检索结果-内蒙古大学图书馆

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Sheng, Leheng Wang, Wenhan Shi, Zhiyi Zhan, Jichao Kong, Youyong Southeast University Chien-Shiung Wu College China Southeast University Jiangsu Provincial Joint International Research Laboratory of Medical Information Processing School of Computer Science and Engineering China Southeast University Ministry of Education Key Laboratory of Computer Network and Information Integration Nanjing China

ISBN: (纸本)9781728163277

Learning about the cognitive state of the brain has always been a popular topic. Based on the fact that fluctuations of brain signals and functional connectome (FC) relate to specific human behaviors, deep learning based methods have shown promising results on the prediction of such behaviors by analyzing biological signals. Existing methods either model from static perspectives or apply spatial-temporal graph convolution to extract dynamic properties. However, the static information and dynamic information can reflect global brain activities and local brain activities respectively. Thus, we propose BrainNetFormer to incorporate both static and dynamic properties for human behavior prediction. To be specific, a spatial cross attention module and a temporal cross attention module are introduced for information fusion. In addition, since a specific behavior of subjects can be decomposed into a series of subtasks, we introduce a sub-task regularization loss to assist in training and empower the model to recognize subtasks at each moment. Experiments on the HCP-Task dataset demonstrate the superior performance of the proposed model. © 2023 IEEE.

关键词： brain states decoding cross attention functional MRI images Spatial-temporal graph convolution

来源：评论

学校读者我要写书评

暂无评论

Calibration bottleneck: over-compressed representations are less calibratable 24

Calibration bottleneck: over-compressed representations are ...

引用

Proceedings of the 41st International Conference on Machine Learning

作者： Deng-Bao Wang Min-Ling Zhang School of Computer Science and Engineering Southeast University Nanjing China and Key Lab. of Computer Network and Information Integration (Southeast University) MOE China

Although deep neural networks have achieved remarkable success, they often exhibit a significant deficiency in reliable uncertainty calibration. This paper focus on model calibratability, which assesses how amenable a model is to be well recalibrated post-hoc. We find that the widely used weight decay regularizer detrimentally affects model calibratability, subsequently leading to a decline in final calibration performance after post-hoc calibration. To identify the underlying causes leading to poor calibratability, we delve into the calibratability of intermediate features across the hidden layers. We observe a U-shaped trend in the calibratability of intermediate features from the bottom to the top layers, which indicates that over-compression of the top representation layers significantly hinders model calibratability. Based on the observations, this paper introduces a weak classifier hypothesis, i.e., given a weak classification head that has not been over-trained, the representation module can be better learned to produce more calibratable features. Consequently, we propose a progressively layer-peeled training (PLP) method to exploit this hypothesis, thereby enhancing model calibratability. Our comparative experiments show the effectiveness of our method, which improves model calibration and also yields competitive predictive performance.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Task Offloading and Resource Allocation with Privacy Constraints in End-Edge-Cloud Environment 17th

Task Offloading and Resource Allocation with Privacy Constra...

引用

17th CCF Conference on computer Supported Cooperative Work and Social Computing, ChineseCSCW 2022

作者： Zhu, Xia Sun, Wei Li, Xiaoping School of Computer Science and Engineering Southeast University Nanjing China Key Laboratory of Computer Network and Information Integration Ministry of Education Southeast University Nanjing China School of Cyber Science and Engineering Southeast University Nanjing China

ISBN: (纸本)9789819923847

The end-edge-cloud architecture brings an efficient solution to the big data processing problem caused by massive IoT devices. The characteristics of this architecture, such as massive devices, heterogeneous resources, and complex layers, bring new challenges to privacy protection issues. This paper considers the cooperation among terminal devices, edge nodes, and remote clouds to solve the task scheduling problem with privacy constraints by optimizing the offloading decision and resource allocation. A heuristic privacy-aware task offloading and resource allocation algorithm is proposed to maximize the number of successful tasks, which offloads low-privacy and non-privacy tasks to find sub-optimal offloading decisions by offloading sequence generation rules and decision adjustment. Task scheduling algorithms are presented by communication and computing resource allocation strategies for low-privacy and non-privacy tasks. ANOVA technique is used to verify the performance of the proposed algorithm. The experimental results show that the proposed algorithm is superior to others in terms of the number of devices, the amount of different task data, and the proportion of privacy tasks at different levels. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Cloud computing architecture

来源：评论

学校读者我要写书评

暂无评论

HIFI-Stego: A High-Fidelity Embedding Audio Steganography Based on Audio Features Decoupling

IEEE Transactions on Audio, Speech and Language Processing

引用

IEEE Transactions on Audio, Speech and Language Processing 2025年 33卷 2032-2044页

作者： Sanfeng Zhang Baiyu Tian Yang Gao Xingyu Liu Wang Yang School of Cyber Science and Engineering Southeast University Nanjing China Key Laboratory of Computer Network and Information Integration Southeast University Ministry of Education Nanjing China

The higher audio quality of steganography is directly correlated with the increased resistance to steganalysis tools. The advancement of generative AI technologies, particularly those that decouple style features from content, has shown promising developments by facilitating the creation of superior media content. This paper introduces the concept of audio decoupling and presents HIFI-Stego, an embedding audio steganography technique that aims to improve security while maintaining elevated stego audio quality. HIFI-Stego comprises a generator based on the encoder-decoder architecture and a secret message extractor. The encoder of the generator decouples the original audio, yielding the content vector, while the vocoder WORLD is employed to extract the style vector to preserve high-quality audio related features such as timbre and tone. Subsequently, the decoder embeds the secret vector into the decoupled content vector and then couples it with the style vector to generate high-fidelity stego audio. As embedding is not done in traditional time domain or frequency domain, existing analysis tools targeted at traditional steganographies fail to effectively detect the presence of the hidden message. The secret message extractor reuses the generator encoder and augments it with a single-layer convolutional neural network, resulting in a simplified structure suitable for lightweight deployment. Experimental results demonstrate that HIFI-Stego outperforms traditional generative and embedding steganographies in terms of audio quality, steganographic capacity, anti-analysis ability, and concealment.

关键词： Steganography Vectors Time-domain analysis Generators Feature extraction Training Resistance Generative adversarial networks Convolutional neural networks Security

来源：评论

学校读者我要写书评

暂无评论

Joint Optimization of UAV Deployment and Task Computation Offloading Decision in UAV-assisted Edge Computing network

Joint Optimization of UAV Deployment and Task Computation Of...

引用

IEEE International Conference on High Performance Computing and Communications (HPCC)

作者： Yichuan Liu Jinbin Tu Yun Wang Key Lab of Computer Network and Information Integration MOE School of Computer Science and Engineering Southeast University Nanjing China

The UAVs' deployment decision and task computation offloading decision in the UAV-assisted edge computing network significantly impact the operating efficiency of edge network. On the basis of this, the Optimization Model for UAV Cluster Deployment and Computation Offloading Decision (OMUCDCOD) is established. The model jointly optimizes the number, location of UAV s, and task computation offloading decision. Different from previous studies, this model regards the terminal devices in the edge network as virtual MEC servers, and introduces the collaborative computing mode. The task computation offloading decision made can effectively utilize the computing capabilities offered by the edge network. Considering that the two problems of UAV deployment decision and task computation offloading decision are intricately interconnected, we propose a two-layer optimization algorithm combining K-Means and ant colony algorithm (ToKmAc) to solve OMUCDCOD. ToKmAc is divided into upper and lower layers to solve this optimization problem. The upper layer uses K-Means to solve the UAV deployment decision, that is, the number and location of UAVS; the lower layer employs ant colony algorithm to solve computation offloading decision. Finally, extensive experiments verify the effectiveness of OMUCDCOD and ToKmAc.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Improving Scene Text Image Super-resolution via Dual Prior Modulation network

arXiv

引用

arXiv 2023年

作者： Zhu, Shipeng Zhao, Zuoyan Fang, Pengfei Xue, Hui School of Computer Science and Engineering Southeast University Nanjing210096 China MOE Key Laboratory of Computer Network and Information Integration Southeast University China

Scene text image super-resolution (STISR) aims to simultaneously increase the resolution and legibility of the text images, and the resulting images will significantly affect the performance of downstream tasks. Although numerous progress has been made, existing approaches raise two crucial issues: (1) They neglect the global structure of the text, which bounds the semantic determinism of the scene text. (2) The priors, e.g., text prior or stroke prior, employed in existing works, are extracted from pre-trained text recognizers. That said, such priors suffer from the domain gap including low resolution and blurriness caused by poor imaging conditions, leading to incorrect guidance. Our work addresses these gaps and proposes a plug-and-play module dubbed Dual Prior Modulation network (DPMN), which leverages dual image-level priors to bring performance gain over existing approaches. Specifically, two types of prior-guided refinement modules, each using the text mask or graphic recognition result of the low-quality SR image from the preceding layer, are designed to improve the structural clarity and semantic accuracy of the text, respectively. The following attention mechanism hence modulates two quality-enhanced images to attain a superior SR result. Extensive experiments validate that our method improves the image quality and boosts the performance of downstream tasks over five typical approaches on the benchmark. Substantial visualizations and ablation studies demonstrate the advantages of the proposed DPMN. Code is available at: https://***/jdfxzzy/DPMN. Copyright © 2023, The Authors. All rights reserved.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Visual Entity Linking via Multi-modal Learning

引用

Data Intelligence 2022年第1期4卷 1-19页

作者： Qiushuo Zheng Hao Wen Meng Wang Guilin Qi School of Cyber Science and Engineering Southeast UniversityNanjing 211189China School of Computer Science and Engineering Southeast UniversityNanjing 211189China Key Laboratory of Computer Network and Information Integration(Southeast University) Ministry of EducationNanjing 211189China

Existing visual scene understanding methods mainly focus on identifying coarse-grained concepts about the visual objects and their relationships,largely neglecting fine-grained scene *** fact,many data-driven applications on the Web(e.g.,news-reading and e-shopping)require accurate recognition of much less coarse concepts as entities and proper linking them to a knowledge graph(KG),which can take their performance to the next *** light of this,in this paper,we identify a new research task:visual entity linking for fine-grained scene *** accomplish the task,we first extract features of candidate entities from different modalities,i.e.,visual features,textual features,and KG ***,we design a deep modal-attention neural network-based learning-to-rank method which aggregates all features and maps visual objects to the entities in *** experimental results on the newly constructed dataset show that our proposed method is effective as it significantly improves the accuracy performance from 66.46%to 83.16%compared with baselines.

关键词： Knowledge graph Multi-modal learning Entity linking Learning to rank Knowledge graph representation

来源：评论

学校读者我要写书评

暂无评论

Identity-based threshold proxy re-encryption scheme from lattices and its applications

引用

Frontiers of information Technology & Electronic Engineering 2022年第2期23卷 258-277页

作者： Liqiang WU Yiliang HAN Xiaoyuan YANG Minqing ZHANG Key Laboratory of Network and Information Security Engineering University of People’s Armed PoliceXi'an 710086China MOE Key Laboratory of Computer Network and Information Security Xidian UniversityXi'an 710071China

Threshold proxy re-encryption(TPRE)can prevent collusion between a single proxy and a delegatee from converting arbitrary files against the wishes of the delegator through multiple proxies,and can also provide normal services even when certain proxy servers are paralyzed or damaged.A non-interactive identity-based TPRE(IB-TPRE)scheme over lattices is proposed which removes the public key *** accomplish this scheme,Shamir’s secret sharing is employed twice,which not only effectively hides the delegator’s private key information,but also decentralizes the proxy power by splitting the re-encryption *** means that a combiner can detect a misbehaving proxy server that has sent an invalid transformed ciphertext *** property is achieved by lattice-based fully homomorphic *** a result,the whole scheme is thoroughly capable of resisting quantum attacks even when they are *** security of the proposed scheme is based on the decisional learning with error hardness assumption in the standard *** typical application scenarios,including a file-sharing system based on a blockchain network and a robust key escrow system with threshold cryptography,are presented.

关键词： Post-quantum cryptography Threshold proxy re-encryption Lattices Robustness Decentralization

来源：评论

学校读者我要写书评

暂无评论

T-Sorokin: A General Mobility Model in Opportunistic networks

T-Sorokin: A General Mobility Model in Opportunistic Network...

引用

International Conference on computer Supported Cooperative Work in Design

作者： Jinbin Tu Qing Li Yun Wang School of Computer Science and Engineering Southeast University Key Lab of Computer Network and Information Integration MOE Nanjing China

The opportunistic networks are a kind of ad hoc networks that rely on the chance of nodes meeting to transmit messages. Acting as an effective supplement to 4G and 5G networks in some special scenarios where hardware devices are limited, the opportunistic networks have a significant application in health monitoring, warning broadcasting, disaster relief, and so on. The mobility model is one of the research focuses on the opportunistic networks. On the basis of the social mobility theory proposed by Sorokin, a general mobility model, which is suited for various scenarios, called T-Sorokin is proposed. This model is described as a seven-tuple and implemented on the Opportunistic network Environment simulator and fits both Infocom06 and Rome taxi data set, which includes different areas ranging from hotel to city and different mobile units ranging from person to taxi. The results of experiments demonstrate that the T-Sorokin model has the advantage of generality, simplicity, and accuracy. It can simply establish movement tracks close to real data under different scenarios.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Non-Intrusive and Real-Time Data Provenance Method for DDS Systems

A Non-Intrusive and Real-Time Data Provenance Method for DDS...

引用

International Conference on Mobile Ad-hoc and Sensor networks, MSN

作者： Siyi Wei Jinbin Tu Yun Wang School of Computer Science and Engineering Southeast University Key Lab of Computer Network and Information Integration MOE Nanjing China

ISBN: (数字)9798350358261

ISBN: (纸本)9798350358278

In the era of big data, because of the escalating data security threats, the importance of data provenance is widely recognized. It facilitates the tracking of data origins and evolution, improving the ability to diagnose errors and detect potential malicious attacks within large-scale software systems. However, conventional log-based provenance methods usually collect data intrusively, leading to strong coupling with the system. Meanwhile, although there are some log-based provenance systems that support runtime analysis, they exhibit inadequate real-time analysis performance. This paper proposes a non-intrusive and real-time approach to data provenance in DDS systems. The data capture is achieved with the help of the Data-Centric Publisher-Subscriber model of DDS, thus avoiding source code modification to the existing system. The collected data is then described in PROV-DDS model to generate RDF data. RDF datasets inherently form a directed and labeled graph. The query over RDF datasets is decomposed into start subqueries to speed up processing with star subgraph signature. The experimental results show that our method provenances effectively with less than 1% time overhead and a lower query response time than RDF-3X and Redland.

关键词： Runtime Source coding Data security Stars Software systems Resource description framework Real-time systems

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：