feature Aggregation and Global and Local Feature Aggregation to capture global context information and detail local information simultaneously. Meanwhile, PM uses edge truth map as supervision imformation to further study the object edge features. The experiments was conducted on three available datasets and the result shows that the proposed framework outperforms state-of-the-arts. Besides, TSPNet is compact with 50% model size saving than existed COD models. © 2021, Springer Nature Switzerland AG.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Bidirectional Logits Tree: Pursuing Granularity Reconcilement in Fine-Grained Classification

arXiv

引用

arXiv 2024年

作者： Lu, Zhiguang Xu, Qianqian Bao, Shilong Yang, Zhiyong Huang, Qingming Key Laboratory of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences China School of Computer Science and Technology University of Chinese Academy of Sciences China Key Laboratory of Big Data Mining and Knowledge Management University of Chinese Academy of Sciences China

This paper addresses the challenge of Granularity Competition in fine-grained classification tasks, which arises due to the semantic gap between multi-granularity labels. Existing approaches typically develop independent hierarchy-aware models based on shared features extracted from a common base encoder. However, because coarse-grained levels are inherently easier to learn than finer ones, the base encoder tends to prioritize coarse feature abstractions, which impedes the learning of fine-grained features. To overcome this challenge, we propose a novel framework called the Bidirectional Logits Tree (BiLT) for Granularity Reconcilement. The key idea is to develop classifiers sequentially from the finest to the coarsest granularities, rather than parallelly constructing a set of classifiers based on the same input features. In this setup, the outputs of finer-grained classifiers serve as inputs for coarser-grained ones, facilitating the flow of hierarchical semantic information across different granularities. On top of this, we further introduce an Adaptive Intra-Granularity Difference Learning (AIGDL) approach to uncover subtle semantic differences between classes within the same granularity. Extensive experiments demonstrate the effectiveness of our proposed method. Copyright © 2024, The Authors. All rights reserved.

关键词： Coarse-grained modeling

来源：评论

学校读者我要写书评

暂无评论

The impact of Russia-Ukraine war on crude oil prices: an EMC framework

HUMANITIES & SOCIAL SCIENCES COMMUNICATIONS

引用

HUMANITIES & SOCIAL SCIENCES COMMUNICATIONS 2024年第1期11卷 1-12页

作者： Zhang, Qi Hu, Yi Jiao, Jianbin Wang, Shouyang Univ Chinese Acad Sci Sch Emergency Management Sci & Engn Beijing 100190 Peoples R China Univ Chinese Acad Sci Sch Econ & Management Beijing 100190 Peoples R China Chinese Acad Sci Key Lab Big Data Min & Knowledge Management Beijing 100190 Peoples R China UCAS MOE Social Sci Lab Digital Econ Forecasts & Policy Sim Beijing 100190 Peoples R China Chinese Acad Sci Acad Math & Syst Sci Beijing 100190 Peoples R China ShanghaiTech Univ Sch Entrepreneurship & Management Shanghai 201210 Peoples R China

As the second-largest oil producer and natural gas exporter, Russia's war with Ukraine has severely impacted the energy market. To what extent has the war influenced crude oil prices, and has it altered the long-term dynamics of oil prices? An objective analysis of the effects of the Russia-Ukraine war on the crude oil market can assist relevant entities in developing both short-term emergency strategies and long-term response plans. This study establishes an analytical framework of the event analysis method based on multiresolution causality testing (EMC). The results of the multiresolution causality testing reveal a significant one-way causality between the Russia-Ukraine war and crude oil prices. Afterward, using the event analysis based on variational mode decomposition (VMD), from October 1, 2021, to August 25, 2022, as the event window, we found that the war and its chain events caused the West Texas Intermediate (WTI) crude oil prices to increase by $37.14, a 52.33% surge, and the Brent crude oil price to rise by $41.49, a 56.33% increase. During the event window, the Russia-Ukraine war can account for 70.72% and 73.62% of the fluctuation in WTI and Brent crude oil prices, respectively. Furthermore, the war amplified oil price volatility and fundamentally altered the trend of crude oil prices. Consequently, this study proposes four recommendations: the establishment of an emergency management mechanism for the oil market, the diversification of oil and gas imports by energy-importing countries, the steady advancement of energy transformation, and the judicious use of financial instruments by enterprises to hedge risks.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Sparse Learning for Neural Networks with A Generalized Sparse Regularization

引用

Procedia Computer Science 2022年 214卷 747-754页

作者： Anda Tang Tongsheng Yao Lingfeng Niu Yong Shi School of Mathematical Sciences University of Chinese Academy of Sciences Beijing 100190 China Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of Sciences Beijing 100190 China

Deep neural networks (DNNs) is very important and have achieved remarkable accuracies in tasks such as image processing. However, the success of DNNs heavily relies on excessive computation and parameter storage costs. To cut down the overheads, a wide range of regularization terms are proposed on netwok compression, which have their own scope of application though. To further reduce computation, structural sparsity learning is a matter of great concern where group sparse regularizations are blossoming in radiant splendour. However, the group sparse regularizations used in network compression is relatively scarce. A majority of structural compression only stem from ℓ 2,1 regularization. In addition, sparse regularization in structural network compression lacks unified form and theoretical guidance. Therefore, we focus our attention on a generalized sparse regularization in sparse learning of deep learning. We put a series of sparse regularization into a framework where some group sparse regularizations with different properties are introduced. The transformation of regularizations can be completed through the selection of hyper-parameters. In this case, we can use the optimization strategy for unification and theoretical guidance, and transform different sparse regularizations to adapt to different tasks. To our knowledge, it is the first work applying the generalized sparse regularization with novel group sparsity in compression of DNNs. The proposed ℓ p , q , r regularization can achieve both neuron-level and connection-level sparsity. And we give the analytical solutions for some specific ( p,q, r ) thus the compression can be achieved throughout the standard optimization process. We perform extensive experiments to illustrate the advantages and characteristics of our new method for further applications.

关键词： Deep neural networks Model compression Group Sparse optimization Regularization Sparse Learning

来源：评论

学校读者我要写书评

暂无评论

knowledge Graph Embedding by Double Limit Scoring Loss

引用

IEEE TRANSACTIONS ON knowledge AND data ENGINEERING 2022年第12期34卷 5825-5839页

作者： Zhou, Xiaofei Niu, Lingfeng Zhu, Qiannan Zhu, Xingquan Liu, Ping Tan, Jianlong Guo, Li Chinese Acad Sci Inst Informat Engn Beijing 100093 Peoples R China Univ Chinese Acad Sci Beijing 100049 Peoples R China Chinese Acad Sci Key Lab Big Data Min & Knowledge Management Beijing 100190 Peoples R China Florida Atlantic Univ Dept Comp & Elect Engn & Comp Sci Boca Raton FL 33431 USA

knowledge graph embedding is an effective way to represent knowledge graph, which greatly enhance the performances on knowledge graph completion tasks, e.g., entity or relation prediction. For knowledge graph embedding models, designing a powerful loss framework is crucial to the discrimination between correct and incorrect triplets. Margin-based ranking loss is a commonly used negative sampling framework to make a suitable margin between the scores of positive and negative triples. However, this loss can not ensure ideal low scores for the positive triplets and high scores for the negative triplets, which is not beneficial for knowledge completion tasks. In this paper, we present a double limit scoring loss to separately set upper bound for correct triplets and lower bound for incorrect triplets, which provides more effective and flexible optimization for knowledge graph embedding. Upon the presented loss framework, we present several knowledge graph embedding models including TransE-SS, TransH-SS, TransD-SS, ProjE-SS and ComplEx-SS. The experimental results on link prediction and triplet classification show that our proposed models have the significant improvement compared to state-of-the-art baselines.

关键词： Semantics Optimization Computational modeling Task analysis Tensors Sparse matrices Predictive models knowledge graph embedding representation learning knowledge graph completion loss function

来源：评论

学校读者我要写书评

暂无评论

Types and Evolutionary Boundaries of Agents 5

Types and Evolutionary Boundaries of Agents

引用

5th Asian Conference on Artificial Intelligence Technology, ACAIT 2021

作者： Liu, Feng Liu, Ying Research Center on Fictitious Economy and Data Science Chinese Academy of Sciences Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences School of Economics and Management Beijing China

ISBN: (纸本)9781665426305

Through the analysis of the characteristics of natural intelligent systems, the three laws of intelligence are exploringly proposed. The most important point is that any agent is a system with knowledge input, output, (dynamic) storage and creation abilities Since an agent has four abilities, and each ability has three states, including zero, infinity, and finite (between zero and infinity), 81 types of agents can be formed. Therefore, it can be inferred that Alpha point (No. 1), where all the four abilities of an agent are zero, and Omega point (No. 81), where all four abilities are infinite are the boundary of the evolution of Agent. This paper designs experiments to detect these 81 agents, and finds that 11 of them can be confirmed by experiments at present. Through the discussion of the states of Agents No. 64 to No. 81, fresh thinking is proposed for the generation and realization of the paradox. © 2021 IEEE.

关键词： Evolution of Agents Types of Agents α Point Ω Point

来源：评论

学校读者我要写书评

暂无评论

Pyramid scheme in stock market:A kind of financial market simulation

引用

Chinese Physics B 2021年第9期30卷 624-632页

作者： Yong Shi Bo Li Guang-Le Du School of Economics and Management University of Chinese Academy of SciencesBeijing 100190China Research Center on Fictitious Economy and Data Science Chinese Academy of SciencesBeijing 100190China Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of SciencesBeijing 100190China Wenzhou Institute University of Chinese Academy of SciencesWenzhou 325001China School of Physical Sciences University of Chinese Academy of SciencesBeijing 100049China

Artificial stock market simulation based on agent is an important means to study financial *** on the assumption that the investors are composed of a main fund,small trend and contrarian investors characterized by four parameters,we simulate and research a kind of financial phenomenon with the characteristics of pyramid *** simulation results and theoretical analysis reveal the relationships between the rate of return of the main fund and the proportion of the trend investors in all small investors,the small investors'parameters of taking profit and stopping loss,the order size of the main fund and the strategies adopted by the main *** work is helpful to explain the financial phenomenon with the characteristics of pyramid schemes in financial markets,design trading rules for regulators and develop trading strategies for investors.

关键词： pyramid scheme agent-based model artificial stock market herd behavior

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：