检索结果-内蒙古大学图书馆

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Chen, Chaoqi Tang, Luyao Huang, Hui College of Computer Science and Software Engineering Shenzhen University China School of Informatics Xiamen University China

Since deep learning models are usually deployed in non-stationary environments, it is imperative to improve their robustness to out-of-distribution (OOD) data. A common approach to mitigate distribution shift is to regularize internal representations or predictors learned from in-distribution (ID) data to be domain invariant. Past studies have primarily learned pairwise invariances, ignoring the intrinsic structure and high-order dependencies of the data. Unlike machines, humans recognize objects by first dividing them into major components and then identifying the topological relation of these components. Motivated by this, we propose Reconstruct and Match (REMA), a general learning framework for object recognition tasks to endow deep models with the capability of capturing the topological homogeneity of objects without human prior knowledge or fine-grained annotations. To identify major components from objects, REMA introduces a selective slot-based reconstruction module to dynamically map dense pixels into a sparse and discrete set of slot vectors in an unsupervised manner. Then, to model high-order dependencies among these components, we propose a hypergraph-based relational reasoning module that models the intricate relations of nodes (slots) with structural constraints. Experiments on standard benchmarks show that REMA outperforms state-of-the-art methods in OOD generalization and test-time adaptation settings. © 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

Contextual Text Mining Framework for Unstructured Textual Judicial Corpora through Ontologies

引用

computer Systems Science & engineering 2022年第12期43卷 1357-1374页

作者： Zubair Nabi Ramzan Talib Muhammad Kashif Hanif Muhammad Awais Department of Computer Science Government College UniversityFaisalabad38000Pakistan Department of Software Engineering Government College UniversityFaisalabad38000Pakistan

Digitalization has changed the way of information processing, and newtechniques of legal data processing are evolving. Text mining helps to analyze andsearch different court cases available in the form of digital text documents toextract case reasoning and related data. This sort of case processing helps professionals and researchers to refer the previous case with more accuracy in reducedtime. The rapid development of judicial ontologies seems to deliver interestingproblem solving to legal knowledge formalization. Mining context informationthrough ontologies from corpora is a challenging and interesting field. Thisresearch paper presents a three tier contextual text mining framework throughontologies for judicial corpora. This framework comprises on the judicial corpus,text mining processing resources and ontologies for mining contextual text fromcorpora to make text and data mining more reliable and fast. A top-down ontologyconstruction approach has been adopted in this paper. The judicial corpus hasbeen selected with a sufficient dataset to process and evaluate the *** experimental results and evaluations show significant improvements incomparison with the available techniques.

关键词： Natural language processing judicial corpora contextual text mining ontologies information extraction information retrieval

来源：评论

学校读者我要写书评

暂无评论

MonkeyGPT: Generative AI in Network Anomaly Detection of Video Conference Applications 22

MonkeyGPT: Generative AI in Network Anomaly Detection of Vid...

引用

22nd IEEE International Symposium on Parallel and Distributed Processing with Applications, ISPA 2024

作者： Xu, Junhao He, Dongbiao Zhang, Chen Zhou, Xu Ming, Zhongxing Cui, Laizhong Sangfor Technologies Inc China Shenzhen University College of Computer Science and Software Engineering China China

ISBN: (纸本)9798331509712

The rapid advancement of generative artificial intelligence (GAI) has led to the creation of transformative applications such as ChatGPT, which significantly boosts text processing efficiency and diversifies audio, image, and video content. Beyond digital content creation, GAI's capability to analyze complex data distributions holds immense potential for next-generation networks and communications, especially given the swift rise of video conferencing applications (VCAs). This paper presents a dynamic, real-time method for detecting anomalous network links in video conferencing applications. The proposed tool, MonkeyGPT, generates tracing representations of network activity and trains a large language model from scratch to serve as a detection system based on network traffic data. Unlike traditional methods, MonkeyGPT provides an unrestricted search space and does not rely on predefined rules or patterns, enabling it to detect a wider range of anomalies. We demonstrate the effectiveness of MonkeyGPT as an anomaly detection tool in real-world VCAs. The results indicate that the model possesses strong detection capabilities, achieving an accuracy rate of over 97%. It is applicable to various platforms, including Zoom, Microsoft Teams, Tencent Meeting, and Feishu, showcasing its robust adaptability. © 2024 IEEE.

关键词： Video conferencing

来源：评论

学校读者我要写书评

暂无评论

UeFormer: A Unifying Encoding Transformer Neural Operator for meteorological downscaling 5

UeFormer: A Unifying Encoding Transformer Neural Operator fo...

引用

5th International Conference on Big Data and Artificial Intelligence and software engineering, ICBASE 2024

作者： Wang, Yichi Jian, Yifan Chen, Xinrui Zhang, Haodi Shenzhen University College of Computer and Software Engineering Shenzhen China Hongling Middle School Shenzhen China

ISBN: (纸本)9798331506612

Meteorological change presents more and more significant challenges to accurate meteorological forecasting. Traditional meteorological forecasting models, constrained by their fixed-resolution capabilities, often fail to provide precise information for specific coordinates meteorological state. To overcome this limitation, downscaling tasks have gained prominence in meteorology. Operator learning is particularly effective for downscaling because it can capture the partial differential equation (PDE) characteristics of meteorological processes, enabling predictions of meteorological states at arbitrary coordinates. However, conventional operator learning methods often neglect the integration of multimodal meteorological data and do not effectively encode the relationship between query coordinates and the current meteorological state, leading to suboptimal predictions. To address these issues, we propose a novel operator learning method, the Unifying encoding Transformer Neural Operator (Ueformer). This method harnesses the powerful multimodal fusion capabilities of transformers and jointly encodes both query coordinates and meteorological states. Our approach demonstrates superior performance across various meteorological variables and metrics, outperforming baseline methods. Furthermore, we provide theoretical proofs of the method's approximation capabilities, contributing significantly to the field of meteorology. © 2024 IEEE.

关键词： Weather forecasting

来源：评论

学校读者我要写书评

暂无评论

A Model Driven Framework Enabling the Development of Standardized Text-Free User Interface for Smartphones 21

A Model Driven Framework Enabling the Development of Standar...

引用

21st International Bhurban Conference on Applied Sciences and Technology, IBCAST 2024

作者： Rasheed, Uzair Khalid, Sadia Mehmood, Rabia Ahmad, Muhammad Muneer, Mishal College of Electrical and Mechanical Engineering NUST Department of Computer and Software Engineering Islamabad Pakistan Lahore College for Women University Department of Computer Science Lahore Pakistan International Univesity of Rostock Department of Computer science Germany

ISBN: (纸本)9798331516680

Smartphones offer a wide range of applications globally, and usability is a key factor in their quality. Human-computer Interaction (HCI) continuously evolves to enhance usability by improving user interface (UI) components. However, most UI trends are designed for literate users, neglecting a significant portion of the population who cannot read or write. These users often rely on others for assistance with smartphone applications, and text-based UIs hinder their learning and usage, reducing overall usability. This research addresses the limitations of text-based standards by introducing a standardized framework for developing text-free user interfaces for smartphones. The framework utilizes static and dynamic images instead of text to convey widget or button functions, aiming to enhance usability for illiterate users, including preschoolers. Developed with Obeo Designer and Sirius, the framework's Platform Independent Metamodel (PIM) was validated through case studies representing the target user groups. Future work involves Model to Text (M2T) transformation, and results show the framework's potential for automating applications with text-free UIs. © 2024 IEEE.

关键词： Smartphones

来源：评论

学校读者我要写书评

暂无评论

Radar style transfer for metric robot localisation on lidar maps

引用

CAAI Transactions on Intelligence Technology 2023年第1期8卷 139-148页

作者： Huan Yin Yue Wang Jun Wu Rong Xiong College of Control Science and Engineering Zhejiang UniversityHangzhouChina Department of Electronic and Computer Engineering Hong Kong University of Science and TechnologyKowloonHong KongChina

Lidar and visual data are affected heavily in adverse weather conditions due to sensing mechanisms,which bring potential safety hazards for vehicle *** sensing is desirable to build a more robust navigation *** this paper,a cross-modality radar localisation on prior lidar maps is ***,the proposed workflow consists of two parts:first,bird's-eye-view radar images are transferred to fake lidar images by training a generative adversarial network *** with online radar scans,a Monte Carlo localisation framework is built to track the robot pose on lidar *** whole online localisation system only needs a rotating radar sensor and a pre-built global lidar *** the experimental section,the authors conduct an ablation study on image settings and test the proposed system on Oxford Radar Robot Car *** promising results show that the proposed localisation system could track the robot pose successfully,thus demonstrating the feasibility of radar style transfer for metric robot localisation on lidar maps.

关键词： lidar local maps

来源：评论

学校读者我要写书评

暂无评论

Counteracting New Attacks in CPS: A Few-shot Class-incremental Adaptation Strategy for Intrusion Detection System

引用

IEEE Transactions on Network and Service Management 2025年第3期22卷 2473-2488页

作者： Dong, Xinrui Lai, Yingxu Zhang, Xiao Xu, Xinyu Beijing University of Technology College of Computer Science Beijing100124 China Ministry of Education Engineering Research Center of Intelligent Perception and Autonomous Control Beijing100124 China

The deep integration of physical devices and communication networks has increased the security risks of cyber-physical systems (CPSs) compared to traditional control systems. Deep learning-based intrusion detection systems (IDSs) play a crucial role in ensuring CPSs security. However, the existing IDSs often rely on known attack features, rendering them unable to withstand emerging new attacks arising from the dynamic evolution of intrusion behaviors. This paper aims to develop an IDSs with high adaptability and strong generalization capabilities, which is capable of rapidly adapting to new attack classes with only a few new samples. To achieve this objective, we propose CAT-IDS, a few-shot class-incremental adaptation strategy for an IDS to counteract new attacks on CPSs. We design a highly symmetric classifier structure for CAT-IDS that can flexibly adjust the classification space to adapt to new attacks. Furthermore, we calibrate the biased distribution formed by a few training samples through statistical feature transfer. In order to prevent the model from forgetting old attack information during the adaptation process, we devise hybrid features for attack detection. These features contain essential information for both old and new class classifications. We demonstrate the effectiveness of CAT-IDS through multiple experiments on three CPSs datasets. The results show that CAT-IDS achieves an average accuracy improvement of approximately 4. 5% compared to the state-of-the-art methods, demonstrating its superior ability to adapt to new attacks while maintaining high performance in classifying existing attacks. © 2004-2012 IEEE.

关键词： Network intrusion

来源：评论

学校读者我要写书评

暂无评论

Named Entity Recognition in Electronic Medical Records Incorporating Pre-trained and Multi-Head Attention

IAENG International Journal of Computer Science

引用

IAENG International Journal of computer Science 2024年第4期51卷 401-408页

作者： Yang, Haotian Wang, Li Yang, Yanpeng School of Computer Science and Software Engineering University of Science and Technology Liaoning CO Anshan114051 China College of Computer Science and Software Engineering University of Science and Technology Liaoning CO Anshan114051 China Network Information Centre University of Science and Technology Liaoning CO Anshan114051 China

Chinese Named Entity Recognition (NER) for Electronic Medical Records (EMR) is a fundamental task in building a digital hospital and is widely considered to be a sequence annotation problem in the Natural Language Processing domain. (NLP). However, existing deep learning sequence annotation models cannot fully use the large amount of unannotated data for Chinese EMRs that contain a vast number of professional unregistered words, named entities, and inter-of-entity relationships carrying rich professional knowledge. Moreover, the syntactic structure of EMR sentences is complex, and the text is long;the features of the EMR documents often cannot be captured deeply. Aiming at these two problems, this paper proposes a deep learning method that combines Multi-Head Attention with a pre-trained language model. (BERT-BiGRU-Att-CRF). The method uses the BERT pre-trained model to obtain dynamic word vectors combined with contextual information, extracts global semantic features through a Bidirectional Gated Recurrent Unit (BiGRU), obtains augmented semantic features by using the Multi-Head Attention. Finally, using Conditional Random Field (CRF) decoding, outputs the globally optimal label sequence with greatest probability. The model is trained using the CCKS2019 Chinese EMR dataset containing six types of entities: anatomical sites, surgeries, diseases and diagnoses, medicines, laboratory tests, and imaging tests, and good results are achieved with an F1 score of 86.97%. © (2024) International Association of Engineers.

关键词： Random processes

来源：评论

学校读者我要写书评

暂无评论

Reliable Task Offloading for 6G-Based IoT Applications

引用

computers, Materials & Continua 2025年第2期82卷 2255-2274页

作者： Usman Mahmood Malik Muhammad Awais Javed Ahmad Naseem Alvi Mohammed Alkhathami Department of Computer Software Engineering National University of Science and Technology(NUST)Islamabad44000Pakistan Department of Electrical and Computer Engineering COMSATS University IslamabadIslamabad45550Pakistan Information Systems Department College of Computer and Information SciencesImam Mohammad Ibn Saud Islamic University(IMSIU)Riyadh11432Saudi Arabia

Fog computing is a key enabling technology of 6G systems as it provides quick and reliable computing,and data storage services which are required for several 6G *** Intelligence(AI)algorithms will be an integral part of 6G systems and efficient task offloading techniques using fog computing will improve their performance and *** this paper,the focus is on the scenario of Partial Offloading of a Task to Multiple Helpers(POMH)in which larger tasks are divided into smaller subtasks and processed in parallel,hence expediting task ***,using POMH presents challenges such as breaking tasks into subtasks and scaling these subtasks based on many interdependent factors to ensure that all subtasks of a task finish simultaneously,preventing resource ***,applying matching theory to POMH scenarios results in dynamic preference profiles of helping devices due to changing subtask sizes,resulting in a difficult-to-solve,externalities *** paper introduces a novel many-to-one matching-based algorithm,designed to address the externalities problem and optimize resource allocation within POMH ***,we propose a new time-efficient preference profiling technique that further enhances time optimization in POMH *** performance of the proposed technique is thoroughly evaluated in comparison to alternate baseline schemes,revealing many advantages of the proposed *** simulation findings indisputably show that the proposed matching-based offloading technique outperforms existing methodologies in the literature,yielding a remarkable 52 reduction in task latency,particularly under high workloads.

关键词： 6G IoT task offloading fog computing

来源：评论

学校读者我要写书评

暂无评论

UAV-Assisted Maritime Data Collection via Optical Communications using Deep Reinforcement Learning 9

UAV-Assisted Maritime Data Collection via Optical Communicat...

引用

9th International Conference on Big Data Computing and Communications, BigCom 2023

作者： Ma, Saisai Li, Mengzhen Deng, Tinglong Wang, Jinglong Ruby, Rukhsana Luo, Hanjiang Shandong University of Science and Technology College of Computer Science and Engineering Qingdao China Shenzhen University College of Computer Science and Software Engineering Shenzhen China

ISBN: (纸本)9798350331240

Ocean data collection via unmanned aerial vehicles (UAVs) has received widespread attention due to its flexibility and low cost. To further improve the efficiency of large maritime data collection between the UAVs and the ocean surface buoys, optical wireless transmission is considered as a promising technique because of its low latency and high bandwidth. However, the waves and other disturbances in the complex oceanic environment result in drift and instability of the buoy, which deteriorate and even interrupt the line-of-sight (LOS) optical transmission. To tackle these challenges, in this paper, we propose a reliable data collection scheme based on a deep reinforcement learning (DRL) approach. We first model the optical channel and calculate the maximum transmission range that satisfies a pre-defined bit error rate (BER) to ensure the quality of service (QoS). Then, we formulate the data collection procedure as a Markov decision process (MDP) aiming at maximizing the received signal intensity and minimizing the energy consumption, in which the system uncertainty caused by the oceanic environmental disturbances is taken into account. Finally, we propose a beam pointing adjustment algorithm based on deep deterministic policy gradient (DDPG) approach to alleviate the performance deterioration and maintain a stable LOS communication. Through extensive simulations, the results demonstrate that the proposed scheme is effective and achieves reliable data collection via the optical links. © 2023 IEEE.

关键词： Reinforcement 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：