检索结果-内蒙古大学图书馆

42nd Chinese Control Conference, CCC 2023

作者： Liu, Wenkang Sun, Mingyi Li, Yang Beihang University The Department of Automation Science and Electrical Engineering Beijing100191 China The Beijing Advanced Innovation Center for Big Data and Brain Computing Beijing100191 China Beihang University State Key Laboratory of Virtual Reality Technology and Systems Beijing100191 China Peking University Advanced Institute of Information Technology Hangzhou310000 China

ISBN: (纸本)9789887581543

In recent years, the demand for facial expression recognition applications has increased rapidly, and its research has received extensive attention from researchers. However, the current recognition methods based on deep learning ignore the idea of multiple head attention and semantic consistency, resulting in the model only paying attention to the local area of the feature map. In addition, the model's inconsistent attention to the images before and after the flip, resulting the poor robustness, poor interpretation, and other shortcomings in the model. To address the above problems, we propose an Affinity Separation Loss (ASLoss), which improves the separability of samples through clustering. Moreover, a Separate Multi-head Attention block (SMA), and a Zonal Loss (ZLoss) are also designed to decentralize the model's attention. Experimental results demonstrate that our proposed MACNet method achieves competitive recognition performance on two public datasets RAF-DB and FERPlus. © 2023 Technical Committee on Control Theory, Chinese Association of Automation.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

Reusing Deep Neural Network Models through Model Re-engineering

Reusing Deep Neural Network Models through Model Re-engineer...

引用

International Conference on Software Engineering (ICSE)

作者： Binhang Qi Hailong Sun Xiang Gao Hongyu Zhang Zhaotian Li Xudong Liu SKLSDE Lab Beihang University Beijing China Beijing Advanced Innovation Center for Big Data and Brain Computing Beihang University Beijing China Chongqing University Chongqing China

Training deep neural network (DNN) models, which has become an important task in today's software development, is often costly in terms of computational resources and time. With the inspiration of software reuse, building DNN models through reusing existing ones has gained increasing attention recently. Prior approaches to DNN model reuse have two main limitations: 1) reusing the entire model, while only a small part of the model's functionalities (labels) are required, would cause much overhead (e.g., computational and time costs for inference), and 2) model reuse would inherit the defects and weaknesses of the reused model, and hence put the new system under threats of security attack. To solve the above problem, we propose SeaM, a tool that re-engineers a trained DNN model to improve its reusability. Specifically, given a target problem and a trained model, SeaM utilizes a gradient-based search method to search for the model's weights that are relevant to the target problem. The re-engineered model that only retains the relevant weights is then reused to solve the target problem. Evaluation results on widely-used models show that the re-engineered models produced by SeaM only contain 10.11% weights of the original models, resulting 42.41% reduction in terms of inference time. For the target problem, the re-engineered models even outperform the original models in classification accuracy by 5.85%. Moreover, reusing the re-engineered models inherits an average of 57% fewer defects than reusing the entire model. We believe our approach to reducing reuse overhead and defect inheritance is one important step forward for practical model reuse.

关键词：

来源：评论

学校读者我要写书评

暂无评论

CLDG: Contrastive Learning on Dynamic Graphs

arXiv

引用

arXiv 2024年

作者： Xu, Yiming Shi, Bin Ma, Teng Dong, Bo Zhou, Haoyi Zheng, Qinghua Department of Computer Science and Technology Xi’an Jiaotong University China Shaanxi Provincial Key Laboratory of Big Data Knowledge Engineering Xi’an Jiaotong University China Department of Distance Education Xi’an Jiaotong University China School of Software Beihang University China Advanced Innovation Center for Big Data and Brain Computing Beihang University China

The graph with complex annotations is the most potent data type, whose constantly evolving motivates further exploration of the unsupervised dynamic graph representation. One of the representative paradigms is graph contrastive learning. It constructs self-supervised signals by maximizing the mutual information between the statistic graph’s augmentation views. However, the semantics and labels may change within the augmentation process, causing a significant performance drop in downstream tasks. This drawback becomes greatly magnified on dynamic graphs. To address this problem, we designed a simple yet effective framework named CLDG. Firstly, we elaborate that dynamic graphs have temporal translation invariance at different levels. Then, we proposed a sampling layer to extract the temporally-persistent signals. It will encourage the node to maintain consistent local and global representations, i.e., temporal translation invariance under the timespan views. The extensive experiments demonstrate the effectiveness and efficiency of the method on seven datasets by outperforming eight unsupervised state-of-the-art baselines and showing competitiveness against four semi-supervised methods. Compared with the existing dynamic graph method, the number of model parameters and training time is reduced by an average of 2,001.86 times and 130.31 times on seven datasets, respectively. The code and data are available at: https://***/yimingxu24/CLDG. © 2024, CC BY.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Prompt-based Unifying Inference Attack on Graph Neural Networks 39

Prompt-based Unifying Inference Attack on Graph Neural Netwo...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Wei, Yuecen Fu, Xingcheng Liu, Lingyun Sun, Qingyun Peng, Hao Hu, Chunming School of Software Beihang University Beijing China Beijing Advanced Innovation Center for Big Data and Brain Computing Beihang University Beijing China Key Lab of Education Blockchain and Intelligent Technology Ministry of Education Guangxi Normal University China

ISBN: (纸本)157735897X

Graph neural networks (GNNs) provide important prospective insights in applications such as social behavior analysis and financial risk analysis based on their powerful learning capabilities on graph data. Nevertheless, GNNs' predictive performance relies on the quality of task-specific node labels, so it is common practice to improve the model's generalization ability in the downstream execution of decision-making tasks through pre-training. Graph prompting is a prudent choice but risky without taking measures to prevent data leakage. In other words, in high-risk decision scenarios, prompt learning can infer private information by accessing model parameters trained on private data (publishing model parameters in pre-training, i.e., without directly leaking the raw data, is a tacitly accepted trend). However, myriad graph inference attacks necessitate tailored module design and processing to enhance inference capabilities due to variations in supervision signals. In this paper, we propose a novel Prompt-based unifying Inference Attack framework on GNNs, named ProIA. Specifically, ProIA retains the crucial topological information of the graph during pre-training, enhancing the background knowledge of the inference attack model. It then utilizes a unified prompt and introduces additional disentanglement factors in downstream attacks to adapt to task-relevant knowledge. Finally, extensive experiments show that ProIA enhances attack capabilities and demonstrates remarkable adaptability to various inference attacks. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

V2I Based Environment Perception for Autonomous Vehicles at Intersections

引用

China Communications 2021年第7期18卷 1-12页

作者： Xuting Duan Hang Jiang Daxin Tian Tianyuan Zou Jianshan Zhou Yue Cao Beijing Advanced Innovation Center for Big Data and Brain Computing Beijing Key Laboratory for Cooperative Vehicle Infrastructure Systems&Safety ControlSchool of Transportation Science and EngineeringBeihang UniversityBeijing 100191China School of Cyber Science and Engineering Wuhan UniversityWuhan 430072China

In recent years,autonomous driving technology has made good progress,but the noncooperative intelligence of vehicle for autonomous driving still has many technical bottlenecks when facing urban road autonomous driving challenges.V2I(Vehicle-to-Infrastructure)communication is a potential solution to enable cooperative intelligence of vehicles and *** this paper,the RGB-PVRCNN,an environment perception framework,is proposed to improve the environmental awareness of autonomous vehicles at intersections by leveraging V2I communication *** framework integrates vision feature based on *** normal distributions transform(NDT)point cloud registration algorithm is deployed both on onboard and roadside to obtain the position of the autonomous vehicles and to build the local map objects detected by roadside multi-sensor system are sent back to autonomous vehicles to enhance the perception ability of autonomous vehicles for benefiting path planning and traffic efficiency at the *** field-testing results show that our method can effectively extend the environmental perception ability and range of autonomous vehicles at the intersection and outperform the PointPillar algorithm and the VoxelRCNN algorithm in detection accuracy.

关键词： V2I environmental perception autonomous vehicles 3D objects detection

来源：评论

学校读者我要写书评

暂无评论

A Low-Code Development Framework for Constructing Industrial Apps 15th

A Low-Code Development Framework for Constructing Industrial...

引用

15th CCF Conference on Computer Supported Cooperative Work and Social computing, Chinese CSCW 2020

作者： Wang, Jingyue Qi, Binhang Zhang, Wentao Sun, Hailong SKLSDE Lab School of Computer Science and Engineering Beihang University Beijing100191 China Beijing Advanced Innovation Center for Big Data and Brain Computing Beijing100191 China

ISBN: (纸本)9789811625398

With the advent of the Industry 4.0, intelligent manufacturing has become a technological highland to conquer in the process of enterprise digitalization. As the core competitiveness of intelligent manufacturing, industrial apps, with new features such as customization and lightweight, has emerged as a new type of industrial software. Traditional development methods and tools can hardly meet the large demand of industrial software on account of its long development cycles while low-code development can greatly improve the productivity of industrial software, lower the barriers and reduce costs for development. Therefore, the research and application of low-code development for industrial apps has received much attention. Industrial Internet platforms such as Siemens, OutSystems have successively launched low-code tools. However, there is still a lack of an open, unified low-code development framework in industry. In response to the above problems, we propose a low-code framework to develop industrial apps quickly and easily, which paves the way for leveraging the crowd intelligence of worldwide developers to improve the productivity of developing industrial apps. Based on BPMN2.0 and Apache Activiti engine, this framework provides drag-and-drop process design, one-click process deployment and operation, data monitoring and other functions. In this paper, we present a prototype system of a low-code development framework and demonstrate its functions through a use case of developing a predictive maintenance application. Finally, the aircraft turbine life data is used to verify the effectiveness of the system. © 2021, Springer Nature Singapore Pte Ltd.

关键词： Productivity

来源：评论

学校读者我要写书评

暂无评论

Instructing the Architecture Search for Spatial-temporal Sequence Forecasting with LLM

arXiv

引用

arXiv 2025年

作者： Xue, Xin Zhou, Haoyi Chen, Tianyu Zhang, Shuai Long, Yizhou Li, Jianxin Department of Computer Science Beihang University Beijing100191 China Beijing Advanced Innovation Center for Big Data and Brain Computing Beijing100191 China Department of Software Beihang University Beijing100191 China Zhongguancun Laboratory Beijing100190 China

Spatial-temporal sequence forecasting (STSF) is a long-standing research problem with widespread real-world applications. Neural architecture search (NAS), which automates the neural network design, has been shown effective in tackling the STSF problem. However, the existing NAS methods for STSF focus on generating architectures in a time-consuming data-driven fashion, which heavily limits their ability to use background knowledge and explore the complicated search trajectory. Large language models (LLMs) have shown remarkable ability in decision-making with comprehensive internal world knowledge, but how it could benefit NAS for STSF remains unexplored. In this paper, we propose a novel NAS method for STSF based on LLM. Instead of directly generate architectures with LLM, We inspire the LLM’s capability with a multi-level enhancement mechanism. Specifically, on the step-level, we decompose the generation task into decision steps with powerful prompt engineering and inspire LLM to serve as instructor for architecture search based on its internal knowledge. On the instance-level, we utilize a one-step tuning framework to quickly evaluate the architecture instance and a memory bank to cumulate knowledge to improve LLM’s search ability. On the task-level, we propose a two-stage architecture search, balancing the exploration stage and optimization stage, to reduce the possibility of being trapped in local optima. Extensive experimental results demonstrate that our method can achieve competitive effectiveness with superior efficiency against existing NAS methods for STSF. © 2025, CC BY-NC-ND.

关键词： Spatio-temporal data

来源：评论

学校读者我要写书评

暂无评论

Triplet-Aware Graph Neural Networks for Factorized Multi-Modal Knowledge Graph Entity Alignment

SSRN

引用

SSRN 2023年

作者： Li, Qian Li, Jianxin Wu, Jia Peng, Xutan Ji, Cheng Peng, Hao Wang, Lihong Yu, Philip S. Beijing Advanced Innovation Center for Big Data and Brain Computing China Beijing China School of Computing Macquarie University Sydney Australia The University of Sheffield South Yorkshire United Kingdom China University of Illinois Chicago Chicago United States

Multi-Modal Entity Alignment (MMEA), aiming to discover matching entity pairs on two multi-modal knowledge graphs (MMKGs), is an essential task in knowledge graph fusion. Through mining feature information of MMKGs, entities are aligned to tackle the issue that an MMKG is incapable of effective integration. The recent attempt at neighbors and attribute fusion mainly focuses on aggregating multi-modal attributes, neglecting the structure effect with multi-modal attributes for entity alignment. This paper proposes an innovative approach, namely TriFac, to exploit embedding refinement for factorizing the original multi-modal knowledge graphs through a two-stage MMKG factorization. Notably, we propose triplet-aware graph neural networks to aggregate multi-relational features. We propose multi-modal fusion for aggregating multiple features and design three novel metrics to measure knowledge graph factorization performance on the unified factorized latent space. Empirical results indicate the effectiveness of TriFac, surpassing previous state-of-the-art models on two MMEA datasets and a power system dataset. © 2023, The Authors. All rights reserved.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Graph algorithms: parallelization and scalability

引用

Science China(Information Sciences) 2020年第10期63卷 234-254页

作者： Wenfei FAN Kun HE Qian LI Yue WANG School of Informatics University of Edinburgh Shenzhen Institute of Computing Sciences Shenzhen University Beijing Advanced Innovation Center for Big Data and Brain Computing Beihang University Guangdong Province Key Laboratory of Popular High Performance Computers Shenzhen University

For computations on large-scale graphs, one often resorts to parallel algorithms. However, parallel algorithms are difficult to write, debug and analyze. Worse still, it is difficult to make algorithms parallelly scalable, such that the more machines are used, the faster the algorithms run. Indeed, it is not yet known whether any PTIME computational problems admit parallelly scalable algorithms on shared-nothing *** it possible to parallelize sequential graph algorithms and guarantee convergence at the correct results as long as the sequential algorithms are correct? Moreover, does a PTIME parallelly scalable problem exist on shared-nothing systems? This position paper answers both questions in the affirmative.

关键词： parallelization parallel scalability PTIME problems graph algorithms shared-nothing systems

来源：评论

学校读者我要写书评

暂无评论

Dual Contrastive Learning: Text Classification via Label-Aware data Augmentation

arXiv

引用

arXiv 2022年

作者： Chen, Qianben Zhang, Richong Zheng, Yaowei Mao, Yongyi SKLSDE School of Computer Science and Engineering Beihang University Beijing China Beijing Advanced Institution on Big Data and Brain Computing Beihang University Beijing China School of Electrical Engineering and Computer Science University of Ottawa Ottawa Canada

Contrastive learning has achieved remarkable success in representation learning via self-supervision in unsupervised settings. However, effectively adapting contrastive learning to supervised learning tasks remains as a challenge in practice. In this work, we introduce a dual contrastive learning (DualCL) framework that simultaneously learns the features of input samples and the parameters of classifiers in the same space. Specifically, DualCL regards the parameters of the classifiers as augmented samples associating to different labels and then exploits the contrastive learning between the input samples and the augmented samples. Empirical studies on five benchmark text classification datasets and their low-resource version demonstrate the improvement in classification accuracy and confirm the capability of learning discriminative representations of DualCL. Copyright © 2022, The Authors. All rights reserved.

关键词： Text processing

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：