检索结果-内蒙古大学图书馆

Attack based on data: a novel perspective to attack sensitive points directly

Cybersecurity 2024年第3期7卷 111-123页

作者： Yuyao Ge Zhongguo Yang Lizhe Chen Yiming Wang Chengyang Li School of Information Science and Technology North China University of TechnologyBeijingChina Bejing Key Laboratory on Integration and Analysis of Large-Scale Stream Data North China University of TechnologyBeijingChina School of Computer Science Peking UniversityBeijingChina

Adversarial attack for time-series classification model is widely explored and many attack methods are *** there is not a method of attack based on the data *** this paper,we innovatively proposed a black-box sparse attack method based on data *** method directly attack the sensitive points in the time-series data accord-ing to statistical features extract from the *** frst,we have validated the transferability of sensitive points among DNNs with different ***,we use the statistical features extract from the dataset and the sensi-tive rate of each point as the training set to train the predictive ***,predicting the sensitive rate of test set by predictive ***,perturbing according to the sensitive *** attack is limited by constraining the LO norm to achieve one-point *** conduct experiments on several datasets to validate the effectiveness of this method.

关键词： Black-box adversarial attack Time series classification data mining

来源：评论

学校读者我要写书评

暂无评论

Improving Entity Linking in Chinese Domain by Sense Embedding Based on Graph Clustering

引用

Journal of computer science & Technology 2023年第1期38卷 196-210页

作者：张照博钟芷漫袁平鹏金海 National Engineering Research Center for Big Data Technology and System Huazhong University of Science and Technology Wuhan 430074China Service Computing Technology and System Laboratory Huazhong University of Science and Technology Wuhan 430074China Cluster and Grid Computing Laboratory Huazhong University of Science and TechnologyWuhan 430074China School of Computer Science and Technology Huazhong University of Science and TechnologyWuhan 430074China

Entity linking refers to linking a string in a text to corresponding entities in a knowledge base through candidate entity generation and candidate entity *** is of great significance to some NLP(natural language processing)tasks,such as question *** English entity linking,Chinese entity linking requires more consideration due to the lack of spacing and capitalization in text sequences and the ambiguity of characters and words,which is more evident in certain *** Chinese domains,such as industry,the generated candidate entities are usually composed of long strings and are heavily *** addition,the meanings of the words that make up industrial entities are sometimes *** semantic space is a subspace of the general word embedding space,and thus each entity word needs to get its exact ***,we propose two schemes to achieve better Chinese entity ***,we implement an ngram based candidate entity generation method to increase the recall rate and reduce the nesting ***,we enhance the corresponding candidate entity ranking mechanism by introducing sense *** the contradiction between the ambiguity of word vectors and the single sense of the industrial domain,we design a sense embedding model based on graph clustering,which adopts an unsupervised approach for word sense induction and learns sense representation in conjunction with *** test the embedding quality of our approach on classical datasets and demonstrate its disambiguation ability in general *** confirm that our method can better learn candidate entities’fundamental laws in the industrial domain and achieve better performance on entity linking through experiments.

关键词： natural language processing(NLP) domain entity linking computational linguistics word sense disambiguation knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Octopus: Experiences with a Hybrid Event-Driven Architecture for Distributed Scientific Computing

Octopus: Experiences with a Hybrid Event-Driven Architecture...

引用

2024 Workshops of the International Conference for High Performance Computing, Networking, Storage and Analysis, SC Workshops 2024

作者： Pan, Haochen Chard, Ryan Zhou, Sicheng Kamatar, Alok Vescovi, Rafael Hayot-Sasson, Valerie Bauer, Andre Gonthier, Maxime Chard, Kyle Foster, Ian Southern University of Science and Technology Department of Computer Science China Illinois Institute of Technology Department of Computer Science ChicagoIL United States University of Chicago Department of Computer Science ChicagoIL United States Argonne National Laboratory Data Science and Learning Division LemontIL United States

ISBN: (纸本)9798350355543

Scientific research increasingly relies on distributed computational resources, storage systems, networks, and instruments, ranging from HPC and cloud systems to edge devices. Event-driven architecture (EDA) benefits applications targeting distributed research infrastructures by enabling the organization, communication, processing, reliability, and security of events generated from many sources. To support the development of scientific EDA, we introduce Octopus, a hybrid, cloud-to-edge event fabric designed to link many local event producers and consumers with cloud-hosted brokers, and to provide a fabric for developing resilient applications. Octopus can be scaled to meet demand, permits the deployment of highly available Triggers for automatic event processing, and enforces fine-grained access control. We identify requirements in self-driving laboratories, scientific data automation, online task scheduling, epidemic modeling, and dynamic workflow management use cases, and present results demonstrating Octopus' ability to meet those requirements. Octopus supports producing and consuming events at a rate of over 4.2 M and 9.6 M events per second, respectively, from distributed clients. © 2024 IEEE.

关键词： Event-driven architecture research automation scientific computing

来源：评论

学校读者我要写书评

暂无评论

Enhanced U-Net with Transformer-Driven Encoder for Medical Image Segmentation

Enhanced U-Net with Transformer-Driven Encoder for Medical I...

引用

2024 International Conference on Brain computer Interface and Healthcare Technologies, iCon-BCIHT 2024

作者： Palaniappan, Ramya Surendran, Suseendran Balakrishnan, Saranya Madanapalle Institute of Technology & Science Department of Computer Science & Engineering Madanapalle India Bannari Amman Institute of Technology Department of Computer Science and Engineering Erode India KGiSL Institute of Technology Department of Artificial Intelligence and Data science Coimbatore India

ISBN: (纸本)9798331540067

The automatic segmentation of tumours or lesions from magnetic resonance imaging (MRI) pictures is a critical but difficult task in clinical settings, sometimes necessitating laborious and time-consuming techniques. Deep learning has made a significant advancement in the medical imaging field in recent years. U-Net has emerged as the predominant model used for segmenting medical images. Although the standard U-Net framework has shown excellent performance in segmenting multimodal medical pictures, it has been noticed that it lags in certain aspects when dealing with tough datasets. We propose specific improvements to the state-of-the-art U-Net model. We have introduced attention block in the skipped connections to highlight relevant features. Additionally, we have incorporated transformers in the encoding path to capture temporal dependencies in medical images. In this research work, we have introduced an innovative architecture that has the potential to replace the U-Net model for segmenting regions of interest (ROI) in medical images. This study utilizes a total of 1760 samples obtained from the ISBI 2015 and PH2 medical repositories. The experimental findings have demonstrated that our proposed model has obtained highly encouraging results of 78.02 Jaccard coefficient and IoU of 65.14. In addition, we conducted tests and compared the performance of the proposed model with the classical U-Net model using the same dataset. As a result, we observed a relative improvement in the performance of both 2.36 and 1.85, respectively. The proposed technology will function as a sophisticated toolkit for radiologists to precisely detect and diagnose anomalous tumour cells. © 2024 IEEE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Comparative Analysis of Hate Speech Detection Using Various Vectorization and Classification Techniques 2

Comparative Analysis of Hate Speech Detection Using Various ...

引用

2nd International Conference on Artificial Intelligence, Blockchain, and Internet of Things, AIBThings 2024

作者： Tali, Sophia Noguera, Daniel Gargees, Rasha S. Central Michigan University Department of Data Science Actuarial Science and Statistics MI United States Central Michigan University Department of Computer Science MI United States

ISBN: (纸本)9798331529529

Hate speech is a prominent, growing problem within our society, and this problem and its effects have only increased with time. According to the Center for Technology and Society, 52% of people reported being harassed online between 2022 and 2023, an increase from 40% in 2022. The new technological advancements make it easier than ever to spread negative, discriminatory, and pejorative comments to attack others with the click of a button. This is unfortunately an issue that affects many and often leads to violence, which is why it is important to identify and remove this online hate speech in order to minimize the spread. In this study, we explore the classification of online hate speech by comparing different vectorization and classification methods. We developed a set of eight classification techniques for identifying hate from a Twitter dataset containing hate speech and benign text examples. We cleaned and vectorized it in two different ways. Each of these sets is used to train deep learning, SVM, SGD, and random forest classifiers. All these methods achieved accuracy above 93%. The machine learning methods (SVM, SGD, and random forest) all achieved a nearperfect classification accuracy of 98.72%, with slight variations in precision and recall. The deep learning approach also showed improved performance with TF-IDF vectorization. Overall, the choice of vectorization had minimal impact on accuracy, and all models performed consistently well. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

A Plastic Waste Hotspot Detection System and Its Application to Measuring Environmental Improvement 1

A Plastic Waste Hotspot Detection System and Its Application...

引用

1st IEEE International Symposium on Consumer Technology, ISCT 2024

作者： Sasaki, Shiori Nakamura, Ryota Wibowo, Prasetyo Musashino University Department of Data Science Tokyo Japan Musashino University Faculty of Data Science Tokyo Japan Politeknik Elektronika Negeri Department of Informatics and Computer Engineering Surabaya Indonesia

ISBN: (纸本)9798350365191

This paper describes a Plastic Waste Hotspot Detection System which has been developed in an international collaborative research project to realize an 'Environmental AI-Human Actions integration' with marine plastic waste reduction actions. In this project, various local conditions are obtained to the system to realize an environmentally resilient society. The nature conservation activities such as marine plastic waste collection and reduction actions are implemented based on these conditions. In the process of integration of Environmental AI and Human Actions, the effects of these activities and activists' contribution to environmental improvement and global-scale knowledge are accumulated, shared, analyzed, and visualized through global-scale analysis. This paper describes the Plastic Waste Hotspot Detection system as one of the systems that integrate AI systems and human actions. This paper examines the feasibility and effectiveness of the system through the fieldwork tests in Phuket, Thailand and Surabaya, Indonesia. © 2024 IEEE.

关键词： Plastic recycling

来源：评论

学校读者我要写书评

暂无评论

Toward Handling Uncertainty-At-Source in AI - A Review and Next Steps for Interval Regression

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第1期5卷 3-22页

作者： Kabir, Shaily Wagner, Christian Ellerby, Zack Lab for Uncertainty in Data and Decision Making School of Computer Science University of Nottingham LUCID NottinghamNG7 2RD United Kingdom

Most of statistics and AI draw insights through modeling discord or variance between sources (i.e., intersource) of information. Increasingly however, research is focusing on uncertainty arising at the level of individual measurements (i.e., within- or intrasource), such as for a given sensor output or human response. Here, adopting intervals rather than numbers as the fundamental data-type provides an efficient, powerful, yet challenging way forward - offering systematic capture of uncertainty-at-source, increasing informational capacity, and ultimately potential for additional insight. Following progress in the capture of interval-valued data in particular from human participants, conducting machine learning directly upon intervals is a crucial next step. This article focuses on linear regression for interval-valued data as a recent growth area, providing an essential foundation for broader use of intervals in AI. We conduct an in-depth analysis of state-of-the-art methods, elucidating their behavior, advantages, and pitfalls when applied to synthetic and real-world datasets with different properties. Specific emphasis is given to the challenge of preserving mathematical coherence, i.e., models maintain fundamental mathematical properties of intervals. In support of real-world applicability of the regression methods, we introduce and demonstrate a novel visualization approach, the interval regression graph (IRG), which effectively communicates the impact of both position and range of variables within the regression models - offering a leap in their interpretability. Finally, this article provides practical recommendations concerning regression-method choice for interval data and highlights remaining challenges and important next steps for developing AI with the capacity to handle uncertainty-at-source. © 2020 IEEE.

关键词： data visualization

来源：评论

学校读者我要写书评

暂无评论

Drug Recommendation based on drug details and Optimization using Bee Algorithm

Drug Recommendation based on drug details and Optimization u...

引用

2024 IEEE International Conference on Cybernetics and Innovations, ICCI 2024

作者： Ariyachaipong, Kornpong Senaluang, Pattayapon Songmuang, Pokpong Kongkachandra, Rachada Thammasat Universtiy Data Science And Innovation College Of Interdisciplinary Studies Pathum Thani Thailand Thammasat Universtiy Department Of Computer Science Bangkok Thailand Thammasat Universtiy Data Science And Innovation College Of Interdisciplinary Studies Bangkok Thailand

ISBN: (纸本)9798350370096

In an effort to bolster the healthcare system in Thailand, particularly in remote areas with limited access to pharmacists, this study proposes a novel drug recommendation system based on drug details. This system aims to address the challenge of medication selection in resource-constrained settings by providing physicians with informed recommendations tailored to patient *** proposed system utilizes drug data sourced from ***, encompassing approximately 34,284 drug entries. To ensure high-quality recommendations, the data undergoes a rigorous preprocessing phase involving null value imputation and feature extraction using techniques like TF-IDF. Following preprocessing, a combination of Drug Recommendation with cosine similarity and the Bee Algorithm is employed. Cosine similarity establishes a baseline for drug similarity, while the Bee Algorithm optimizes the selection process by considering additional factors beyond just similarity, such as potential side effects and cost-effectiveness. © 2024 IEEE.

关键词： Recommender systems

来源：评论

学校读者我要写书评

暂无评论

Synthetically Augmented Self-Supervised Fine-Tuning for Diverse Text OCR Correction 27

Synthetically Augmented Self-Supervised Fine-Tuning for Dive...

引用

27th European Conference on Artificial Intelligence, ECAI 2024

作者： Guan, Shuhao Greene, Derek Insight Centre for Data Analytics Dublin School of Computer Science University College Dublin Ireland

ISBN: (纸本)9781643685489

The adoption of Optical Character Recognition (OCR) tools has been central to the increased digitization of historical documents. However, the errors introduced during OCR, particularly in texts with a specialized vocabulary (SV), necessitate effective post-OCR correction methodologies. This study introduces a novel approach that leverages weak supervision and self-supervised finetuning to enhance post-OCR correction without the need for substantial manual annotations. By using multi-noise-level synthetic data, generated through automatically-extracted OCR errors and applied to clean texts, we can train robust models tailored for post-OCR tasks. Furthermore, we propose a unique self-supervised fine-tuning strategy, applied specifically to long texts, enables models to adeptly handle out-of-vocabulary problems and SV. Additionally, we tested the performance of the GPT model on post-OCR tasks. © 2024 The Authors.

关键词： Optical character recognition

来源：评论

学校读者我要写书评

暂无评论

Layer-Wise Learning Rate Optimization for Task-Dependent Fine-Tuning of Pre-Trained Models: An Evolutionary Approach

引用

ACM Transactions on Evolutionary Learning and Optimization 2024年第4期4卷 1-23页

作者： Bu, Chenyang Liu, Yuxin Huang, Manzong Shao, Jianxuan Ji, Shengwei Luo, Wenjian Wu, Xindong Key Laboratory of Knowledge Engineering with Big Data Ministry of Education and School of Computer Science and Information Engineering Hefei University of Technology Hefei China School of Artificial Intelligence and Big Data Hefei University Hefei China Guangdong Provincial Key Laboratory of Novel Security Intelligence Technologies School of Computer Science and Technology Harbin Institute of Technology Shenzhen China

The superior performance of large-scale pre-Trained models, such as Bidirectional Encoder Representations from Transformers (BERT) and Generative Pre-Trained Transformer (GPT), has received increasing attention in both academic and industrial research and has become one of the current research hotspots. A pre-Trained model refers to a model trained on large-scale unlabeled data, whose purpose is to learn general language representation or features for fine-Tuning or transfer learning in subsequent tasks. After pre-Training is complete, a small amount of labeled data can be used to fine-Tune the model for a specific task or domain. This two-stage method of "pre-Training+fine-Tuning"has achieved advanced results in natural language processing (NLP) tasks. Despite widespread adoption, existing fixed fine-Tuning schemes that adapt well to one NLP task may perform inconsistently on other NLP tasks given that different tasks have different latent semantic structures. In this article, we explore the effectiveness of automatic fine-Tuning pattern search for layer-wise learning rates from an evolutionary optimization perspective. Our goal is to use evolutionary algorithms to search for better task-dependent fine-Tuning patterns for specific NLP tasks than typical fixed fine-Tuning patterns. Experimental results on two real-world language benchmarks and three advanced pre-Training language models show the effectiveness and generality of the proposed framework. © 2024 held by the owner/author(s).

关键词： Semantics

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：