检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Gao, Yinghua Li, Yiming Gong, Xueluan Li, Zhifeng Xia, Shu-Tao Wang, Qian The Tsinghua Shenzhen International Graduate School Tsinghua University Shenzhen518055 China The State Key Laboratory of Blockchain and Data Security Zhejiang University Hangzhou311200 China College of Computing and Data Science Nanyang Technological University 639798 Singapore The School of Computer Science Wuhan University China Tencent Data Platform Shenzhen518057 China The Research Center of Artificial Intelligence Peng Cheng Laboratory Shenzhen518000 China The School of Cyber Science and Engineering Wuhan University China

Deep neural networks (DNNs) are vulnerable to backdoor attacks, where the adversary manipulates a small portion of training data such that the victim model predicts normally on the benign samples but classifies the triggered samples as the target class. The backdoor attack is an emerging yet threatening training-phase threat, leading to serious risks in DNN-based applications. In this paper, we revisit the trigger patterns of existing backdoor attacks. We reveal that they are either visible or not sparse and therefore are not stealthy enough. More importantly, it is not feasible to simply combine existing methods to design an effective sparse and invisible backdoor attack. To address this problem, we formulate the trigger generation as a bi-level optimization problem with sparsity and invisibility constraints and propose an effective method to solve it. The proposed method is dubbed sparse and invisible backdoor attack (SIBA). We conduct extensive experiments on benchmark datasets under different settings, which verify the effectiveness of our attack and its resistance to existing backdoor defenses. The codes for reproducing main experiments are available at https://***/YinghuaGao/SIBA. © 2023, CC BY.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Contamination-resilient anomaly detection via adversarial learning on partially-observed normal and anomalous data 24

Contamination-resilient anomaly detection via adversarial le...

引用

Proceedings of the 41st International Conference on Machine Learning

作者： Wenxi Lv Qinliang Su Hai Wan Hongteng Xu Wenchao Xu School of Computer Science and Engineering Sun Yat-sen University Guangzhou China School of Computer Science and Engineering Sun Yat-sen University Guangzhou China and Guangdong Key Laboratory of Big Data Analysis and Processing Guangzhou China Gaoling School of Artifical Intelligence Renmin University of China Beijing China Department of Computing The Hong Kong Polytechnic University Hong Kong SAR

Many existing anomaly detection methods assume the availability of a large-scale normal dataset. But for many applications, limited by resources, removing all anomalous samples from a large unlabeled dataset is unrealistic, resulting in contaminated datasets. To detect anomalies accurately under such scenarios, from the probabilistic perspective, the key question becomes how to learn the normal-data distribution from a contaminated dataset. To this end, we propose to collect two additional small datasets that are comprised of partially-observed normal and anomaly samples, and then use them to help learn the distribution under an adversarial learning scheme. We prove that under some mild conditions, the proposed method is able to learn the correct normal-data distribution. Then, we consider the overfitting issue caused by the small size of the two additional datasets, and a correctness-guaranteed flipping mechanism is further developed to alleviate it. Theoretical results under incomplete observed anomaly types are also presented. Extensive experimental results demonstrate that our method outperforms representative baselines when detecting anomalies under contaminated datasets.

关键词：

来源：评论

学校读者我要写书评

暂无评论

JTEA: Implementing Semi-Supervised Entity Alignment Using Joint Teaching Strategies

JTEA: Implementing Semi-Supervised Entity Alignment Using Jo...

引用

IEEE International Conference on Systems, Man and Cybernetics

作者： Kai Lu Jing Zhao Lichao Ding Zenghao Hao Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center Qilu University of Technology (Shandong Academy of Sciences) Jinan China Faculty of Computer Science and Technology Shandong Engineering Research Center of Big Data Applied Technology Qilu University of Technology (Shandong Academy of Sciences) Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China

ISBN: (数字)9781665410205

ISBN: (纸本)9781665410212

Entity alignment(EA) aims to identify equivalent entities in different knowledge graphs or data sources and establish correspondences between them. In recent years, embedding-based methods have made significant progress in the field of entity alignment, yet the shortage of adequate training data (i.e., seed entity pairs) continues to be a primary challenge. Traditional semi-supervised methods typically address this issue by generating potential mappings (i.e., pseudo-mappings) for unlabeled entities. However, these methods often suffer from the influence of erroneous potential mappings or overlook the uncertainty of potential mappings. Therefore, we propose a novel semi-supervised entity alignment method called JTEA. This method combines seed entity pairs with potential mappings through end-to-end joint training, providing strong guidance for model training. In constructing potential mappings, we propose a bidirectional matching strategy (BMS), which generates more detailed potential mappings by jointly adjusting decisions in different directions and uses a joint matching confidence score to characterize its uncertainty. Furthermore, we also introduce a method of matching diversity alignment selection, aiming to screen out more reliable entity alignments to further improve the performance and stability of the model. We conducted a large number of experiments on the benchmark dataset to fully validate the effectiveness of the JTEA method.

关键词： Training Uncertainty Soft sensors Noise Training data Knowledge graphs data models Stability analysis Reliability Thermal stability

来源：评论

学校读者我要写书评

暂无评论

HGTA: News Recommendation Based on Hierarchical Granular Semantic Embeddings and Threshold Attention

HGTA: News Recommendation Based on Hierarchical Granular Sem...

引用

International Conference on Computer Supported Cooperative Work in Design

作者： Yiwen Li Zhenyu Yang Zhijun Li Xiaoyang Li Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center Qilu University of Technology (Shandong Academy of Sciences) Jinan China Shandong Engineering Research Center of Big Data Applied Technology Faculty of Computer Science and Technology Qilu University of Technology (Shandong Academy of Sciences) Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China

ISBN: (数字)9798350349184

ISBN: (纸本)9798350349191

Personalized news recommendation helps users find interesting content among much news information. Most existing news recommendation techniques densely interact a user’s historical clicked news with candidate news to capture their reading interest. However, during this process, even if certain historical clicked news is weakly correlated with the candidate news, the model still interacts with it, generating more noise information when constructing user interests. Simultaneously, the news representation learning process using neural networks is a fine-grained to coarse-grained learning process; i.e., as the network deepens, the fine-grained original news information is easily lost. Therefore, to solve the above problems, this paper proposes a news recommendation model based on hierarchical granular semantic embeddings and threshold attention (HGTA). In HGTA, we propose a threshold attention module to filter unimportant historical clicked news by setting an attention score threshold, i.e., interacting only with historical clicked news that is strongly correlated with the candidate news when capturing user interests. Thus, the model attention is focused, and the noisy interaction information is reduced. Furthermore, when training the neural network, to prevent the loss of fine-grained information, we design a hierarchical granular information extraction module to capture both coarse- and fine-grained news features to better enrich the news semantics. Experiments conducted on the large real-world Microsoft News dataset (MIND) show that our model achieves highly competitive results in terms of popular metrics. Our code is published at the following URL: https://***/liyiweneven/HGTA.

关键词： Uniform resource locators Training Representation learning Federated learning Semantics Neural networks Noise

来源：评论

学校读者我要写书评

暂无评论

LESaET: Low-Dimensional Embedding Method for Link Prediction Combining Self-Attention and Enhanced-TuckER

LESaET: Low-Dimensional Embedding Method for Link Prediction...

引用

IEEE International Conference on Systems, Man and Cybernetics

作者： Lichao Ding Jing Zhao Kai Lu Zenghao Hao Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center Qilu University of Technology (Shandong Academy of Sciences) Jinan China Shandong Engineering Research Center of Big Data Applied Technology Faculty of Computer Science and Technology Qilu University of Technology (Shandong Academy of Sciences) Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China

ISBN: (数字)9781665410205

ISBN: (纸本)9781665410212

Knowledge graphs (KGs) provide a structured representation of the real world through entity-relation triples. However, current KGs are often incomplete, typically containing only a small fraction of all possible facts. This involves inferring missing content from existing information, a task known as link prediction. Existing methods in the field of link prediction struggle with controlling the dimensionality of embedding vectors or suffer from overly complex models. In order to tackle these challenges, we introduce a method in this paper, named Low-Dimensional Embedding Method for Link Prediction Combining Self-Attention and Enhanced-TuckER (LESaET). LESaET leverages both self-attention mechanisms and tensor factorization to learn expressive contextual-enhanced representations of KGs. Specifically, LESaET employs the multi-head self-attention mechanism of Transformer as an encoder to capture the mutual information between entities and relationships, and utilizes Enhanced-TuckER as a decoder, ultimately achieving expressive low-dimensional embeddings for link prediction tasks. LESaET demonstrates competitive performance compared to advanced methods on standard datasets.

关键词： Tensors Stacking Knowledge graphs Predictive models Transformers Vectors Decoding Standards Mutual information Periodic structures

来源：评论

学校读者我要写书评

暂无评论

Security Enhancing for IRS-Aided NOMA Networks via Constructive Interference Precoding

Security Enhancing for IRS-Aided NOMA Networks via Construct...

引用

International Symposium on Microwave, Antenna, Propagation and EMC Technologies for Wireless Communications (MAPE)

作者： Jingying Bao Yang Cao Xiaoqi Qin Lexi Xu Nan Zhao Dusit Niyato School of Information and Communication Engineering Dalian University of Technology Dalian China Engineering Systems and Design Pillar Singapore University of Technology and Design Singapore State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China Research Institute China United Network Communications Corporation Beijing China College of Computing and Data Science Nanyang Technological University Singapore

ISBN: (数字)9798331533922

ISBN: (纸本)9798331533939

In this paper, we propose a constructive interference precoding (CIP) enabled secure intelligent reflecting surface (IRS)-non-orthogonal multiple access (NOMA) scheme, lever-aging both the inter-user interference and artificial noise to boost the legitimate transmission of far user, while inducing the eavesdropper to decode the deceptive information. Based on this CIP-NOMA scheme, we minimize the transmit power under the perfect channel state information (CSI), subject to the CIP constraint for the far user and eavesdropper, while guaranteeing the quality of service and successive interference cancellation for the near user and the IRS unit modulus constraint. To handle this non-convex problem, an alternating optimization algorithm is proposed. Specifically, by alternately optimizing the precoding vectors at the base station and the IRS reflecting matrix via the successive convex approximation and the penalty-based algorithm, respectively, a reliable solution can be obtained. Simulation results demonstrate that the proposed scheme can achieve better security performance with less energy consumption compared to the conventional NOMA.

关键词： Microwave antennas Wireless communication NOMA Precoding Simulation Noise Approximation algorithms Vectors Security Reliability

来源：评论

学校读者我要写书评

暂无评论

Overcoming data Limitations and Cross-Modal Interaction Challenges in Medical visual Question Answering

Overcoming Data Limitations and Cross-Modal Interaction Chal...

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Yurun Bi Xingang Wang Qingao Wang Jinxiao Yang Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center Qilu University of Technology (Shandong Academy of Sciences) Jinan China Shandong Engineering Research Center of Big Data Applied Technology Faculty of Computer Science and Technology Qilu University of Technology (Shandong Academy of Sciences) Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China

ISBN: (数字)9798350359312

ISBN: (纸本)9798350359329

Medical visual Question Answering (Med-VQA) is a sophisticated multimodal technology within the medical domain that utilizes natural language processing and artificial intelligence techniques. It serves as a tool in the medical field to provide information and answers, aiming to correctly respond to medical questions related to specific professional medical images. The primary goal is to assist patients in understanding their physical condition and provide decision support for healthcare professionals. While Generalized visual Question Answering has seen significant advancements, challenges persist in the medical domain. Firstly, medical data has inherent limitations, such as the scarcity of annotated medical datasets and the complexity of medical images (including factors like noise, artifacts, and low contrast). Secondly, in terms of cross-modal fusion, there exists an intricate relationship between medical images and textual information. The complexity of the relationship between medical images and medical text, as well as the incorporation of global contextual details, pose challenges. To address these issues, this paper proposes the Feature-Enhanced Network with Conditional Mixing. This innovative approach aims to overcome the scarcity and complexity of medical images without the need for additional datasets. Simultaneously, to tackle the challenge of cross-modal fusion and information interaction and to address the complex relationship between medical images and text, we propose the Cross-Adaptive Fusion Network. Experimental results conducted on the VQA-RAD dataset demonstrate noticeable improvements in accuracy compared to mainstream Med-VQA models.

关键词： visualization Accuracy Noise Neural networks Medical services Question answering (information retrieval) data models

来源：评论

学校读者我要写书评

暂无评论

Opinion-Tree-guided Contrastive Learning for Aspect Sentiment Quadruple Prediction

Opinion-Tree-guided Contrastive Learning for Aspect Sentimen...

引用

International Conference on Computer Supported Cooperative Work in Design

作者： Zhijun Li Zhenyu Yang Yiwen Li Xiaoyang Li Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center Qilu University of Technology (Shandong Academy of Sciences) Jinan China Shandong Engineering Research Center of Big Data Applied Technology Faculty of Computer Science and Technology Qilu University of Technology (Shandong Academy of Sciences) Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China

ISBN: (数字)9798350349184

ISBN: (纸本)9798350349191

In recent years, aspect sentiment quadruple prediction (ASQP) has become a popular task in the field of aspect-based sentiment analysis. When modeling comment statements, it is often necessary to consider structural information between sentiment labels. Existing approaches categorize these labels by encoding text and dependency tree structures separately and then incorporating their representations, but this approach fails to effectively capture the structural information between sentiment labels. To address this problem, we propose a method for ASQP called Opinion-Tree-guided Contrastive Learning (OTCL), which embeds structural information directly into the text instead of modeling it separately. Specifically, OTCL utilizes the opinion tree structure to guide the construction of positive samples of the input text during training. By associating the input text with its positive samples, the text encoder can learn to generate textual representations associated with the opinion tree structure. We validate the effectiveness of our method on two commonly used datasets.

关键词： Training Sentiment analysis Federated learning Contrastive learning Predictive models Encoding Task analysis

来源：评论

学校读者我要写书评

暂无评论

Denial-of-Service or Fine-Grained Control: Towards Flexible Model Poisoning Attacks on Federated Learning

arXiv

引用

arXiv 2023年

作者： Zhang, Hangtao Yao, Zeming Zhang, Leo Yu Hu, Shengshan Chen, Chao Liew, Alan Li, Zhetao School of Cyber Science and Engineering Huazhong University of Science and Technology China Swinburne University of Technology Australia Griffith University Australia National Engineering Research Center for Big Data Technology and System China Services Computing Technology and System Lab. Hubei Key Laboratory of Distributed System Security China Hubei Engineering Research Center on Big Data Security China RMIT University Australia Xiangtan University China

Federated learning (FL) is vulnerable to poisoning attacks, where adversaries corrupt the global aggregation results and cause denial-of-service (DoS). Unlike recent model poisoning attacks that optimize the amplitude of malicious perturbations along certain prescribed directions to cause DoS, we propose a Flexible Model Poisoning Attack (FMPA) that can achieve versatile attack goals. We consider a practical threat scenario where no extra knowledge about the FL system (e.g., aggregation rules or updates on benign devices) is available to adversaries. FMPA exploits the global historical information to construct an estimator that predicts the next round of the global model as a benign reference. It then fine-tunes the reference model to obtain the desired poisoned model with low accuracy and small perturbations. Besides the goal of causing DoS, FMPA can be naturally extended to launch a fine-grained controllable attack, making it possible to precisely reduce the global accuracy. Armed with precise control, malicious FL service providers can gain advantages over their competitors without getting noticed, hence opening a new attack surface in FL other than DoS. Even for the purpose of DoS, experiments show that FMPA significantly decreases the global accuracy, outperforming six state-of-the-art attacks. © 2023, CC BY.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

FaceBench: A Multi-View Multi-Level Facial Attribute VQA dataset for Benchmarking Face Perception MLLMs

arXiv

引用

arXiv 2025年

作者： Wang, Xiaoqin Ma, Xusen Hou, Xianxu Ding, Meidan Li, Yudong Chen, Junliang Chen, Wenting Peng, Xiaoyang Shen, Linlin Computer Vision Institute College of Computer Science and Software Engineering Shenzhen University China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University China Guangdong Provincial Key Laboratory of Intelligent Information Processing China AIAC Xi’an Jiaotong-Liverpool University China Tsinghua University China The Hong Kong Polytechnic University Hong Kong City University of Hong Kong Hong Kong Sun Yat-sen University China

Multimodal large language models (MLLMs) have demonstrated remarkable capabilities in various tasks. However, effectively evaluating these MLLMs on face perception remains largely unexplored. To address this gap, we introduce FaceBench, a dataset featuring hierarchical multi-view and multi-level attributes specifically designed to assess the comprehensive face perception abilities of MLLMs. Initially, we construct a hierarchical facial attribute structure, which encompasses five views with up to three levels of attributes, totaling over 210 attributes and 700 attribute values. Based on the structure, the proposed FaceBench consists of 49,919 visual question-answering (VQA) pairs for evaluation and 23,841 pairs for fine-tuning. Moreover, we further develop a robust face perception MLLM baseline, Face-LLaVA, by training with our proposed face VQA data. Extensive experiments on various mainstream MLLMs and Face-LLaVA are conducted to test their face perception ability, with results also compared against human performance. The results reveal that, the existing MLLMs are far from satisfactory in understanding the fine-grained facial attributes, while our Face-LLaVA significantly outperforms existing open-source models with a small amount of training data and is comparable to commercial ones like GPT-4o and Gemini. The dataset will be released at https://***/CVI-SZU/FaceBench. Copyright © 2025, The Authors. All rights reserved.

关键词：

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：