检索结果-内蒙古大学图书馆

9th IEEE Smart World Congress, SWC 2023

作者： Wang, Zumin Xu, Chang Chen, Liming Qin, Jing Ji, Changqing School of Information Engineering Dalian University Dalian China School of Computing Ulster University Belfast United Kingdom School of Software Engineering Dalian University Dalian China School of Physical Science and Technology School of Information Engineering Dalian University Dalian China

ISBN: (纸本)9798350319804

The improvement of intention classification of medical query can effectively improve the performance of search engines and Question-Answering systems, and further improve medical services. The Pre-Trained language models has achieved good performance in various NLP tasks, but the slow reasoning speed and high requirements for storing computing resources of the model make it difficult for medical institutions to deploy relevant models. In this paper, we propose a knowledge distillation model based on ERNIE to solve the Chinese medical intention classification task, in order to provide a Small-Scale and more efficient intention classification model. We propose a Two-Stage framework, which is distilled on specific domain knowledge and tasks, and we use data augmentation methods to minimize the loss of accuracy caused by model compression. Specifically, we distill the Embedding layer and Transformer layer of the teacher model in the Domain-Knowledge-Specific-Distillation stage, so that our student model can better capture the general domain and Medical specific domain knowledge of the teacher model. In the Intent-Classification Task-specific Distillation stage, we also distilled the Prediction layer and added Soft-Label and Hard-Label to calculate the Prediction layer loss to ensure that the student model acquires knowledge of specific tasks. Considering the lack of labeled data available for training in the medical field and intention classification task, we use word exchange and whole entity masking methods to augment the labeled data and augment the generalization ability of student models. We conducted experiments on a publicly available CBLUE dataset, and the experimental results showed that our proposed 4-layer student model retained more than 98.6% of the language comprehension capabilities of the original 12-layer teacher model - ERNIE, while reducing the computational resources required and significantly accelerating reasoning speed. © 2023 IEEE.

关键词： Intent Classification Knowledge Distillation Medical NLP Pre-Trained Language Models

来源：评论

学校读者我要写书评

暂无评论

FedCSA: Enhancing Federated Unlearning Efficiency Through Adaptive Clustering Under Data Heterogeneity

引用

Chinese Journal of Electronics 2025年第3期34卷 970-979页

作者： Yang Yang Wang Zhen Alghazzawi Daniyal M Cheng Li Liu Gaoyang Wang Chen Zeng Cheng Li Yuying Key Laboratory of Intelligent Sensing System and Security (Ministry of Education) School of Artificial Intelligence Hubei University Wuhan 430000 China Faculty of Computing and Information Technology King Abdulaziz University Jeddah 22427 Saudi Arabia Hubei Key Laboratory of Smart Internet Technology Huazhong University of Science and Technology Wuhan 430000 China

In the digital era, escalating concerns over personal privacy and social security have led to the advocacy for the “right to be forgotten”, a principle that empowers individuals to request the deletion of their personal data from online platforms. Consequently, machine unlearning (MU) has been proposed as a method for targeted data deletion within machine learning models. However, MU encounters difficulties in distributed learning environments, such as federated learning (FL), where direct access to data is restricted. Federated unlearning (FU) has been developed in response, aiming to facilitate the process of data deletion requests from clients within FL frameworks. Despite advancements, FU methods based on approximate unlearning present a risk of potential data breaches, while methods reliant on retraining necessitate either complete or repeated retraining of clients, which is inefficient. Addressing these challenges, we introduce the federated cluster slicing algorithm (FedCSA), a novel FU strategy that achieves precision and efficiency in data unlearning. FedCSA organizes clients into distinct slices based on model deviation values, facilitating targeted retraining of local models upon unlearning requests. This method not only ensures consistency in the independent and identically distributed degree across slices but also improves unlearning efficiency and maintains global model accuracy. Moreover, FedCSA features an adaptive clustering mechanism that autonomously determines the optimal number of slices, optimizing the unlearning process. Our empirical analysis, conducted across the MNIST, Fashion-MNIST, and CIFAR-10 datasets, underscores FedCSA’s superior performance. FedCSA exhibits a fourfold increase in unlearning efficiency compared to traditional retraining methods. Furthermore, when juxtaposed with the sharded, isolated, sliced, and aggregated technique, FedCSA demonstrates a 4%–5% enhancement in global model accuracy. These findings corroborate the effi

关键词： Federated learning Federated unlearning Machine unlearning

来源：评论

学校读者我要写书评

暂无评论

Cyberbullying Detection Using Deep Learning: A Comparative Study 2

Cyberbullying Detection Using Deep Learning: A Comparative S...

引用

2nd International Conference on Cyber Resilience, ICCR 2024

作者： Alkasassbeh, Mouhammd Almomani, Ammar Aldweesh, Amjad Al-Qerem, Ahmad Alauthman, Mohammad Nahar, Khalid Mago, Beenu Princess Sumaya University For Technology Computer Science Department Amman Jordan School Of Computing Skyline University College Sharjah United Arab Emirates AlHuson University College AlBalqa Applied University I.T. Department Irbid Jordan College Of Computing And It Shaqra University Shaqra Saudi Arabia Zarqa University Faculty Of Information Technology Computer Science Department Zarqa13110 Jordan University Of Petra Faculty Of Information Technology Department Of Information Security Amman Jordan Yarmouk University Faculty Of I.T Computer Science Department Irbid Jordan School Of Computing Skyline University College University City Of Sharjah Sharjah United Arab Emirates

ISBN: (纸本)9798350394962

Cyberbullying has become a serious issue with the rise in social media usage, especially among adolescents. Victims of cyberbullying often suffer from psychological distress, anxiety, depression and even suicidal tendencies. With the rapid growth of user-generated content online, manual detection of cyberbullying is extremely difficult. Hence, automated cyber-bullying detection systems are needed. In recent years, deep learning models have emerged as a promising technique for cyberbullying detection. This paper provides a comprehensive review of the use of deep learning for cyberbullying detection. The key contributions are: (1) a taxonomy of cyberbullying detection techniques, (2) a review of datasets used for model training and evaluation, (3) a comparative analysis of various deep learning models including CNN, RNN, LSTM, GRU and Transformer- based models, (4) a discussion of key challenges and limitations, and (5) suggestions for future research directions. The paper provides a critical perspective on model performance across multiple datasets. Key findings indicate RNNs and specifically LSTM and GRU models demonstrate state-of-the-art performance for sequential cyberbullying detection. The study concludes that deep ensemble models combining LSTMs and pretrained contextual embeddings show significant promise to advance the field. © 2024 IEEE.

关键词： Social networking (online)

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Learning Approach to Solve Dynamic Bi-objective Police Patrol Dispatching and Rescheduling Problem 32

Reinforcement Learning Approach to Solve Dynamic Bi-objectiv...

引用

32nd International Conference on Automated Planning and Scheduling, ICAPS 2022

作者： Joe, Waldy Lau, Hoong Chuin Pan, Jonathan School of Computing and Information Systems Singapore Management University Singapore Home Team Science and Technology Agency Singapore

ISBN: (纸本)9781577358749

Police patrol aims to fulfill two main objectives namely to project presence and to respond to incidents in a timely manner. Incidents happen dynamically and can disrupt the initially-planned patrol schedules. The key decisions to be made will be which patrol agent to be dispatched to respond to an incident and subsequently how to adapt the patrol schedules in response to such dynamically-occurring incidents whilst still fulfilling both objectives;which sometimes can be conflicting. In this paper, we define this real-world problem as a Dynamic Bi-Objective Police Patrol Dispatching and Rescheduling Problem and propose a solution approach that combines Deep Reinforcement Learning (specifically neural networks-based Temporal-Difference learning with experience replay) to approximate the value function and a rescheduling heuristic based on ejection chains to learn both dispatching and rescheduling policies jointly. To address the dual objectives, we propose a reward function that implicitly tries to maximize the rate of successfully responding to an incident within a response time target while minimizing the reduction in patrol presence without the need to explicitly set predetermined weights for each objective. The proposed approach is able to compute both dispatching and rescheduling decisions almost instantaneously. Our work serves as the first work in the literature that takes into account these dual patrol objectives and real-world operational consideration where incident response may disrupt existing patrol schedules. © 2022, Association for the Advancement of Artificial Intelligence.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Multiclass Classification of Brain Tumor for MR Images Using Shallow Autoencoder Based Neural Network 2

Multiclass Classification of Brain Tumor for MR Images Using...

引用

2nd IEEE International Conference on Mobile Networks and Wireless Communications, ICMNWC 2022

作者： Jyothi, Parvathy Dhanasekaran, S. Singh, Robert A School of Computing Kalasalingam Academy of Research and Education Krishnankoil India Kalasalingam Academy of Research and Education Department of Information Technology Krishnankoil India School of Computing SRM Institute of Science and Technology Chennai India

ISBN: (纸本)9781665491112

Brain tumor is an abnormal growth of cells, that may be cancerous or non-cancerous. The earlier prediction, identification, and classification of tumor is essential for rapid diagnosis. In brain MRI, the size and location of tumors can be diverse for different patients. Because of the increased flow of patients in scan centers, patients must now wait for a long time to collect their reports from the radiologists, as it ends up taking the radiologists a long time to classify the images. The proposed methodology in this work can classify tumors from MR brain images into three categories. At first, a shallow autoencoder network is designed for image reconstruction. The encoder segment is made up of three convolutional layers, and in decoder segment, four layers are used for reconstruction. Autoencoder offer excellent noise robustness and feature reduction thereby reducing the possibility of over-fitting. Secondly, to perform classification, an additional convolutional layer is added to the encoder part of neural network along with 2×2 filter. The features extracted from the encoder part were given to a single layer dense neural network and finally testing is performed on SoftMax layer for the classification. The developed algorithm was trained and evaluated on the Cheng dataset, and achieved an accuracy of 95.26%. The developed methods' outcomes outperform well than the conventional techniques. © 2022 IEEE.

关键词： Tumors

来源：评论

学校读者我要写书评

暂无评论

Secure Multi-Keyword Retrieval with Integrity Guarantee for Outsourced ADS-B Data in Clouds 29

Secure Multi-Keyword Retrieval with Integrity Guarantee for ...

引用

29th IEEE International Conference on Parallel and Distributed Systems, ICPADS 2023

作者： Yang, Shangru Ding, Yong Wang, Yujue Liang, Hai Yang, Changsong Wang, Huiyong Guilin University of Electronic Technology Guangxi Key Laboratory of Cryptography and Information Security School of Computer Science and Information Security Guilin China Hkct Institute for Higher Education Institute of Cyberspace Technology Hong Kong Hong Kong Hangzhou Innovation Institute of Beihang University Hangzhou China Guilin University of Electronic Technology School of Mathematics and Computing Science Guilin China

ISBN: (纸本)9798350330717

With the development of Automatic Dependent Surveillance-Broadcast (ADS-B) technology in the aviation industry, a large amount of data are generated in ADS-B systems everyday. Cloud storage can satisfy the requirement of storing a large amount of ADS-B data well, however, the data stored on cloud server also raises problems about the security and confidentiality. This paper proposes a secure ADS-B data outsourcing scheme with integrity guarantee and multi-keyword retrieval (MRDP) to address these issues. In our MRDP, an index is generated for each piece of ADS-B data through Bloom filters to enable multi-keyword retrieval from clouds. Also, a unique label is produced for each piece of ADS-B data for integrity verification of query results. Our MRDP solution enjoys completeness property in that all query results with regard to the multiple keywords would be returned by the cloud server, otherwise it could be detected by the user. Security analysis indicated that our scheme offers integrity, completeness and privacy protection on outsourced ADS-B data, under the computational Diffie-Hellman (CDH) assumption and the discrete logarithm (DL) assumption. Theoretical and experimental analyses demonstrate the practicality of our proposed MRDP construction compared to existing solutions. © 2023 IEEE.

关键词： ADS-B systems Cloud storage Integrity audit Multi-keyword search

来源：评论

学校读者我要写书评

暂无评论

UZNER: A Benchmark for Named Entity Recognition in Uzbek 12th

UZNER: A Benchmark for Named Entity Recognition in Uzbek

引用

12th National CCF Conference on Natural Language Processing and Chinese computing, NLPCC 2023

作者： Yusufu, Aizihaierjiang Jiang, Liu Ainiwaer, Abidan Teng, Chong Yusufu, Aizierguli Li, Fei Ji, Donghong Key Laboratory of Aerospace Information Security and Trusted Computing Ministry of Education School of Cyber Science and Engineering Wuhan University Wuhan China School of Information Management Wuhan University Wuhan China School of Computer Science and Technology Xinjiang Normal University Urumqi China

ISBN: (纸本)9783031446924

Named entity recognition (NER) is a key task in natural language processing, and entity recognition can provide necessary semantic information for many downstream tasks. However, the performance of NER is often limited by the richness of language resources. For low-resource languages, NER usually performs poorly due to the lack of sufficient labeled data and pre-trained models. To address this issue, we manually constructed a large-scale, high-quality Uzbek NER corpus of Uzbek, and experimented with various NER methods. We improved state-of-the-art baseline models by introducing additional features and data translations. Data translation enables the model to learn richer syntactic structure and semantic information. Affix features provide knowledge at the morphological level and play an important role in identifying oversimplified low-frequency entity labels. Our data and models will be available to facilitate low-resource NER. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

A Nature-Inspired Energy-Efficient Clustering Protocol Based on Improved Hybrid Grey Wolf Optimization for FANETs 7

A Nature-Inspired Energy-Efficient Clustering Protocol Based...

引用

7th IEEE International Conference on Electronics and Communication Engineering, ICECE 2024

作者： Xu, Mengying Zu, Yunxiao Zhou, Jie Dai, Jiapeng Chen, Jing Liu, Meiru School of Electronic Engineering Beijing University of Posts and Telecommunications Beijing China Xinjiang Production and Construction Corps Key Laboratory of Computing Intelligence and Network Information Security Shihezi China Shihezi University Department of Information Science and Technology Shihezi China

ISBN: (纸本)9798331517755

Flying ad hoc networks (FANETs) are composed of multiple UA V nodes that are distributed across a wide area and linked via wireless communication to gather and transmit environmental data. However, the battery-powered UAVs in FANETs imposes significant energy limitations, thereby shortening the network's operational lifespan. This paper proposes an energy-efficient clustering protocol based on improved hybrid grey wolf optimization (EEC-IHGWO), which can optimally select cluster heads (CHs), significantly extend network lifetime and maximize energy efficiency. The selection criteria include factors like remaining energy, the distance from cluster members to CHs, and node density. The performance of the EEC-IHGWO method, measured by network lifetime, the number of living nodes, and remaining energy, is assessed against traditional algorithms such as LEACH, as well as the more recent PSO and FCM. The simulation results demonstrate that EEC-IHGWO sustains a longer network lifetime and higher energy efficiency. © 2024 IEEE.

关键词： Ad hoc networks

来源：评论

学校读者我要写书评

暂无评论

VisTA: vision transformer-attention enhanced CNN ensemble for optimized classification of acute lymphoblastic leukemia benign and progressive malignant stages

引用

International Journal of information technology (Singapore) 2024年 1-18页

作者： Nunna, Hasmitha Krishna Altable, Ali Gundala, Pallavi Rangarajan, Prasanna Kumar Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Chennai India Department of Computer Engineering Prince Sattam bin Abdulaziz University Al-Kharj 11942 Saudi Arabia Faculty of Computing and Information Technology King Abdulaziz University Jeddah 21589 Saudi Arabia

Acute lymphoblastic leukemia (ALL), a rare blood disorder marked by excessive immature lymphocyte proliferation in the bone marrow, primarily affects children but poses a more challenging prognosis in adults due to late detection. Pathologists employ bone marrow morphology and immunophenotyping to establish cell lineage based on antigens to diagnose acute lymphoblastic leukemia. These manual approaches are time-consuming and need specialized expertise. It is critical for the decision support system to train an accurate classification model to discriminate between mature and immature white blood cells along with the stratification of its phases which is a crucial component in diagnostic accuracy. Another major challenge in ALL classification arises from the morphological resemblance between leukemic B-lymphoblast cells and normal B-lymphoid progenitors. To improve the detection of leukemia, this work integrates an ensemble model of Vision transformer with attention-based pre-trained architectures that use high-resolution blood smear images, to accelerate the prompt detection of this insidious illness and its stages. The ensemble model, which combines ViT with attention-based CNNs, uses dual feature extraction methods from cell images to improve classification results. Our suggested model demonstrates promising performance with an accuracy of 99.96% on a test dataset with 1600 images. The suggested model provides a promising methodology that could be extremely useful in medical settings, allowing for the accurate classification and diagnosis of acute lymphoblastic leukemia. © Bharati Vidyapeeth's Institute of Computer Applications and Management 2024.

关键词： Acute lymphoblastic leukemia classification Attention mechanism Benign Blood smears CLAHE Ensemble model Leukemia detection Malignant Vision transformers

来源：评论

学校读者我要写书评

暂无评论

Personalized Federated Learning with Collaborative Aggregation Networks for Multi-Site Brain Disorder Diagnosis 4

Personalized Federated Learning with Collaborative Aggregati...

引用

4th International Conference on Industrial Automation, Robotics and Control Engineering, IARCE 2024

作者： Si, Qian Li, Yang School of Cyber Science and Technology Beihang University Beijing China Department of Automation Science and Electrical Engineering The Beijing Advanced Innovation Center for Big Data and Brain Computing State Key Laboratory of Virtual Reality Technology and Systems Beijing China Advanced Institute of Information Technology Peking University Beijing China Beihang University Beijing China

ISBN: (纸本)9798350380323

In multi-site brain disease diagnosis studies, traditional centralized training methods necessitate sharing medical data, posing significant privacy risks. Federated learning (FL) offers a privacy-preserving solution by enabling global model training through aggregating locally trained models from multiple data centers without sharing raw data. However, current FL approaches rely on a server-based network topology, where central server failure disrupts training. Additionally, data heterogeneity across sites often slows convergence and reduces accuracy. To overcome these issues, we introduce a decentralized personalized federated learning collaborative aggregation network (pFedCAN). This framework has two core components: (1) separating local models into shared and personalized layers, and (2) forming a collaborative aggregation network via similarity detection in the shared layers. Specifically, each center trains its local model, then separates it into shared and personalized layers. The shared layer is exchanged with other centers, while the personalized layer remains local. Data centers analyze similarities in received shared layers to build a collaborative network, where shared layers from similar centers are aggregated to refine the model. This approach flexibly adapts to varying levels of data heterogeneity, enhancing model training efficiency. Validation on public datasets, ABIDE I and ADHD, shows that the proposed method outperforms current leading techniques. © 2024 IEEE.

关键词： Federated learning

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：