检索结果-内蒙古大学图书馆

IEEE Transactions on Artificial Intelligence 2025年第2期6卷 405-415页

作者： Cai, Zinuo Ma, Rongbo Fu, Yicheng Zhang, Weishan Ma, Ruhui Guan, Haibing Shanghai Jiao Tong University School of Electronic Information and Electrical Engineering Shanghai200240 China Qingdao Institute of Software College of Computer Science and Technology Qingdao266555 China

In recent years, the emergence of large-language models (LLMs) has profoundly transformed our production and lifestyle. These models have shown tremendous potential in fields, such as natural language processing, speech recognition, and recommendation systems, and are increasingly playing crucial roles in applications such as human-computer interaction and intelligent customer service. Efficient inference solutions for LLMs in data centers have been extensively researched, with a focus on meeting users' quality of service requirements. In this article, we focus on two additional requirements that responsible LLM inference should meet under QoS conditions: security throughout the model execution process and low maintenance requirements for the inference system. Therefore, we propose LLMaaS, a trusted model inference platform based on a serverless computing platform aimed at providing inference as a service for LLMs. First, we design a trusted serverless computing platform based on software guard extension (SGX), which includes distributed identity verification and SGX device plugins to ensure the security and trustworthiness of the inference process. Additionally, to reduce the maintenance requirements of the system, we enhance the SGX-based deep learning computing framework, including replacing PyTorch and using a greedy algorithm for graph partitioning. We conduct tests on four typical large models, and the experimental results demonstrate that, with minimal overhead and user code modifications, we can ensure the security of model execution. © 2020 IEEE.

关键词： Quality of service

来源：评论

学校读者我要写书评

暂无评论

DA-FR-UNet: A Dual Attention-Enhanced Approach for Retinal Vessel Image Segmentation

IAENG International Journal of Computer Science

引用

IAENG International Journal of computer Science 2025年第5期52卷 1325-1335页

作者： Li, Yuanyuan Wu, Jiansheng Liu, Yang School of Computer Science and Software Engineering University of Science and Technology Liaoning Anshan114051 China University of Science and Technology Liaoning Anshan114051 China

Retinal vessel image segmentation is crucial for the early diagnosis and treatment of ophthalmic diseases. However, accurate segmentation remains challenging due to the complexity and diversity of retinal images, as well as the intricate nature of vascular structures. Existing methods often underperform when dealing with fine vessels and complex backgrounds, limiting their effectiveness in clinical applications. This study proposes a novel retinal vessel image segmentation method, Dual Attention FR-UNet (DA-FR-UNet), which incorporates two innovative attention mechanisms: Enhanced Channel Prior Attention (ECPA) and Hard Triplet Attention (HTA). The ECPA module enhances the model's ability to recognize fine blood vessels by integrating global average pooling and global maximum pooling to effectively capture multi-scale features. In contrast, the HTA module improves the model's understanding of complex vascular structures by simultaneously considering attention information across three dimensions: channel, spatial height, and width. Additionally, the study introduces the testtime augmentation (TTA) technique during the experimental phase to enhance the model's generalizability and robustness. To validate the performance of the DA-FR-UNet model in retinal vessel image segmentation tasks, a series of comparative experiments were conducted on the FIVES and RVD datasets. The experimental results demonstrate that the DA-FR-UNet model, combining the dual-attention mechanism and the TTA strategy, outperforms existing state-of-the-art methods across several key evaluation metrics. On the FIVES dataset, DAFR- UNet achieves a sensitivity of 88.42%, an IoU value of 82.27%, and a Dice coefficient of 89.72%. On the RVD dataset, these metrics are 65.46%, 44.53%, and 61.43%, respectively. These findings highlight the model's significant advantages in processing fine vessels and complex backgrounds, improving its sensitivity and robustness in blood vessel image segmentation. © (2025)

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

An Efficient Dialogue Policy Agent with Model-Based Causal Reinforcement Learning 31

An Efficient Dialogue Policy Agent with Model-Based Causal R...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Xu, Kai Wang, Zhenyu Zhao, Yangyang Fang, Bopeng School of Software Engineering South China University of Technology Guangdong China Department of Computer and Communication Engineering Changsha University of Science and Technology Changsha China School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu China

ISBN: (纸本)9798891761964

Dialogue policy trains an agent to select dialogue actions frequently implemented via deep reinforcement learning (DRL). The model-based reinforcement methods built a world model to generate simulated data to alleviate the sample inefficiency. However, traditional world model methods merely consider one-step dialogues, leading to an inaccurate environmental simulation. Furthermore, different users may have different intention preferences, while most existing studies lack consideration of the intention-preferences causal relationship. This paper proposes a novel framework for dialogue policy learning named MCA, implemented through model-based reinforcement learning with automatically constructed causal chains. The MCA model utilizes an autoregressive Transformer to model dialogue trajectories, enabling a more accurate simulation of the environment. Additionally, it constructs a causal chains module that outputs latent preference distributions for intention-action pairs, thereby elucidating the relationship between user intentions and agent actions. The experimental results show that MCA can achieve state-of-the-art performances on three dialogue datasets over the compared dialogue agents, highlighting its effectiveness and robustness. © 2025 Association for Computational Linguistics.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Training Latency Minimization for Model-Splitting Allowed Federated Edge Learning

引用

IEEE Transactions on Network Science and engineering 2025年第3期12卷 2081-2092页

作者： Wen, Yao Zhang, Guopeng Wang, Kezhi Yang, Kun China University of Mining and Technology School of Computer Science and Technology Xuzhou221116 China Brunel University London Department of Computer Science MiddlesexUB8 3PH United Kingdom Nanjing University State Key Laboratory of Novel Software Technology Nanjing210008 China School of Intelligent Software and Engineering Nanjing University Suzhou Campus Suzhou215163 China

To alleviate the shortage of computing power faced by clients in training deep neural networks (DNNs) using federated learning (FL), we leverage the edge computing and split learning to propose a model-splitting allowed FL (SFL) framework, with the aim to minimize the training latency without loss of test accuracy. Under the synchronized global update setting, the latency to complete a round of global training is determined by the maximum latency for the clients to complete a local training session. Therefore, the training latency minimization problem (TLMP) is modelled as a minimizing-maximum problem. To solve this mixed integer nonlinear programming problem, we first propose a regression method to fit the quantitative-relationship between the cut-layer and other parameters of an AI-model, and thus, transform the TLMP into a continuous problem. Considering that the two subproblems involved in the TLMP, namely, the cut-layer selection problem for the clients and the computing resource allocation problem for the parameter-server are relative independence, an alternate-optimization-based algorithm with polynomial time complexity is developed to obtain a high-quality solution to the TLMP. Extensive experiments are performed on a popular DNN-model EfficientNetV2 using dataset MNIST, and the results verify the validity and improved performance of the proposed SFL framework. © 2013 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Message Control for Blockchain Rewriting

引用

IEEE Transactions on Dependable and Secure Computing 2025年第3期22卷 1864-1876页

作者： Li, Yingjiu Sengupta, Binanda Tian, Yangguang Yuan, Jiaming Yuen, Tsz Hon School of Computer and Information Science University of Oregon EugeneOR United States School of Computer Science and Electronic Engineering University of Surrey Guilford Surrey United Kingdom Indian Institute of Engineering Science and Technology Shibpur India Department of Software Systems and Cybersecurity Monash University MelbourneVIC Australia

Blockchain rewriting is necessary for modifying illegal or messages included in blockchain transactions, while maintaining the consistency of subsequent blocks in the blockchain. However, arbitrary blockchain rewriting is not desirable as it defeats the purpose of blockchain rewriting. In this work, we propose a new security primitive named message-controlled chameleon hash (MCH) and apply it for message control in blockchain rewriting to ensure that no unspecified messages are generated from blockchain rewriting. The proposed MCH enables permitted parties to select candidate messages from designated message sets for blockchain rewriting at the message level. Our evaluation shows that the proposed MCH can be easily integrated into both permissioned and permissionless blockchains. © 2024 IEEE.

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

IPT-ILR: Image Pyramid Transformer Coupled With Information Loss Regularization for All-in-One Image Restoration

引用

IEEE Transactions on Circuits and Systems for Video technology 2025年第5期35卷 4341-4356页

作者： Yang, Sai Hu, Bin Liu, Fan Wu, Xiaoxin Ding, Weiping Zhou, Jun Nantong University School of Electrical Engineering and Automation Nantong226019 China Nantong University School of Artificial Intelligence and Computer Science Nantong226019 China Hohai University College of Computer Science and Software Engineering Nanjing210098 China Griffith University School of Information and Communication Technology NathanQLD4111 Australia

All-in-one image restoration has recently developed to be a new research trend in the low-level computer vision field, aiming to tackle multiple image degradation types simultaneously in a unified model. As a typical multi-task learning, existing approaches focus on modeling either the specificity or commonality among different image restoration tasks. To exploit the unique strengths of both worlds, we propose a method of Image Pyramid Transformer coupled with Information Loss Regularization (IPT-ILR), in which the multi-scale architecture structure can excavate more information for multiple restoration tasks concurrently, while the learning strategy can identify the difference among multiple restoration tasks depending on the degree of information loss in each restoration task. Specifically, it first establishes a new Image Pyramid Transformer Network (IPT-Network) to accommodate multiple image restoration tasks. Given original degraded images, the IPT-Network exploits the image pyramid technique to establish a series of images with different scales, which are then restored by transformer-like auto-encoders. Moreover, the restored image on a low-level scale is referenced to assist restoring the degraded image on a high-level scale. Next, Information Loss Regularization (ILR) is presented to optimize the IPT-Network. ILR calculates the average distance between degraded images and their clean counterparts as the weights, which automatically implement different penalties for different image restoration tasks, thus avoiding the short-cut phenomenon for the easy task while encouraging the hard task. Extensive experiments have been conducted with 6 image restoration tasks in the all-in-one setting. The results show our method performs favorably against numerous state-of-the-art methods across most tasks, including image denoising, image deblurring, image dehazing, image deraining, image desnowing, as well as low-light enhancement. © 1991-2012 IEEE.

关键词： Restoration

来源：评论

学校读者我要写书评

暂无评论

Research On Fatigue Driving Detection Based On Deep Learning

引用

engineering Letters 2025年第2期33卷 348-356页

作者： Zhou, Huinan Dai, Hong University of Science and Technology Liaoning Liaoning AnshanCO 114051 China School of Computer Science and Software Engineering University of Science and Technology Liaoning Liaoning AnshanCO 114051 China

Fatigue driving is one of the main causes of traffic accidents. Effective fatigue driving detection technology can reduce traffic accidents caused by fatigue driving. Traditional fatigue driving detection methods usually use a two-stage detection method, which has the problems of low efficiency and poor accuracy. To solve these problems, this paper designs a single-stage fatigue driving detection model based on separable convolution Road Safety Net (RSNet), which extracts the features in the driver's image or video and classifies these features to determine whether the driver is in an abnormal driving state. RSNet borrows the structure of Swin Transformer and adopts a hierarchical design, which makes the model better able to extract features and classify. Each level is composed of a GDS Layer and multiple RS Blocks. To solve the problem of low efficiency of traditional methods, we use separable convolution in RS Block to replace the traditional convolutional neural network, which reduces the complexity of the model and effectively improves the computational efficiency of the model. Experimental results show that the Top-1 and Top-5 Accuracy of RSNet on the HMDB51 dataset reaches 85 % and 97.35 %, respectively, and the delay is 5.97 ms. Compared with ResNet, EfficientNet, and Swin Transformer, RSNet has higher accuracy and efficiency. In addition, this paper takes the YOLACT algorithm as an example, selects different backbone networks for object detection, and compares the indicators of mean Average Precision (mAP) and Frames Per Second (FPS). Experiments show that RSNet performs better when used as the backbone network for object detection, with mAP reaching 32.4 % and FPS of 39.6 ms. It indicates that RSNet performs well in action recognition tasks. © 2025, International Association of Engineers. All rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Robust copy-move detection and localization of digital audio based CFCC feature

引用

Multimedia Tools and Applications 2025年第11期84卷 9573-9589页

作者： Wang, Dongyu Li, Xiaojie Shi, Canghong Niu, Xianhua Xiong, Ling Wu, Hanzhou Qian, Qing Qi, Chao School of Computer and Software Engineering Xihua University Chengdu610039 China The College of Computer Science Chengdu University of Information Technology Chengdu610225 China School of Information Shanghai University Shanghai200444 China School of Information Guizhou University of Finance and Economics Guiyang550000 China

Copy-move forgery is a common audio tampering technique in which users copy the contents of one speech and paste them into another region of the same speech signal, thus achieving the effect of tampering with the semantics. To verify the authenticity of the audio, this paper proposes a method to detect and localize audio copy-move forgery based on the cochlear filter of cochlear filter cepstral coefficients (CFCC) feature. The pitch tracking algorithm is used to distinguish the voiced and unvoiced segments in the audio, and then the CFCC coefficients are extracted for each voiced segment. The CFCC feature simulates the entire transmission process of signals in the cochlear basilar membrane using wavelet transformation. Finally, we use Pearson correlation coefficients (PCCs) and dynamic time warping (DTW) in combination to compare the similarity of voiced segments, accurately determining the tampered locations in the audio through threshold judgment. Through extensive experiments on relevant datasets, this algorithm achieves a precision rate of 98.39% and a recall rate of 98.00% in detecting audio without post-processing. Even when detecting audio that has undergone different post-processing, the precision and recall rates remain above 90%. Compared to existing methods, this approach not only achieves precise localization of replicated segments but also demonstrates superior experimental results. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Runtime Verification of Neural-Symbolic Systems 10th

Runtime Verification of Neural-Symbolic Systems

引用

10th International Symposium on Dependable software engineering: Theories, Tools and Applications, SETTA 2024

作者： Deng, Shaojun Liu, Wanwei Zhang, Miaomiao College of Computer Science and Technology National University of Defense Technology Changsha China School of Software Engineering Tongji Uniersity Shanghai China

ISBN: (纸本)9789819606016

Neural-symbolic systems (NSSs), which are typically cyber-physical systems integrated with artificial intelligence modules, have received much attention in both academic and industrial fields. However, thorough verification (such as model checking) of such systems in general leads to prohibitively high costs due to the scale of the state space, and this makes it barely accomplishable within desired time-span. Consequently, light-weight verification techniques are introduced to deal with such cases, e.g., runtime verification (RV, for short). In this paper, we investigate an RV framework for NSSs against signal temporal logic (STL) properties. To guarantee that the on-line monitoring could be accomplished in time, we utilize Euler-prediction to sample the system under scrutiny at discrete time-steps. Our framework supports both qualitative and quantitative on-line monitoring of STL specifications, and thus it terminates whenever a property is satisfied and/or violated by a finite prefix of the trajectory, and it can also provide the bounds of the quantitative satisfaction/violation w.r.t. the remaining execution. Notably, for the so-called piece-wise linear NSSs against affine STL formulas, our approach is exact. On top of it, a prototype tool is implemented, and is experimentally evaluated. The results demonstrate the feasibility of the presented runtime verification approach. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Model checking

来源：评论

学校读者我要写书评

暂无评论

BiEfficient: Bidirectionally Prompting Vision-Language Models for Parameter-Efficient Video Recognition 17th

BiEfficient: Bidirectionally Prompting Vision-Language Model...

引用

17th Asian Conference on computer Vision, ACCV 2024

作者： He, Haichen Liu, Weibin Xing, Weiwei School of Computer Science and Technology Beijing Jiaotong University Beijing China School of Software Engineering Beijing Jiaotong University Beijing China

ISBN: (纸本)9789819609079

Vision-language models (VLMs) pre-trained on large-scale image-text pairs have shown great success in various image tasks. However, how to efficiently transfer such powerful VLMs into video domain is still an open problem. Given that full finetuning VLMs for video tasks could be computationally expensive, recent studies turn their focus on parameter-efficient finetuning (PEFT). The great potential of VLMs lies in leveraging the bidirectional semantic connections between the two modalities of vision and language. Nevertheless, most current PEFT methods use the vision-only framework and usually ignore the semantic connections between vision and language. In this paper, we propose a novel method called BiEfficient, which use bidirectional prompting schemes to efficiently transfer the VLM to video recognition task with a small number of tunable parameters: 1) Vision-to-Language: we propose two prompt mechanisms, Pre-Prompt and Post-Prompt, which act before and after the text encoder respectively to generate discriminative video-level text representation for each input video. 2) Language-to-Vision: we propose Word-Guided Visual-Prompt, which enhances the temporal modeling of videos using textual knowledge in an almost parameter-free manner. Experiments on Kinetics-400, UCF-101, HMDB-51 demonstrate that the proposed method can achieve comparable or even better performance to the full finetuning methods with much fewer tunable parameters across closed-set and zero-shot video recognition benchmarks. Code is available here: https://***/WbLiuBJTULab/BiEfficient. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Semantics

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：