检索结果-内蒙古大学图书馆

IIN-FFD:Intra-Inter Network for Face Forgery Detection

Tsinghua science and technology 2024年第6期29卷 1839-1850页

作者： Qihua Zhou Zhili Zhou Zhipeng Bao Weina Niu Yuling Liu School of Software Nanjing University of Information Science and TechnologyNanjing 210044China Institute of Artificial Intelligence Guangzhou UniversityGuangzhou 510006China School of Computer Science University of Electronic Science and Technology of ChinaChengdu 611731SichuanChina School of Computer Science Hunan UniversityChangsha 410082HunanChina

Since different kinds of face forgeries leave similar forgery traces in videos,learning the common features from different kinds of forged faces would achieve promising generalization ability of forgery ***,to accurately detect known forgeries while ensuring high generalization ability of detecting unknown forgeries,we propose an intra-inter network(IIN)for face forgery detection(FFD)in videos with continual *** proposed IIN mainly consists of three modules,i.e.,intra-module,inter-module,and forged trace masking module(FTMM).Specifically,the intra-module is trained for each kind of face forgeries by supervised learning to extract special features,while the inter-module is trained by self-supervised learning to extract the common *** a result,the common and special features of the different forgeries are decoupled by the two feature learning modules,and then the decoupled common features can be utlized to achieve high generalization ability for ***,the FTMM is deployed for contrastive learning to further improve detection *** experimental results on FaceForensic++dataset demonstrate that the proposed IIN outperforms the state-of-the-arts in ***,the generalization ability of the IIN verified on DFDC and Celeb-DF datasets demonstrates that the proposed IIN significantly improves the generalization ability for FFD.

关键词： deep learning information security image classfication neural networks face forgery face forgery detection

来源：评论

学校读者我要写书评

暂无评论

Controllable chaos in the coupled waveguide-optomechanical system with phase shifter

引用

science China(Physics,Mechanics & Astronomy) 2025年第4期68卷 26-34页

作者： Deng-Wei Zhang Pei-Qi Qin Li-Li Zheng Zhi-Ming Zhan Xin-You Lü Department of Mathematics and Physics Luoyang Institute of Science and TechnologyLuoyang 471023China School of Computer and Information Engineering School of SoftwareLuoyang Institute of Science and TechnologyLuoyang 471023China School of Artificial Intelligence Jianghan UniversityWuhan 430056China School of Physics Huazhong University of Science and TechnologyWuhan 430074China

We theoretically investigate chaotic dynamics in an optomechanical system composed of a whispering-gallery-mode(WGM)microresonator and a *** find that tuning the optical phase using a phase shifter and modifying the coupling strength via a unidirectional waveguide(IWG)can induce chaotic *** underlying reason for this phenomenon is that adjusting the phase and coupling strength via the phase shifter and IWG bring the system close to an exceptional point(EP),where field localization dynamically enhances the optomechanical nonlinearity,leading to the generation of chaotic *** addition,due to the sensitivity of chaos to phase in the vicinity of the EP,we propose a theoretical scheme to measure the optical phase perturbations using *** work may offer an alternative approach to chaos generation with current experimental technology and provide theoretical guidance for optical signal processing and chaotic secure communication.

关键词： cavity optomechanics chaos exceptional point

来源：评论

学校读者我要写书评

暂无评论

Learning from Highly Sparse Spatio-temporal data 38

Learning from Highly Sparse Spatio-temporal Data

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Deng, Leyan Lian, Defu Wu, Chenwang Chen, Enhong School of Artificial Intelligence and Data Science University of Science and Technology of China China School of Computer Science and Technology University of Science and Technology of China China

Incomplete spatio-temporal data in the real world has spawned much research. However, existing methods often utilize iterative message-passing across temporal and spatial dimensions, resulting in substantial information loss and high computational cost. We provide a theoretical analysis revealing that such iterative models are susceptible to data and graph sparsity, causing unstable performances on different datasets. To overcome these limitations, we introduce a novel method named One-step Propagation and Confidence-based Refinement (OPCR). In the first stage, OPCR leverages inherent spatial and temporal relationships by employing a sparse attention mechanism. These modules propagate limited observations directly to the global context through one-step imputation, which is theoretically affected only by data sparsity. Following this, we assign confidence levels to the initial imputations by correlating missing data with valid data. This confidence-based propagation refines the separate spatial and temporal imputation results through spatio-temporal dependencies. We evaluate the proposed model across various downstream tasks involving highly sparse spatio-temporal data. Empirical results indicate that our model outperforms state-of-the-art imputation methods, demonstrating its effectiveness and robustness. © 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

Efficient Multimodal Selection for Retrieval in Knowledge-based Visual Question Answering

引用

IEEE Transactions on Circuits and Systems for Video technology 2025年第6期35卷 5195-5207页

作者： Luo, Linyin Lai, Hanjiang Pan, Yan Yin, Jian Sun Yat-Sen University School of Computer Science and Engineering China School of Artificial Intelligence China

Retrieval plays an important role in knowledge-based visual question answering (KB-VQA), which relies on external knowledge to answer questions related to an image. However, not all information in the external knowledge is beneficial in retrieval, e.g., the knowledge that is only semantically similar to the query but is not useful for question answering. To improve the effectiveness and efficiency of retrieval, in this paper, we propose efficient multimodal selection to filter out irrelevant information and increase the retriever performance for KB-VQA. First, to exclude most irrelevant knowledge from the large external knowledge, multimodal selection uses a query-aware sample selection method, which uses the pretrained answer generator's prediction to obtain better positive and negative training samples to help retrievers distinguish knowledge that is semantically relevant to the multimodal query. Then, question-aware visual feature selection is proposed to select the distinguishable visual information related to the question: where cross-attention to questions and images is proposed to obtain question-aware visual features. These visual features are used to perform fine-grained multimodal retrieval within the small set to obtain the final top-related knowledge. The experimental results show that the proposed approach achieves state-of-the-art retrieval performance on the OK-VQA and FVQA datasets, indicating the effectiveness of our selection strategy for retrieval. © 1991-2012 IEEE.

关键词： Question answering

来源：评论

学校读者我要写书评

暂无评论

AI-Enhanced Secure data Aggregation for Smart Grids with Privacy Preservation

引用

computers, Materials & Continua 2025年第1期82卷 799-816页

作者： Congcong Wang Chen Wang Wenying Zheng Wei Gu School of Software Nanjing University of Information Science and TechnologyNanjing210044China School of Information Science and Engineering Zhejiang Sci-Tech UniversityHangzhou310018China State Key Laboratory of Public Big Data Guizhou UniversityGuiyang550025China School of Computer Science and Technology(School of Artificial Intelligence) Zhejiang Sci-Tech UniversityHangzhou310018China School of Computer Science(School of Cyber Science and Engineering) Nanjing University of Information Science and TechnologyNanjing210044China

As smart grid technology rapidly advances,the vast amount of user data collected by smart meter presents significant challenges in data security and privacy *** research emphasizes data security and user privacy concerns within smart ***,existing methods struggle with efficiency and security when processing large-scale *** efficient data processing with stringent privacy protection during data aggregation in smart grids remains an urgent *** paper proposes an AI-based multi-type data aggregation method designed to enhance aggregation efficiency and security by standardizing and normalizing various data *** approach optimizes data preprocessing,integrates Long Short-Term Memory(LSTM)networks for handling time-series data,and employs homomorphic encryption to safeguard user *** also explores the application of Boneh Lynn Shacham(BLS)signatures for user *** proposed scheme’s efficiency,security,and privacy protection capabilities are validated through rigorous security proofs and experimental analysis.

关键词： Smart grid data security privacy protection artificial intelligence data aggregation

来源：评论

学校读者我要写书评

暂无评论

Adaptive Wavelet Convolution Stochastic Analysis Graph Neural Network for Remaining Useful Life Prediction

引用

IEEE Sensors Journal 2025年第11期25卷 19428-19441页

作者： Hu, Qidi Su, Binyi Kong, Jinzhen Liu, Weipeng Hebei University of Technology School of Artificial Intelligence and Data Science Tianjin300130 China Hebei University of Technology School of Mechanical Engineering Tianjin300130 China

The prediction of the Remaining Useful Life (RUL) of stochastically degrading devices is of critical importance for Prognostics and Health Management (PHM). However, most of the existing approaches overlook the inherent coupling of trends and fluctuations in the degradation process. Meanwhile, these methods are struggling to effectively model the random fluctuations component in the degradation process and may lead to inaccurate prediction. To address these limitations, we propose a modified RUL prediction framework WTSAGNN. The framework extends the spatiotemporal dependencies of multi-sensor signals into the frequency domain. By leveraging the proposed adaptive orthogonal wavelet convolution to more effectively decouple the trend and fluctuation components of the degradation process. Furthermore, we introduce a stochastic differential network (SDN) to replace the traditional multilayer perceptron (MLP). The proposed method enhances the robustness of neural networks against stochastic variations in degradation processes through explicit stochastic modeling. Experimental results demonstrate that WTSAGNN outperforms existing state-of-the-art(SOTA) methods on benchmark datasets, achieving up to a 9.1% improvement in prediction accuracy with only a limited increase in computational overhead. In addition, the proposed modules are plug-and-play, allowing seamless integration with existing data-driven prediction frameworks. © 2025 IEEE.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Measuring code maintainability with deep neural networks

引用

Frontiers of computer science 2023年第6期17卷 61-75页

作者： Yamin HU Hao JIANG Zongyao HU School of Computer Science and Technology Beijing Institute of TechnologyBeijing 100081China School of Artificial Intelligence Anhui UniversityHefei 230601China

The maintainability of source code is a key quality characteristic for software *** approaches have been proposed to quantitatively measure code *** approaches rely heavily on code metrics,e.g.,the number of Lines of Code and McCabe’s Cyclomatic *** employed code metrics are essentially statistics regarding code elements,e.g.,the numbers of tokens,lines,references,and branch ***,natural language in source code,especially identifiers,is rarely exploited by such *** a result,replacing meaningful identifiers with nonsense tokens would not significantly influence their outputs,although the replacement should have significantly reduced code *** this end,in this paper,we propose a novel approach(called DeepM)to measure code maintainability by exploiting the lexical semantics of text in source *** leverages deep learning techniques(e.g.,LSTM and attention mechanism)to exploit these lexical semantics in measuring code *** key rationale of DeepM is that measuring code maintainability is complex and often far beyond the capabilities of statistics or simple ***,DeepM leverages deep learning techniques to automatically select useful features from complex and lengthy inputs and to construct a complex mapping(rather than simple heuristics)from the input to the output(code maintainability index).DeepM is evaluated on a manually-assessed *** evaluation results suggest that DeepM is accurate,and it generates the same rankings of code maintainability as those of experienced programmers on 87.5%of manually ranked pairs of Java classes.

关键词： code maintainability lexical semantics deep learning neural networks

来源：评论

学校读者我要写书评

暂无评论

Semantic-Aware Late-Stage Supervised Contrastive Learning for Fine-Grained Action Recognition

引用

IEEE Transactions on Circuits and Systems for Video technology 2025年第6期35卷 5912-5924页

作者： Pan, Yijun Zhao, Quan Zhang, Yueyi Wang, Zilei Sun, Xiaoyan Wu, Feng University of Science and Technology of China School of Artificial Intelligence and Data Science Hefei230026 China University of Science and Technology of China School of Information Science and Technology Hefei230026 China University of Science and Technology of China National Engineering Laboratory for Brain-Inspired Intelligence Technology and Application Hefei230026 China

Fine-grained action recognition typically faces challenges with lower inter-class variances and higher intra-class variances. Supervised contrastive learning is inherently suitable for this task, as it can decrease intra-class feature distances while increasing inter-class ones. However, directly applying it into fine-grained action recognition encounters two main problems. The first problem stems from the heavy training cost associated with supervised contrastive learning, which requires numerous training epochs, each involving double augmentation views per instance. To address this issue, we propose the late-stage supervised contrastive learning (late-SC) strategy, which effectively reduces the number of training epochs needed for the contrastive learning process. The second problem is that supervised contrastive loss does not explicitly consider the semantic distances between fine-grained actions when adjusting representation distances. This results in less reasonable and efficient adjustments to the representation space. To overcome this limitation, we introduce the semantic-aware temperature adaptation (STA) mechanism, enhancing the suitability of the supervised contrastive loss for fine-grained action recognition. We conduct experiments on several benchmark datasets for fine-grained action recognition, including Epic-Kitchens-55/100, SomethingSomething-V1, and Diving48-V2. The results demonstrate that our proposed method (referred to as LSC-STA) consistently enhances performance across various base feature extractors, without introducing additional inference overhead and incurring only a marginal increase in training expenses. © 1991-2012 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Reconnaissance-Strike Complex: A Network-Layer Solution to the Natural Forking in Blockchain

引用

IEEE Transactions on Information Forensics and Security 2025年 20卷 2022-2034页

作者： Chen, Anlin Wang, Shengling Shi, Hongwei Guo, Yu Cheng, Xiuzhen Beijing Normal University School of Artificial Intelligence Beijing100875 China School of Computer Science and Technology Shandong Qingdao266237 China

The natural forking severely compromises the security and wastes resources of Blockchain. Current analyses of the natural forking are carried out from microscale and macroscale perspectives, each facing challenges in generality and accuracy respectively. This results in dire straits that the forking arising from the network layer cannot be solved within the same layer, and existing defense schemes concentrate on the consensus layer, unfortunately coming at the expense of diminishing decentralized trust. Hence, to overcome these shortcomings, we propose the first reconnaissance-strike solution to the natural forking where the issue is recognized at the network layer and further struck on-site. Specifically, our endeavors encompass 1) analyzing the spatial-temporal transmission dynamics of the main chain. We exert mesoscale perspective by transforming the behavioral analysis of the transmitters (i.e., the nodes) into the movement analysis of the transmitted object (i.e., the main chain) which mitigates following Lévy mobility. Based on this, we quantify the dynamics of long-range leaping and short-range diffusion of the main chain transmission;2) proposing a cost-effective anti-forking mechanism. This mechanism combats the forking with low cost by configuring logical connections at the network management level, based on the quantitative relationship between Blockchain network topology and the natural forking rate we have derived. Both theoretical analysis and extensive experiments verify that our scheme can maintain the natural forking rate not more than the given threshold in most cases. © 2005-2012 IEEE.

关键词： Cost effectiveness

来源：评论

学校读者我要写书评

暂无评论

NeuroProbe: A Performance Profiling Feed Forward Neural Network Simulator for Hardware Interaction Analysis

引用

SN computer science 2025年第5期6卷 1-22页

作者： Ayyappan, Binu Kumar, G. Santhosh Department of Computer Science Cochin University of Science and Technology Kochi India Department of Artificial Intelligence and Data Science Rajagiri School of Engineering & Technology Kochi India

NeuroProbe is a simple neural network simulator designed by authors specifically for educational purposes focusing on simulating inference phase on a computationally capable embedded hardware, aiming to provide a deeper understanding of hardware interactions and performance improvement. This simulator incorporates performance profiling capabilities, allowing students and educators to analyze and optimize the performance of a simulated trained neural network. By simulating the intricate hardware interactions, NeuroProbe enables users to gain insights into how various components impact the overall performance of neural networks. Using this simulator, learners can experiment with different configurations and architectures, identify potential bottlenecks, and explore optimization strategies to enhance neural network performance before deploying the model on an Edge Computing environment. It also enriches computer science education by enabling learners to gain deeper insights into hardware interactions and performance improvement. By identifying performance bottlenecks, learners can refine their understanding of neural networks and develop skills in optimizing their models for enhanced performance. In conclusion, NeuroProbe empowers users to bridge the gap between theoretical knowledge and practical implementation of neural network on an embedded device, fostering a more comprehensive understanding of neural networks and their hardware requirements. Currently, it is focused solely on single-threaded CPU-based computation enabling SSE, AVX. At present, NeuroProbe does not support SIMD extensions such as AVX-FMA, AVX-512, multi-threaded computation, and accelerators such as NPU, TPU, and GPU. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： Hardware–software interactions Neural network simulator NeuroProbe Performance bottlenecks Performance profiling Simulations

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：