检索结果-内蒙古大学图书馆

Layer-Wise Learning Rate Optimization for Task-Dependent Fine-Tuning of Pre-Trained Models: An Evolutionary Approach

引用

ACM Transactions on Evolutionary Learning and Optimization 2024年第4期4卷 1-23页

作者： Bu, Chenyang Liu, Yuxin Huang, Manzong Shao, Jianxuan Ji, Shengwei Luo, Wenjian Wu, Xindong Key Laboratory of Knowledge Engineering with Big Data Ministry of Education and School of Computer Science and Information Engineering Hefei University of Technology Hefei China School of Artificial Intelligence and Big Data Hefei University Hefei China Guangdong Provincial Key Laboratory of Novel Security Intelligence Technologies School of Computer Science and Technology Harbin Institute of Technology Shenzhen China

The superior performance of large-scale pre-Trained models, such as Bidirectional Encoder Representations from Transformers (BERT) and Generative Pre-Trained Transformer (GPT), has received increasing attention in both academic and industrial research and has become one of the current research hotspots. A pre-Trained model refers to a model trained on large-scale unlab.led data, whose purpose is to learn general language representation or features for fine-Tuning or transfer learning in subsequent tasks. After pre-Training is complete, a small amount of lab.led data can be used to fine-Tune the model for a specific task or domain. This two-stage method of "pre-Training+fine-Tuning"has achieved advanced results in natural language processing (NLP) tasks. Despite widespread adoption, existing fixed fine-Tuning schemes that adapt well to one NLP task may perform inconsistently on other NLP tasks given that different tasks have different latent semantic structures. In this article, we explore the effectiveness of automatic fine-Tuning pattern search for layer-wise learning rates from an evolutionary optimization perspective. Our goal is to use evolutionary algorithms to search for better task-dependent fine-Tuning patterns for specific NLP tasks than typical fixed fine-Tuning patterns. Experimental results on two real-world language benchmarks and three advanced pre-Training language models show the effectiveness and generality of the proposed framework. © 2024 held by the owner/author(s).

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Aggregation-based dual heterogeneous task allocation in spatial crowdsourcing

引用

Frontiers of computer science 2024年第6期18卷 215-227页

作者： Xiaochuan LIN Kaimin WEI Zhetao LI Jinpeng CHEN Tingrui PEI College of Information Science and Technology&Cyberspace Security Jinan UniversityGuangzhou 510632China National&Local Joint Engineering Research Center of Network Security Detection and Protection Technology Guangzhou 510632China Guangdong Provincial Key Laboratory of Data Security and Privacy Protection Guangzhou 510632China School of Computer Science Beijing University of Posts and TelecommunicationsBeijing 100876China

Spatial crowdsourcing(SC)is a popular data collection paradigm for numerous *** the increment of tasks and workers in SC,heterogeneity becomes an unavoidable difficulty in task *** researches only focus on the single-heterogeneous task ***,a variety of heterogeneous objects coexist in real-world SC *** dramatically expands the space for searching the optimal task allocation solution,affecting the quality and efficiency of data *** this paper,an aggregation-based dual heterogeneous task allocation algorithm is put *** investigates the impact of dual heterogeneous on the task allocation problem and seeks to maximize the quality of task completion and minimize the average travel *** problem is first proved to be ***,a task aggregation method based on locations and requirements is built to reduce task ***,a time-constrained shortest path planning is also developed to shorten the travel distance in a *** that,two evolutionary task allocation schemes are ***,extensive experiments are conducted based on real-world datasets in various *** with baseline algorithms,our proposed schemes enhance the quality of task completion by up to 25% and utilize 34% less average travel distance.

关键词： task allocation aggregation shortest path dual heterogeneous spatial crowdsourcing

来源：评论

学校读者我要写书评

暂无评论

Unsigned Road Incidents Detection Using Improved RESNET From Driver-View Images

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2025年第5期6卷 1203-1216页

作者： Li, Changping Wang, Bingshu Zheng, Jiangbin Zhang, Yongjun Chen, C.L. Philip Northwestern Polytechnical University School of Software Xi’an710129 China Shenzhen University Guangdong Provincial Key Laboratory of Intelligent Information Processing Shenzhen Key Laboratory of Media Security Shenzhen518060 China Guizhou University State Key Laboratory of Public Big Data College of Computer Science and Technology Guiyang550025 China South China University of Technology School of Computer Science and Engineering Guangzhou510641 China

Frequent road incidents cause significant physical harm and economic losses globally. The key to ensuring road safety lies in accurately perceiving surrounding road incidents. However, the highly dynamic nature of traffic introduces significant challenges, particularly in detecting sudden and temporary incidents. In this article, we propose a novel detection framework, multihead attention ResNet with dynamic bottleneck (DB_RESNET_MHA), to identify physical unsigned road incidents. Our approach introduces three key innovations. First, we develop a tailored data augmentation strategy to create images that closely mimic the complex variations found in real-world road environments. Second, we enhance model expressiveness by employing attention mechanisms to nonlinearly integrate convolutional kernels within the residual network. Furthermore, we refine the prediction head by applying spatially distinct attention weights, enabling the model to capture intricate correlations between different features more effectively. To demonstrate the effectiveness of our method, we create a dataset for unsigned road incidents (UnsignRI), comprising a total of 16 323 images that capture 12 distinct types of incidents. It stands out as the most comprehensive dataset in the field, encompassing a wide range of geographical features and incident categories. Experimental results show that DB_RESNET_MHA achieves an average accuracy of 96.2% and a f1-score of 0.955 across various categories of unsigned incidents, surpassing other models. © 2020 IEEE.

关键词： Highway accidents

来源：评论

学校读者我要写书评

暂无评论

New results on quantum boomerang attacks

引用

Quantum information Processing 2023年第4期22卷 1-27页

作者： Zou, Hongkai Zou, Jian Luo, Yiyuan College of Computer and Data Science Fuzhou University Fuzhou China Key Lab of Information Security of Network Systems Fuzhou University Fuzhou China School of Information Sciences and Technology Huizhou University Huizhou China

At SAC 2021, Frixons et al. proposed quantum boomerang attacks that can effectively recover the keys of block ciphers in the quantum setting. Based on their work, we further consider how to quantize the generic boomerang attacks proposed by Biham et al. at FSE 2002, so as to obtain more generic quantum boomerang attacks. Similar to Frixons et al.’s work, we only consider quantum key recovery attacks in the single-key setting. With the help of some famous quantum algorithms, this paper presents two methods to convert the attacks of Biham et al. into some new quantum key recovery attacks. In order to proof our methods, we apply our new ideas to attack Serpent-256 and ARIA-196. To sum up, for Serpent-256, we give valid 9-round and 10-round quantum key recovery attacks respectively. The quantum time complexity of 9-round and 10-round of Serpent-256 is 2 115.43 and 2 126.6 respectively. Furthermore, we show a valid quantum key attack on 6-round ARIA-196 which has a time complexity of 2 89.8 with negligible memory. The time complexity of the above quantum attacks are better than the corresponding classical attacks and quantum generic key recovery attack via Grover’s algorithm. © 2023, The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature.

关键词： Quantum theory

来源：评论

学校读者我要写书评

暂无评论

Graph Prompt Clustering

引用

IEEE Transactions on Pattern Analysis and Machine Intelligence 2025年第7期47卷 5794-5805页

作者： Chen, Man-Sheng Lai, Pei-Yuan Liao, De-Zhang Wang, Chang-Dong Lai, Jian-Huang Sun Yat-sen University School of Computer Science and Engineering Guangzhou510275 China Guangxi Zhuang Autonomous Region Information Center Guangxi Key Laboratory of Digital Infrastructure Nanning530219 China Guangdong Key Laboratory of Big Data Analysis and Processing Guangzhou510006 China Ministry of Education Key Laboratory of Machine Intelligence and Advanced Computing Beijing100816 China South China Technology Commercialization Center Guangzhou510006 China Guangdong University of Technology School of Information Engineering Guangzhou510006 China Guangdong Key Laboratory of Information Security Technology Guangzhou510033 China Ministry of Education Beijing100044 China

Due to the wide existence of unlab.led graph-structured data (e.g., molecular structures), the graph-level clustering has recently attracted increasing attention, whose goal is to divide the input graphs into several disjoint groups. However, the existing methods habitually focus on learning the graphs embeddings with different graph reguralizations, and seldom refer to the obvious differences in data distributions of distinct graph-level datasets. How to characteristically consider multiple graph-level datasets in a general well-designed model without prior knowledge is still challenging. In view of this, we propose a novel Graph Prompt Clustering (GPC) method. Within this model, there are two main modules, i.e., graph model pretraining as well as prompt and finetuning. In the graph model pretraining module, the graph model is pretrained by a selected source graph-level dataset with mutual information maximization and self-supervised clustering regularization. In the prompt and finetuning module, the network parameters of the pretrained graph model are frozen, and a groups of learnable prompt vectors assigned to each graph-level representation are trained for adapting different target graph-level datasets with various data distributions. Experimental results across six benchmark datasets demonstrate the impressive generalization capability and effectiveness of GPC compared with the state-of-the-art methods. © 1979-2012 IEEE.

关键词： Graph embeddings

来源：评论

学校读者我要写书评

暂无评论

Joint Optimization of Compression, Transmission and Computation for Cooperative Perception Aided Intelligent Vehicular Networks

引用

IEEE Transactions on Vehicular technology 2025年第5期74卷 8201-8214页

作者： Lu, Binbin Huang, Xumin Wu, Yuan Qian, Liping Zhou, Sheng Niyato, Dusit University of Macau State Key Laboratory of Internet of Things for Smart City Department of Computer and Information Science 999078 China Guangdong University of Technology School of Automation Guangzhou510006 China University of Macau State Key Lab of Internet of Things for Smart City 999078 China University of Macau State Key Laboratory of Internet of Things for Smart City Department of Computer Information Science 999078 China Zhuhai UM Science and Technology AQ2 Research Institute Zhuhai519301 China Zhejiang University of Technology Institute of Cyberspace Security Hangzhou310023 China Tsinghua University Beijing National Research Center for Information Science and Technology Department of Electronic Engineering Beijing100084 China Nanyang Technological University School of Computer Science and Engineering 639815 Singapore

Cooperative perception is a promising paradigm to tackle the perception limitations of a single intelligent vehicle (IV) to enhance the driving safety and efficiency in intelligent vehicular networks. However, the real-time transmission and computation-intensive fusion of raw sensing data raise new challenges for satisfying the stringent delay requirement of delay-sensitive applications. In this paper, we formulate a joint optimization problem of the cooperative IV selection, compression ratio selection, transmit power control, task offloading decision and computation allocation to minimize the end-to-end delay consisting of compression, transmission and computation. In particular, to meet the perception probability requirement, the quality and quantity-aware matching algorithm is proposed to optimize the cooperative IV selection. To guarantee the queue stability of offloading tasks, we develop the Lyapunov optimization algorithm to determine the upper bound of the offloading tasks and the corresponding optimal computation allocation. The Lyapunov aided deep reinforcement learning algorithm is further proposed to dynamically adjust the task offloading decision and compression ratio selection to minimize the end-to-end delay with the transmit power control being optimized by the successive convex approximation algorithm. Simulation results demonstrate that, compared to several benchmark algorithms, our proposed algorithm achieves the lowest end-to-end delay while guaranteeing the requirements on perception probability and queue stability effectively. © 1967-2012 IEEE.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

EPRFL:An Efficient Privacy-Preserving and Robust Federated Learning Scheme for Fog Computing

引用

China Communications 2025年第4期22卷 202-222页

作者： Ke Zhijie Xie Yong Syed Hamad Shirazi Li Haifeng School of Computer Technology and Application Qinghai UniversityXining 810016China School of Computer and Information Science Qinghai Institute of TechnologyXining 810016China Guangdong Key Laboratory of Blockchain Security Guangzhou UniversityGuangzhou 510006China Department of Information Technology Hazara UniversityBaffa 21110Pakistan Qinghai Provincial Key Laboratory of Big Data in Finance and Artificial Intelligence Application Technology Xining 810016China

Federated learning(FL)is a distributed machine learning paradigm that excels at preserving data privacy when using data from multiple *** combined with Fog Computing,FL offers enhanced capabilities for machine learning applications in the Internet of Things(IoT).However,implementing FL across large-scale distributed fog networks presents significant challenges in maintaining privacy,preventing collusion attacks,and ensuring robust data *** address these challenges,we propose an Efficient Privacy-preserving and Robust Federated Learning(EPRFL)scheme for fog computing ***,we first propose an efficient secure aggregation strategy based on the improved threshold homomorphic encryption algorithm,which is not only resistant to model inference and collusion attacks,but also robust to fog node ***,we design a dynamic gradient filtering method based on cosine similarity to further reduce the communication *** minimize training delays,we develop a dynamic task scheduling strategy based on comprehensive *** analysis demonstrates that EPRFL offers robust security and low *** experimental results indicate that EPRFL outperforms similar strategies in terms of privacy preserving,model performance,and resource efficiency.

关键词： federated learning fog computing internet of things privacy-preserving robustness

来源：评论

学校读者我要写书评

暂无评论

DarkSAM: Fooling Segment Anything Model to Segment Nothing 38

DarkSAM: Fooling Segment Anything Model to Segment Nothing

引用

38th Conference on Neural information Processing Systems, NeurIPS 2024

作者： Zhou, Ziqi Song, Yufei Li, Minghui Hu, Shengshan Wang, Xianlong Zhang, Leo Yu Yao, Dezhong Jin, Hai National Engineering Research Center for Big Data Technology and System China Services Computing Technology and System Lab China Cluster and Grid Computing Lab China Hubei Engineering Research Center on Big Data Security China Hubei Key Laboratory of Distributed System Security China School of Computer Science and Technology Huazhong University of Science and Technology China School of Cyber Science and Engineering Huazhong University of Science and Technology China School of Software Engineering Huazhong University of Science and Technology China School of Information and Communication Technology Griffith University Australia

Segment Anything Model (SAM) has recently gained much attention for its outstanding generalization to unseen data and tasks. Despite its promising prospect, the vulnerabilities of SAM, especially to universal adversarial perturbation (UAP) have not been thoroughly investigated yet. In this paper, we propose DarkSAM, the first prompt-free universal attack framework against SAM, including a semantic decoupling-based spatial attack and a texture distortion-based frequency attack. We first divide the output of SAM into foreground and background. Then, we design a shadow target strategy to obtain the semantic blueprint of the image as the attack target. DarkSAM is dedicated to fooling SAM by extracting and destroying crucial object features from images in both spatial and frequency domains. In the spatial domain, we disrupt the semantics of both the foreground and background in the image to confuse SAM. In the frequency domain, we further enhance the attack effectiveness by distorting the high-frequency components (i.e., texture information) of the image. Consequently, with a single UAP, DarkSAM renders SAM incapable of segmenting objects across diverse images with varying prompts. Experimental results on four datasets for SAM and its two variant models demonstrate the powerful attack capability and transferability of DarkSAM. Our codes are availab.e at: https://***/CGCL-codes/DarkSAM. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Contrastive Learning for Robust Android Malware Familial Classification

引用

IEEE Transactions on Dependable and Secure Computing 2022年 1-14页

作者： Wu, Yueming Dou, Shihan Zou, Deqing Yang, Wei Qiang, Weizhong Jin, Hai National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Hubei Engineering Research Center on Big Data Security School of Cyber Science and Engineering Huazhong University of Science and Technology Wuhan China Shanghai Key Laboratory of Intelligent Information Processing School of Computer Science Fudan University Shanghai China University of Texas at Dallas Dallas USA National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China

Due to its open-source nature, Android operating system has been the main target of attackers to exploit. Malware creators always perform different code obfuscations on their apps to hide malicious activities. Features extracted from these obfuscated samples through program analysis contain many useless and disguised features, which leads to many false negatives. To address the issue, in this paper, we demonstrate that obfuscation-resilient malware family analysis can be achieved through contrastive learning. The key insight behind our analysis is that contrastive learning can be used to reduce the difference introduced by obfuscation while amplifying the difference between malware and other types of malware. Based on the proposed analysis, we design a system that can achieve robust and interpretable classification of Android malware. To achieve robust classification, we perform contrastive learning on malware samples to learn an encoder that can automatically extract robust features from malware samples. To achieve interpretable classification, we transform the function call graph of a sample into an image by centrality analysis. Then the corresponding heatmaps can be obtained by visualization techniques. These heatmaps can help users understand why the malware is classified as this family. We implement IFDroid and perform extensive evaluations on two datasets. Experimental results show that IFDroid is superior to state-of-the-art Android malware familial classification systems. Moreover, IFDroid is capable of maintaining a 98.4% F1 on classifying 69,421 obfuscated malware samples. IEEE

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Frozen-DETR: Enhancing DETR with Image Understanding from Frozen Foundation Models 38

Frozen-DETR: Enhancing DETR with Image Understanding from Fr...

引用

38th Conference on Neural information Processing Systems, NeurIPS 2024

作者： Fu, Shenghao Yan, Junkai Yang, Qize Wei, Xihan Xie, Xiaohua Zheng, Wei-Shi School of Computer Science and Engineering Sun Yat-sen University China Peng Cheng Laboratory Shenzhen518055 China Tongyi Lab Alibaba Group China Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education China Guangdong Province Key Laboratory of Information Security Technology China Guangdong Guangzhou510555 China

Recent vision foundation models can extract universal representations and show impressive abilities in various tasks. However, their application on object detection is largely overlooked, especially without fine-tuning them. In this work, we show that frozen foundation models can be a versatile feature enhancer, even though they are not pre-trained for object detection. Specifically, we explore directly transferring the high-level image understanding of foundation models to detectors in the following two ways. First, the class token in foundation models provides an in-depth understanding of the complex scene, which facilitates decoding object queries in the detector's decoder by providing a compact context. Additionally, the patch tokens in foundation models can enrich the features in the detector's encoder by providing semantic details. Utilizing frozen foundation models as plug-and-play modules rather than the commonly used backbone can significantly enhance the detector's performance while preventing the problems caused by the architecture discrepancy between the detector's backbone and the foundation model. With such a novel paradigm, we boost the SOTA query-based detector DINO from 49.0% AP to 51.9% AP (+2.9% AP) and further to 53.8% AP (+4.8% AP) by integrating one or two foundation models respectively, on the COCO validation set after training for 12 epochs with R50 as the detector's backbone. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

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

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

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：