检索结果-内蒙古大学图书馆

Fake Face Detection Based on Fusion of Spatial Texture and High-Frequency Noise

Chinese Journal of Electronics 2025年第1期34卷 212-221页

作者： Dengyong Zhang Feifan Qi Jiahao Chen Jiaxin Chen Rongrong Gong Yuehong Tian Lebing Zhang Hunan Provincial Key Laboratory of Intelligent Processing of Big Data on Transportation School of Computer and Communication Engineering Changsha University of Science and Technology School of Computer and Communication Engineering Changsha University of Science and Technology Changsha Social Work College Changkuangao Beijing Technology Co. Ltd. School of Computer and Artificial Intelligence Huaihua University

The rapid development of the Internet has led to the widespread dissemination of manipulated facial images, significantly impacting people's daily lives. With the continuous advancement of Deepfake technology, the generated counterfeit facial images have become increasingly challenging to distinguish. There is an urgent need for a more robust and convincing detection method. Current detection methods mainly operate in the spatial domain and transform the spatial domain into other domains for analysis. With the emergence of transformers, some researchers have also combined traditional convolutional networks with transformers for detection. This paper explores the artifacts left by Deepfakes in various domains and, based on this exploration, proposes a detection method that utilizes the steganalysis rich model to extract high-frequency noise to complement spatial features. We have designed two main modules to fully leverage the interaction between these two aspects based on traditional convolutional neural networks. The first is the multi-scale mixed feature attention module, which introduces artifacts from high-frequency noise into spatial textures, thereby enhancing the model's learning of spatial texture features. The second is the multi-scale channel attention module, which reduces the impact of background noise by weighting the features. Our proposed method was experimentally evaluated on mainstream datasets, and a significant amount of experimental results demonstrate the effectiveness of our approach in detecting Deepfake forged faces, outperforming the majority of existing methods.

关键词： Deepfakes Adaptation models Attention mechanisms Noise Transforms Transformers Feature extraction Internet Convolutional neural networks Faces

来源：评论

学校读者我要写书评

暂无评论

An overview on IRS-enabled sensing and communications for 6G: architectures, fundamental limits, and joint beamforming designs

引用

Science China(Information Sciences) 2025年第5期68卷 170-193页

作者： Xianxin SONG Yuan FANG Feng WANG Zixiang REN Xianghao YU Ye ZHANG Fan LIU Jie XU Derrick Wing Kwan NG Rui ZHANG Shuguang CUI School of Science and Engineering (SSE) Shenzhen Future Network of Intelligence Institute (FNii-Shenzhen)and Guangdong Provincial Key Laboratory of Future Networks of Intelligence The Chinese University of Hong Kong Department of Electrical Engineering City University of Hong Kong School of Information Engineering Guangdong University of Technology Key Laboratory of Wireless-Optical Communications Chinese Academy of Sciences School of Information Science and Technology University of Science and Technology of China Computer School Beijing Information Science and Technology University National Mobile Communications Research Laboratory School of Information Science and Engineering Southeast University School of Electrical Engineering and Telecommunications University of New South Wales School of Science and Engineering Shenzhen Research Institute of Big Data The Chinese University of Hong Kong Department of Electrical and Computer Engineering National University of Singapore

This study presents an overview on intelligent reflecting surface(IRS)-enabled sensing and communication for the forthcoming sixth-generation(6G) wireless networks, in which IRSs are strategically deployed to proactively reconfigure wireless environments to improve both sensing and communication(S&C) performance. First, we exploit a single IRS to enable wireless sensing in the base station's(BS's) non-line-of-sight(NLoS) area. In particular, we present three IRS-enabled NLoS target sensing architectures with fully-passive, semi-passive, and active IRSs, respectively. We compare their pros and cons by analyzing the fundamental sensing performance limits for target detection and parameter estimation. Next, we consider a single IRS to facilitate integrated sensing and communication(ISAC), in which the transmit signals at the BS are used for achieving both S&C functionalities, aided by the IRS through reflective beamforming. We present joint transmit signal and receiver processing designs for realizing efficient ISAC, and jointly optimize the transmit beamforming at the BS and reflective beamforming at the IRS to balance the fundamental performance tradeoff between S&C. Furthermore, we discuss multi-IRS networked ISAC, by particularly focusing on multi-IRS-enabled multi-link ISAC, multi-region ISAC, and ISAC signal routing, respectively. Finally, we highlight various promising research topics in this area to motivate future work.

关键词： integrated sensing and communication (ISAC) intelligent reflecting surface (IRS) non-line-of-sight (NLoS) sensing sensing and communication tradeoff

来源：评论

学校读者我要写书评

暂无评论

HTOTP: Honey Time-based One-Time Passwords

引用

IEEE Transactions on Information Forensics and Security 2025年 20卷 4438-4453页

作者： Ding, Zixuan Wang, Ding Nankai University College of Cryptology and Cyber Science Tianjin300350 China Chinese Academy of Sciences Key Laboratory of Cyberspace Security Defense [Institute of Information Engineering Beiiing100085 China Nankai University Ministry of Education Key Laboratory of Data and Intelligent System Security Tianjin300350 China

One-Time Passwords (OTPs) play a crucial role in Two-Factor Authentication (2FA) and Multi-Factor Authentication (MFA) by adding an additional layer of security. OTPs effectively mitigate the risk of static passwords being intercepted and reused. Nevertheless, both academic schemes and industrial solutions still encounter security threats stemming from server/device compromises and OTP factor forgery. Chain-based asymmetric OTP schemes are a promising approach to addressing server compromise but still face threats from device compromise and pre-generated chain leakage. We emphasize that since devices directly store the OTP seed, OTP authentication is essentially equivalent to verifying device possession. This means that in existing OTP schemes, OTP forgery and device compromise remain prevalent and difficult to overcome. In this work, we propose a brand new scheme to address OTP factor forgery and server/device compromises. For the first time, our scheme constructs a tightly coupled architecture between the password factor and the OTP factor. The OTP seed is derived from a password and a device-stored salt, preventing OTP seed extraction and OTP forgery even in the event of a device compromise. Through the integration of "honeywords"with the tightly coupled OTP architecture, the server stores decoy OTP seeds generated by decoy passwords, providing resistance against server compromises and partial password guessing from devices. We conduct a comprehensive evaluation of our OTP schemes. The computational overhead is correlated with the number of honeywords, and with the recommended set size of 20, the total verification overhead is approximately 0.24ms. Additionally, we propose formal security properties and application metrics, and rigorously prove our scheme's resistance against server/device compromise attacks and guessing attacks. Our scheme is the first to achieve comprehensive OTP security with low overhead. © 2025 IEEE.

关键词： Authentication

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Learning Based data-Driven Control Method for Dynamic Systems and Its Application to FLUIDMechatronix Platform 16th

Reinforcement Learning Based Data-Driven Control Method for...

引用

16th International Conference on Modelling, Identification and Control, ICMIC 2024

作者： Zou, Zonghan Gao, Jingjing Yang, Xu Huang, Jian Zhang, Tao Key Laboratory of Knowledge Automation for Industrial Processes of Ministry of Education School of Automation and Electrical Engineering University of Science and Technology Beijing Beijing100083 China

ISBN: (纸本)9789819617760

With the increasing complexity of industrial process, accurate models of automatic control systems are difficult to be established, which makes model-based control methods become less effective. On this premise, a reinforcement learning (RL) based data-driven control method for dynamic systems is proposed in this paper. For our purposes, a model-free RL controller with an Actor-Critic structure is developed by using neural networks (NN), combining the merits of both value-based and policy-based RL algorithms. Policy of the actor is assessed by the critic, updating parameters of RL controller in real-time for online optimization. The effectiveness of the RL controller is demonstrated by an experimental study on FLUIDMechatronix platform. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Dynamic-prototype Contrastive Fine-tuning for Continual Few-shot Relation Extraction with Unseen Relation Detection 31

Dynamic-prototype Contrastive Fine-tuning for Continual Few-...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Zhao, Simiao Tan, Zhen Pang, Ning Xiao, Weidong Zhao, Xiang National Key Laboratory of Information Systems Engineering National University of Defense Technology China Laboratory for Big Data and Decision National University of Defense Technology China

ISBN: (纸本)9798891761964

Continual Few-shot Relation Extraction (CFRE) aims to continually learn new relations from limited labeled data while preserving knowledge about previously learned relations. Facing the inherent issue of catastrophic forgetting, previous approaches predominantly rely on memory replay strategies. However, they often overlook task interference in continual learning and the varying memory requirements for different relations. To address these shortcomings, we propose a novel framework, DPC-FT, which features: 1) a lightweight relation encoder for each task to mitigate negative knowledge transfer across tasks;2) a dynamic prototype module to allocate less memory for easier relations and more memory for harder relations. Additionally, we introduce the None-Of-The-Above (NOTA) detection in CFRE and propose a threshold criterion to identify relations that have never been learned. Extensive experiments demonstrate the effectiveness and efficiency of our method in CFRE, making our approach more practical and comprehensive for real-world scenarios. © 2025 Association for Computational Linguistics.

关键词： knowledge transfer

来源：评论

学校读者我要写书评

暂无评论

Semantic and Sentiment Dual-Enhanced Generative Model for Script Event Prediction 31

Semantic and Sentiment Dual-Enhanced Generative Model for Sc...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Wu, Feiyang Huang, Peixin Hu, Yanli Tan, Zhen Zhao, Xiang Laboratory for Big Data and Decision National University of Defense Technology China National Key Laboratory of Information Systems Engineering National University of Defense Technology China

ISBN: (纸本)9798891761964

Script Event Prediction (SEP) aims to forecast the next event in a sequence from a list of candidates. Traditional methods often use pre-trained language models to model event associations but struggle with semantic ambiguity and embedding bias. Semantic ambiguity arises from the multiple meanings of identical words and insufficient consideration of event arguments, while embedding bias results from assigning similar word embeddings to event pairs with similar lexical features, despite their different meanings. To address above issues, we propose a the Semantic and Sentiment Dual-enhanced Generative Model (SSD-GM). SSD-GM leverages two types of script event information to enhance the generative model. Specifically, it employs a GNN-based semantic structure aggregator to integrate the event-centric structure information, thereby mitigating the impact of semantic ambiguity. Furthermore, we find that local sentiment variability effectively reduces biases in event embeddings, while maintaining global sentiment consistency enhances predictive accuracy. As a result, SSD-GM adeptly captures both global and local sentiment of events through its sentiment information awareness mechanism. Extensive experiments on the Multi-Choice Narrative Cloze (MCNC) task demonstrate that our approach achieves better results than other state-of-the-art baselines. © 2025 Association for Computational Linguistics.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

CoT-based Synthesizer: Enhancing LLM Performance through Answer Synthesis

arXiv

引用

arXiv 2025年

作者： Zhang, Bohan Zhang, Xiaokang Zhang, Jing Yu, Jifan Luo, Sijia Tang, Jie School of Information Renmin University of China China Tsinghua University China Key Laboratory of Data Engineering and Knowledge Engineering Beijing China

Current inference scaling methods, such as Self-consistency and Best-of-N, have proven effective in improving the accuracy of LLMs on complex reasoning tasks. However, these methods rely heavily on the quality of candidate responses and are unable to produce correct answers when all candidates are incorrect. In this paper, we propose a novel inference scaling strategy, CoT-based Synthesizer, which leverages CoT reasoning to synthesize superior answers by analyzing complementary information from multiple candidate responses, even when all candidate responses are flawed. To enable a lightweight and cost-effective implementation, we introduce an automated data generation pipeline that creates diverse training data. This allows smaller LLMs trained on this data to improve the inference accuracy of larger models, including API-based LLMs. Experimental results across four benchmark datasets with seven policy models demonstrate that our method significantly enhances performance, with gains of 11.8% for Llama3-8B and 10.3% for GPT-4o on the MATH dataset. The corresponding training data and code are publicly available on the repository. Copyright © 2025, The Authors. All rights reserved.

关键词： data accuracy

来源：评论

学校读者我要写书评

暂无评论

Dynamic Scaling of Unit Tests for Code Reward Modeling

arXiv

引用

arXiv 2025年

作者： Ma, Zeyao Zhang, Xiaokang Zhang, Jing Yu, Jifan Luo, Sijia Tang, Jie School of Information Renmin University of China China Tsinghua University China Key Laboratory of Data Engineering and Knowledge Engineering Beijing China

Current large language models (LLMs) often struggle to produce accurate solutions on the first attempt for code generation. Prior research tackles this challenge by generating multiple candidate solutions and validating them with LLM-generated unit tests. The execution results of unit tests serve as reward signals to identify correct solutions. As LLMs always confidently make mistakes, these unit tests are not reliable, thereby diminishing the quality of reward signals. Motivated by the observation that scaling the number of solutions improves LLM performance, we explore the impact of scaling unit tests to enhance reward signal quality. Our pioneer experiment reveals a positive correlation between the number of unit tests and reward signal quality, with greater benefits observed in more challenging problems. Based on these insights, we propose CodeRM-8B, a lightweight yet effective unit test generator that enables efficient and high-quality unit test scaling. Additionally, we implement a dynamic scaling mechanism that adapts the number of unit tests based on problem difficulty, further improving efficiency. Experimental results show that our approach significantly improves performance across various models on three benchmarks (e.g., with gains of 18.43% for Llama3-8B and 3.42% for GPT-4o-mini on HumanEval Plus). © 2025, CC BY-SA.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A quinolinyl analog of resveratrol improves neuronal damage after ischemic stroke by promoting Parkin-mediated mitophagy

引用

Chinese Journal of Natural Medicines 2025年第2期23卷 214-224页

作者： Qingqi Meng Yan Mi Libin Xu Yeshu Liu Dong Liang Yongping Wang Yan Wang Yueyang Liu Guoliang Chen Yue Hou Key Laboratory of Bioresource Research and Development of Liaoning Province College of Life and Health SciencesNational Frontiers Science Center for Industrial Intelligence and Systems OptimizationKey Laboratory of Data Analytics and Optimization for Smart IndustryMinistry of EducationNortheastern UniversityShenyang 110167China State Key Laboratory for Chemistry and Molecular Engineering of Medicinal Resources School of Chemistry and Pharmaceutical SciencesGuangxi Normal UniversityGuilin 541004China Shenyang Key Laboratory of Vascular Biology Science and Research CenterDepartment of PharmacologyShenyang Medical CollegeShenyang 110034China Key Laboratory of Structure-Based Drug Design&Discovery of Ministry of Education School of Pharmaceutical EngineeringShenyang Pharmaceutical UniversityShenyang 110016China

Ischemic stroke(IS)is a prevalent neurological disorder often resulting in significant disability or ***,extracted from Polygonum cuspidatum *** Zucc.(commonly known as Japanese knotweed),has been recognized for its potent neuroprotective ***,the neuroprotective efficacy of its derivative,(E)-4-(3,5-dimethoxystyryl)quinoline(RV02),against ischemic stroke remains inadequately *** study aimed to evaluate the protective effects of RV02 on neuronal ischemia-reperfusion injury both in vitro and in *** research utilized an animal model of middle cerebral artery occlusion/reperfusion and SH-SY5Y cells subjected to oxygen-glucose deprivation and reperfusion to simulate ischemic *** findings demonstrate that RV02 attenuates neuronal mitochondrial damage and scavenges reactive oxygen species(ROS)through mitophagy ***,Parkin knockdown was found to abolish RV02's ability to activate mitophagy and neuroprotection in *** results suggest that RV02 shows promise as a neuroprotective agent,with the activation of Parkin-mediated mitophagy potentially serving as the primary mechanism underlying its neuroprotective effects.

关键词： (E)-4-(3 5-dimethoxystyryl)quinoline Resveratrol Ischemic stroke Mitophagy Parkin

来源：评论

学校读者我要写书评

暂无评论

Dual-antenna GNSS/MEMS INS Tightly Coupled Algorithm for Agricultural Machinery Based on Adaptive Federated Filtering

引用

IEEE Sensors Journal 2025年第9期25卷 14780-14792页

作者： Feng, Yihang Huang, Guanwen Li, Xin Li, Zhenhong Zhang, Kai Li, Hang Jing, Ce Chang'an University College of Geological Engineering and Geomatics Xi'an710054 China Ministry of Natural Resources Key Laboratory of Ecological Geology and Disaster Prevention Xi'an710054 China Universities of Shaanxi Province Engineering Research Center of Intelligent Bds Xi'an710054 China Key Laboratory of Western China's Mineral Resource Geological Engineering Ministry of Education Xi'an710054 China Chang'an University Big Data Center for Geosciences and Satellites Xi'an710054 China

Modern agricultural machinery relies on high accuracy navigation systems, however, the common loosely coupled (LC) solution of dual-antenna global navigation satellite system (GNSS) and micro-electro-mechanical systems inertial navigation system (MEMS INS) often fails to meet accuracy requirements in complex environments. Theoretically, the tightly coupled (TC) solution of the dual-antenna baseline constraint and MEMS INS offers better attitude accuracy. However, its state space is incomplete, comprising only attitude, gyro biases, and ambiguity. Moreover, the previous studies have not conducted a state observability analysis on this model, which is essential for understanding its state estimation capabilities. Therefore, we derived the TC model of dual-antenna baseline constraint and MEMS INS within a complete state space and performed an observability analysis. Based on these results and considering computational efficiency, we integrated this model into the GNSS/MEMS INS TC model using federated filtering. To further improve the algorithm's accuracy in complex agricultural environments, an adaptive robust positioning algorithm is proposed based on turning state detection. The proposed algorithm was validated through three sets of experiments, demonstrating position accuracy within 2 cm in both open and slightly occluded environments, with heading accuracy within 0.6°, and maintaining optimal accuracy even in severely occluded environments.(Figure presented.) © 2001-2012 IEEE.

关键词： Adaptive filtering

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：