检索结果-内蒙古大学图书馆

Teacher Probability Reconstruction based knowledge distillation within intelligent network compression

International Journal of Intelligent Networks 2025年 6卷 47-56页

作者： Chen, Han Teng, Xuyang Su, Jiajie Li, Chunhao Hu, Chang Han, Meng School of Communication Engineering Hangzhou Dianzi University Hangzhou China Department of Computer Science National University of Singapore Singapore College of Computer Science and Technology Zhejiang University Hangzhou China

In the optimization of intelligent network architecture, limited resources at each node, including edge computing devices, have posed challenges for deploying large models in performance-demanding scenarios. Knowledge distillation serves as a model compression method that extracts knowledge from a large-scale teacher model and transfers it to a more lightweight student model. Previous knowledge distillation methods mainly focus on the intermediate layers of the network. However, due to privacy protection regulations that limit data sharing and access as well as computational efficiency requirements in practical scenarios, feature-based distillation encounters challenges in practical applications. We start with logit-based distillation to address these issues, enabling students to learn more representative knowledge from the teacher's output probability distribution. Due to the structural limitations of the teacher network such as insufficient depth or width, and potential issues in the training data like noise and imbalance, the output probability distribution contains many errors. Therefore, we propose a knowledge distillation method that improves the student by correcting errors in the teacher model. Nevertheless, teacher's errors not only bring mistakes to students but also give students greater subjectivity, enabling them to break free from the limitations of the teacher. We also retain the teacher's thinking to prevent students from becoming biased on the remaining (non-target) categories while correcting teacher errors for students. Extensive experiments demonstrate that our method achieves competitive performance on multiple benchmarks without extra parameters. © 2025 The Authors

关键词： Normal distribution

来源：评论

学校读者我要写书评

暂无评论

Fake News Detection Based on Cross-Modal Ambiguity Computation and Multi-Scale Feature Fusion

引用

computers, Materials & Continua 2025年第5期83卷 2659-2675页

作者： Jianxiang Cao Jinyang Wu Wenqian Shang Chunhua Wang Kang Song Tong Yi Jiajun Cai Haibin Zhu State Key Laboratory of Media Convergence and Communication Communication University of ChinaBeijing100024China School of Computer Science and Engineering Guangxi Normal UniversityGuilin541004China Department of Computer Science Nipissing UniversityNorth BayON P1B 8L7Canada

With the rapid growth of socialmedia,the spread of fake news has become a growing problem,misleading the public and causing significant *** social media content is often composed of both images and text,the use of multimodal approaches for fake news detection has gained significant *** solve the problems existing in previous multi-modal fake news detection algorithms,such as insufficient feature extraction and insufficient use of semantic relations between modes,this paper proposes the MFFFND-Co(Multimodal Feature Fusion Fake News Detection with Co-Attention Block)***,the model deeply explores the textual content,image content,and frequency domain ***,it employs a Co-Attention mechanism for cross-modal ***,a semantic consistency detectionmodule is designed to quantify semantic deviations,thereby enhancing the performance of fake news *** verified on two commonly used datasets,Twitter and Weibo,the model achieved F1 scores of 90.0% and 94.0%,respectively,significantly outperforming the pre-modified MFFFND(Multimodal Feature Fusion Fake News Detection with Attention Block)model and surpassing other baseline *** improves the accuracy of detecting fake information in artificial intelligence detection and engineering software detection.

关键词： Fake news detection multimodal cross-modal ambiguity computation multi-scale feature fusion

来源：评论

学校读者我要写书评

暂无评论

Multi-scale feature fusion for facial expression recognition

引用

Neural Computing and Applications 2025年 1-22页

作者： Guo, Jiatao Peng, Junjie Huang, Yansong Chen, Gan Cai, Zesu Tan, Shuhua School of Computer Engineering and Science Shanghai University Shanghai China Shanghai Institute for Advanced Communication and Data Science Shanghai University Shanghai China School of Computer Science and Technology Harbin Institute of Technology Harbin China National Engineering Laboratory for Logistics Information Technology YTO Express Co. Ltd. Shanghai China

As an important computer vision task that can be used in many areas, facial expression recognition (FER) has been widely studied which much progress has been obtained especially when deep learning (DL) approaches have been introduced in FER. However, existing methods are restricted to extracting localized information or utilizing attention mechanism to focus on the region most relevant to the emotion, possibly resulting in errors in facial expression interpretation. Moreover, the inconsistency of image sizes in the wild FER datasets further increases the difficulty of the FER task. To address these issues, we propose a multi-scale feature fusion (MSFF) for FER. Specifically, our method first extracts features from facial images using multi-scale feature extraction to obtain rich multi-dimensional information. Meanwhile, extensive feature extraction at different scales is achieved by expanding the receptive field of the model, which provides the basis for more detailed analysis of facial expressions. Eventually, features at different scales are fused to enhance their synergistic effect. Subsequently, the adaptive extraction of comprehensive and valuable information from the fused features is performed, optimizing the performance of expression recognition. Extensive experimental results demonstrate that the proposed method outperforms the state-of-the-art methods on in-the-wild datasets. It is promising for FER in realistic and complex scenarios. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2025.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

A Distributed Routing Algorithm for LEO Satellite Networks: A Multi-Agent Transformer-MIX Learning Approach

引用

IEEE Internet of Things Journal 2025年第11期12卷 15748-15763页

作者： Chen, Xiao Ji, Zhe Wu, Sheng Jia, Haoge Xiao, Ailing Jiang, Chunxiao Beijing University of Posts and Telecommunications School of Information and Communication Engineering Beijing100876 China Tsinghua University Tsinghua Space Center Beijing National Research Center for Information Science and Technology Beijing100084 China

As a complement to terrestrial networks, low Earth orbit (LEO) satellite networks are promising to provide ubiquitous and continuous services. To accommodate the dynamic topology of LEO satellite networks and increasing traffic demands, this paper proposes a distributed routing approach to optimize the end-to-end delay relying on the multi-agent deep reinforcement learning(MADRL), where each satellite node is deployed with an autonomous agent and makes its real-time next-hop decisions independently with local observation information. To promote the inner cooperation between decentralized agents, a centralized training scheme with a unified load-balancing reward is utilized by adopting a novel multi-agent Transformer-MIX architecture. Moreover, to derive a better decision for each agent, we design an attention-involved agent network to capture more hidden information, and a Transformer-based parameter recurrent mechanism to generate the joint action-value function is used to enhance a more stable training. The simulation results indicate that our proposed scheme achieves faster convergence and demonstrates superior performance across several key performance metrics compared to existing benchmark schemes. Specifically, when the inter-satellite link (ISL) failure rate in the network reaches 18%, our scheme achieves a reduction in end-to-end delay by 13.6% and an increase in packet successful delivery rate by 5.4% compared to the benchmark schemes. © 2014 IEEE.

关键词： Satellite links

来源：评论

学校读者我要写书评

暂无评论

Real-Time Road Damage Detection Using an Optimized YOLOv9s-Fusion in IoT Infrastructure

引用

IEEE Internet of Things Journal 2025年第11期12卷 17649-17660页

作者： Waseem Khan, Muhammad Obaidat, Mohammad S. Mahmood, Khalid Sadoun, Balqies Sanaullah Badar, Hafiz Muhammad Gao, Wu Center for Bio-inspired System-on-Chip School of Computer Science Northwestern Polytechnical University Xi’an China Sahiwal Campus Pakistan The King Abdullah II School of Information Technology The University of Jordan Amman11942 Jordan The School of Communication and Computing Engineering University of Science and Technology Beijing Beijing100083 China The Amity School of Engineering and Technology Amity University Uttar Pradesh Noida201301 India The Graduate School of Intelligent Data Science National Yunlin University of Science and Technology Douliu Yunlin64002 Taiwan School of Computer and Information Engineering Henan University Kaifeng475001 China The College of Engineering Al-Balqa’ Applied University Al-Salt19117 Jordan

In IoT-enabled smart infrastructure, accurate and real-time road damage detection is crucial for enhancing road safety and optimizing maintenance processes. However, detecting road damage in complex and dynamic environments presents significant challenges, such as varying lighting conditions, diverse damage types, and the need for fast processing to enable real-time decision-making. This study introduces an advanced approach utilizing the YOLOv9s-Fusion model to overcome these challenges. Leveraging the RDD2022 dataset, which comprises 1,976 annotated images of road damage from China, we employ comprehensive data preprocessing to create optimal conditions for model training. The YOLOv9s-Fusion model integrates innovative features, including a Transformer-based auxiliary module and enhanced feature extraction layers, specifically designed to detect fine-grained damage patterns accurately. Experimental results demonstrate that the model outperforms existing approaches, achieving notable improvements in mean Average Precision (mAP) and F1-score. Ablation studies further validate the impact of our modifications, highlighting the model’s robustness in real-time detection across diverse conditions. This IoT-centric approach sets a new standard for autonomous road damage detection, significantly advancing vehicle navigation and smart infrastructure management capabilities. © 2014 IEEE.

关键词： Intelligent systems

来源：评论

学校读者我要写书评

暂无评论

Generative Replay and Multi-steps Knowledge Distillation in Class-Incremental Learning 24

Generative Replay and Multi-steps Knowledge Distillation in ...

引用

3rd International Conference on Artificial Intelligence and Intelligent Information Processing, AIIIP 2024

作者： Meng, Yicheng Ping, Jinming Shi, Jingye Zhi, Ruicong School of Computer and Communication Engineering University of Science and Technology Beijing Beijing China Beijing Key Laboratory of Security and Privacy in Intelligent Transportation Beijing Jiaotong University Beijing China

ISBN: (纸本)9798400707308

Data-free generative modeling (GM) methods have gained much attention because pseudo data avoids security and privacy issues, and no amount limitation for storing real data in Class-Incremental Learning (CIL). However, GM-based methods have three drawbacks: un stable quality of generated image, semantic sifting, and domain bias for knowledge and distillation. To tackle with these problems, we propose an effective two-stage training method based on generative replay distillation, namely GA-RCT in this paper. It mainly consists of three modules. The Generative Adaptive Replay (GAR) module in stage one obtainsa well-trained generator for every task, then freezes generator for generating stable images, and trains feature extractor with mixture of real data and generated data. The Global Average Class (GAC) module is designed to reserve class prototypes to reduce semantic drift in the training stage. And the Generative Auxiliary Teaching (GAT) module in stage two performs generative distillation in dual-teacher to student way, for solving domain bias. Experiments on four datasets have shown that GA-RCT performs better than other generative replay methods. © 2024 Copyright held by the owner/author(s).

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

A Locality-Sensitive-Hashing-Based Collaborative Recommendation Method for Responsible AI-Driven Recommender Systems

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2025年第2期6卷 393-404页

作者： Lin, Wenmin Zhou, Xinyi Sun, Lu Qi, Lianyong Tsai, Sang-Bing Yang, Yihong Liu, Hanwen Kou, Huaizhen Kong, Lingzhen Hangzhou Normal University Alibaba Business School Hangzhou311121 China College of Computer Science and Technology Qingdao266580 China International Engineering and Technology Institute Hong Kong999077 Hong Kong China University of Geosciences School of Information Engineering Beijing100083 China Nanjing University of Science and Technology School of Computer Science and Engineering Nanjing210094 China

As one of the most representative recommendation solutions, traditional collaborative filtering (CF) models typically have limitations in dealing with large-scale, sparse data to capture complex relationships between users and items. The rise of artificial intelligence (AI) provides powerful tools such as deep neural networks to overcome the data sparsity issue with typical CF models. Existing works on AI-driven recommender systems are focusing on improving recommendation accuracy by extracting rich user/item features from multisource data with deep learning tools to build more accurate user preference model. How to ensure the responsibility of AI-driven recommender systems is still a big challenge. On one hand, AI-driven recommender systems need to access raw data such as user profiles to train recommendation model, which face the risk of leaking user privacy information. On the other hand, the ever-increasing volume of user/item interaction records raises the efficiency challenge to provide recommendation results in a real-time manner. In view of those observations, we propose a collaborative recommendation method by adopting the locality-sensitive-hashing (LSH) technique (i.e., DisRecLSH), which aims to achieve the goal of privacy protection and calculation efficiency for AI-driven recommender systems. More specifically, we split DisRecLSH into three phases: 1) offline feature extraction: deep neural networks are applied to extract user/item features from multisource user/item data on each local dataset;2) offline user index building: LSH technique is adopted to map user features to hash codes to quickly recall correlated similar users for a given target user;and 3) online top-N items calculation: with similar users selected by phase 2, top-N items are calculated based on the ranking of predicted user-item rating score. Finally, extensive experiments are conducted on three public datasets to evaluate the efficiency of our proposal. © 2020 IEEE.

关键词： Recommender systems

来源：评论

学校读者我要写书评

暂无评论

Noise-Mitigated Variational Quantum Eigensolver with Pre-training and Zero-Noise Extrapolation

Noise-Mitigated Variational Quantum Eigensolver with Pre-tra...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Sun, Wanqi Xu, Jungang Duan, Chenghua School of Electronic Electrical and Communication Engineering University of Chinese Academy of Sciences Beijing China School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China

ISBN: (纸本)9798350368741

As a hybrid quantum-classical algorithm, the variational quantum eigensolver is widely applied in quantum chemistry simulations, especially in computing the electronic structure of complex molecular systems. However, on existing noisy intermediate-scale quantum devices, some factors such as quantum decoherence, measurement errors, and gate operation imprecisions are unavoidable. To overcome these challenges, this study proposes an efficient noise-mitigating variational quantum eigensolver for accurate computation of molecular ground state energies in noisy environments. We design the quantum circuit with reference to the structure of matrix product states and utilize it to pre-train the circuit parameters, which ensures circuit stability and mitigates fluctuations caused by initialization. We also employ zero-noise extrapolation to mitigate quantum noise and combine it with neural networks to improve the accuracy of the noise-fitting function, which significantly eliminates noise interference. Furthermore, we implement an intelligent grouping strategy for measuring Hamiltonian Pauli strings, which not only reduces measurement errors but also improves sampling efficiency. We perform numerical simulations to solve the ground state energy of the H4 molecule by using MindSpore Quantum framework, and the results demonstrate that our algorithm can constrain noise errors within the range of O(10-2) ∼ O(10-1), outperforming mainstream variational quantum eigensolvers. This work provides a new strategy for high-precision quantum chemistry calculations on near-term noisy quantum hardware. © 2025 IEEE.

关键词： Error mitigation Matrix product states Variational quantum eigensolver Zero-noise extrapolation

来源：评论

学校读者我要写书评

暂无评论

Multi-IRS-Aided Secure communication in UAV-MEC Networks

引用

IEEE Transactions on Vehicular technology 2025年第5期74卷 7327-7338页

作者： Gao, Yuan Wang, Zhenyu Zhang, Yu Lu, Weidang Tang, Jie Zhao, Nan Gao, Feifei Academy of Military Science of the PLA Beijing China Zhejiang University of Technology College of Information Engineering Hangzhou310023 China Dalian University of Technology School of Information and Communication Engineering Dalian116024 China Tsinghua National Laboratory for Information Science and Technology Beijing100084 China

With the merits of high mobility and easy deployment, mounting mobile edge computing (MEC) servers on unmanned aerial vehicles(UAVs) can efficiently fulfill the task offloading of ground users (GUs) over a large area. Nevertheless, data security is a challenging issue for the computation offloading in UAV-MEC networks, especially when there exist flying eavesdroppers. An intelligent reflecting surface (IRS) assisted secure communication scheme for a UAV-MEC network is proposed in this paper, wherein multiple IRSs are utilized to help the secure computation offloading from GUs against a UAV eavesdropper. Our aim is to maximize the secure computation capacity through the joint optimization of the IRS phase-shift, allocation of communication and computing resources and trajectory of UAV. We firstly solve the problem under a fixed UAV trajectory by alternating optimization to obtain the resource allocation and IRS phase-shift, wherein Dinkebach and Taylor expansion methods are used to transform the subproblems into tractable forms. Then, by adopting the proximal policy optimization, a joint optimization approach which further incorporates the UAV trajectory optimization is proposed. Numerical results verify that compared with benchmarks, the proposed scheme efficiently improves the system secure computation capacity. © 1967-2012 IEEE.

关键词： Computation offloading

来源：评论

学校读者我要写书评

暂无评论

An Efficient Hybrid Quantum Variational Classifier With Matrix Product State

An Efficient Hybrid Quantum Variational Classifier With Matr...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

ISBN: (纸本)9798350368741

Matrix Product States have been extensively explored as a powerful tool for simulating quantum states in image classification task. However, most research has focused on classical simulations or computations involving high-dimensional unitaries, and significant challenges still exist in the practical preparation of Matrix Product States on quantum computers. This paper proposes a novel and practically feasible quantum variational algorithm based on Matrix Product States for image classification task. We design a hardware-efficient quantum circuit with several adjustable entangling operators to prepare the local tensors in Matrix Product States and integrate minimal residuals to ensure computational stability. We demonstrate that our algorithm can reduce the parameter complexity from growing exponentially with the system size to a linear scale. To validate the effectiveness of this quantum variational algorithm, we conducted experiments on the MNIST dataset, achieving accuracies of 99.95% and 95.96% for binary and ten-class classification tasks, which outperforms other related quantum algorithms. This work advances the practical application of quantum machine learning in resource-constrained environments of quantum computing. © 2025 IEEE.

关键词： Image classification Matrix product states Quantum circuit Variational quantum algorithm

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：