检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Yang, Minmin Ren, Huantao Velipasalar, Senem Electrical Engineering and Computer Science Dept. Syracuse University SyracuseNY United States

Existing zero-shot 3D point cloud segmentation methods often struggle with limited transferability from seen classes to unseen classes and from semantic to visual space. To alleviate this, we introduce 3D-PointZshotS, a geometry-aware zero-shot segmentation framework that enhances both feature generation and alignment using latent geometric prototypes (LGPs). Specifically, we integrate LGPs into a generator via a cross-attention mechanism, enriching semantic features with fine-grained geometric details. To further enhance stability and generalization, we introduce a self-consistency loss, which enforces feature robustness against point-wise perturbations. Additionally, we re-represent visual and semantic features in a shared space, bridging the semantic-visual gap and facilitating knowledge transfer to unseen classes. Experiments on three real-world datasets, namely ScanNet, SemanticKITTI, and S3DIS, demonstrate that our method achieves superior performance over four baselines in terms of harmonic mIoU. The code is available at Github. © 2025, CC BY.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Reconsider Time Series Analysis for Insider Threat Detection

Reconsider Time Series Analysis for Insider Threat Detection

引用

IEEE International Conference on Big Data

作者： Chia-Cheng Chen Hsing-Kuo Pao Dept. of Computer Science and Information Engineering National Taiwan University of Science and Technology Taipei Taiwan

ISBN: (数字)9798350362480

ISBN: (纸本)9798350362497

Insider threat detection (ITD) presents a significant challenge in cybersecurity, particularly within large and complex organizations. Traditionally, ITD has been overshadowed by the focus of external threats, resulting in less attention and development in this critical area. Conventional ITD approaches often rely heavily on event-driven approaches. On top of that, researchers developed various rule-based methods to conquer the tasks. Based on that, we often ignore the intrinsic temporal relationships that are naturally built in between events that occur in different moments. For instance, we may easily understand events with causality such as one anomalous event followed by another specific event to complete a malicious action; however, may not be aware of events that occur around 9 am every morning during working hours. In our opinion, we attempt to re-consider the temporal behavior to extract the information hidden in cyberspace activities. Specifically, some effective sentence embeddings can assist us in providing informative internal representations to summarize temporal behaviors in the temporal activity sequences to make the right judgment on insider threat detection. In this paper, we propose a novel methodology for insider threat detection that emphasizes temporal relationship modeling on top of already-matured event sequence analysis to effectively catch insider threats. The proposed approach leverages contrastive sentence embeddings to learn users’ intentions in sequences, followed by the deployment of a user-level and event-level Contrastive Learning (euCL) model to incorporate temporal behaviors with user behavior embeddings. To validate the proposed methodology, we conduct extensive analyses and experiments using the publicly available CERT dataset. The results demonstrate the effectiveness and robustness of the proposed method in detecting insider threats and identifying malicious scenarios, highlighting its potential for enhancing cybersecurity measu

关键词： Analytical models Sequences Time series analysis Semantics Organizations Threat assessment Robustness Data models computer security Standards

来源：评论

学校读者我要写书评

暂无评论

Exploring Voice-Activated AI's Impact on Workplace Productivity with Virtual Assistants

Exploring Voice-Activated AI's Impact on Workplace Productiv...

引用

Integration of Emerging Technologies for the Digital World (ICIETDW), International Conference on

作者： Veeramalai Sankaradass S. Sudhakar Dept. of Computer Science and Engineering Chennai Institute of Technology Chennai Tamil Nadu India Dept. of Computer Science and Engineering Chennai Institute of Technology Anna University Chennai Tamil Nadu India

ISBN: (数字)9798331541170

ISBN: (纸本)9798331541187

This paper presents a revolutionary computer desktop computer aide encouraged by GPT-3, a cutting-edge language version along with support discovering strategies. The chatbot functioning as a detailed device for computer individuals, effortlessly incorporates AI to give user-friendly support in jobs varying from standard procedures to progressed data administration and also setups changes. By using the capacities of GPT-3 the chatbot involves individuals in all-natural language discussions, supplying thorough reactions and also carrying out jobs with very little individual input. In addition, an assistance discovering based suggestion system adjusts to individual actions, continually refining ideas to maximize individual experience. This paper likewise represents a substantial innovation in individual computer showcasing the transformative possibility of AI in boosting performance coupled with individualizing the computer experience. With this cutting-edge technique we picture a future where AI-driven individual aides redefine the requirements of customer communication and also effectiveness in computer procedures.

关键词： Training Productivity Technological innovation Refining Employment Personal voice assistants Reinforcement learning Chatbots Reliability Standards

来源：评论

学校读者我要写书评

暂无评论

Facial Forgery-Based Deepfake Detection Using Fine-Grained Features 22

Facial Forgery-Based Deepfake Detection Using Fine-Grained F...

引用

22nd IEEE International Conference on Machine Learning and Applications, ICMLA 2023

作者： Nadimpalli, Aakash Varma Rattani, Ajita School of Computing Wichita State University Kansas United States Uni. of North Texas at Denton Dept. of Computer Science and Engineering TX United States

ISBN: (纸本)9798350345346

Facial forgery by deepfakes has caused major se-curity risks and raised severe societal concerns. As a counter-measure, a number of deepfake detection methods have been proposed. Most of them model deepfake detection as a binary classification problem using a backbone convolutional neural network (CNN) architecture pretrained for the task. These CNN-based methods have demonstrated very high efficacy in deepfake detection with the Area under the Curve (AUC) as high as 0.99. However, the performance of these methods degrades signifi-cantly when evaluated across datasets and deepfake manipulation techniques. This draws our attention towards learning more subtle, local, and discriminative features for deepfake detection. In this paper, we formulate deepfake detection as a fine-grained classification problem and propose a new fine-grained solution to it. Specifically, our method is based on learning subtle and generalizable features by effectively suppressing background noise and learning discriminative features at various scales for deepfake detection. Through extensive experimental validation, we demonstrate the superiority of our method over the published research in cross-dataset and cross-manipulation generalization of deepfake detectors for the majority of the experimental scenarios. © 2023 IEEE.

关键词： Cross-Manipulation Generalization Deepfakes Facial Manipulations Fine-Grained Classification

来源：评论

学校读者我要写书评

暂无评论

EEG-Based Alzheimer’s Disease Detection Using Deep Learning Techniques

EEG-Based Alzheimer’s Disease Detection Using Deep Learning...

引用

IEEE Region 10 Humanitarian Technology Conference (R10-HTC)

作者： Ancy S.A. Jerrin Thomas Panachakel Dept. of Computer Science and Engineering College of Engineering Trivandrum India Dept. of Electronics and Comm. Engg. College of Engineering Trivandrum India

ISBN: (数字)9798350340204

ISBN: (纸本)9798350340211

Alzheimer’s disease (AD) is a neurological condition characterized by the degeneration of neuron cells, leading to cognitive impairment. This project investigates the effectiveness of EEG-based features for detecting Alzheimer’s Disease using deep learning techniques. The dataset comprises EEG recordings from healthy controls, Frontotemporal Dementia (FTD) patients, and patients diagnosed with AD. Data preprocessing involved segmenting the EEG signals into fixed-length epochs and filtering the baseline EEG into alpha, beta, and theta frequency bands. Features such as Mean Phase Coherence (MPC), Differential Entropy (DE), and Mean Squared Coherence (MSC) were computed pairwise for each epoch and frequency band. A feature matrix was constructed for each subject by stacking these features, which were then converted into images and used as input data for a ResNet50-based pre-trained model. The model was tested in leave-one-out cross-validation setting and achieved results comparable to the state-of-the-art in distinguishing between EEG recordings from healthy and AD-affected subjects.

关键词： Deep learning Computational modeling Coherence Brain modeling Feature extraction Electroencephalography Data models Recording Reliability Alzheimer's disease

来源：评论

学校读者我要写书评

暂无评论

Comparative Advances in Financial Sentiment Analysis:A Review of BERT,FinBert, and Large Language Models

Comparative Advances in Financial Sentiment Analysis:A Revie...

引用

Intelligent Data Communication Technologies and Internet of Things (IDCIoT), International Conference on

作者： Manish Barath Mahendran Aswin Kumar Gokul Poornima Lakshmi S. Pavithra Dept. of Computer Science and Engineering Chennai Institute of Technology Chennai Tamil Nadu India

ISBN: (数字)9798331527549

ISBN: (纸本)9798331527556

Capturing market sentiments and supporting well-informed financial decision-making depend on the developing field of Financial Sentiment Analysis (FSA). Natural language processing (NLP) has made significant strides in comprehending and categorizing sentiment in intricate financial texts, especially with the use of Large Language Models (LLMs). The application of various LLMs, such as Bidirectional Encoder Representations form Transformers(BERT) and Financial BERT (FinBERT), as well as distilled models, such as DistilBERT and DistilRoBERTa, on a variety of financial datasets, including Financial Phrase Bank and LexisNexis news articles, was the main focus of this review article. highlighting different approaches such as model fine-tuning, zero-shot and few-shot learning, and prompt engineering. The study focuses on practical model predictions through a case study of sentimental analysis of cryptocurrencies. While FinBERT, a financial variant of BERT, exhibits high accuracy and robustness, other LLMs exhibit varying degrees of success based on the dataset and domain requirements. The analysis concentrates on the difficulties, compromises, and potential paths for improving LLMs for financial sentiment analysis.

关键词： Sentiment analysis Technological innovation Reviews Large language models Decision making Bidirectional control Market research Encoding Risk management Indexing

来源：评论

学校读者我要写书评

暂无评论

Child and Women Safety Measures using iBeacon Technology

Child and Women Safety Measures using iBeacon Technology

引用

Trends in Material science and Inventive Materials (ICTMIM), International Conference on

作者： Shanker Shalini Keerthana E Nishanthi P Dept. of Computer Science and Engineering St. Joseph’s Institute of Technology Chennai

ISBN: (数字)9798331501488

ISBN: (纸本)9798331501495

A smart real-time tracking system featuring a compact wearable device is designed to enhance the safety of children and women by utilizing IoT-driven technology and a secure database for efficient movement monitoring. This system operates on four key pillars: outdoor localization, cross-platform communication, server integration, and user engagement. The tracking mechanism comprises a wearable device embedded with an iBeacon and a built-in MEMS accelerometer, seamlessly linked to a smartphone application. The system begins with sensor calibration to establish reference values, followed by continuous tracking of sensor variations, particularly accelerometer data, to detect motion patterns with precision. Once paired with the designated Android application, the wearable device analyze movement data and transmits GPS coordinates to a centralized server, ensuring real-time location tracking of both the device and the associated iBeacon. Planned future improvements include the integration of Google Maps for enhanced real-time visualization, interactive monitoring, geofencing alerts, and emergency response notifications, making the system more robust and responsive to safety concerns.

关键词： Accelerometers Tracking Web services Real-time systems Threat assessment software Safety Servers Wearable devices Monitoring

来源：评论

学校读者我要写书评

暂无评论

OrcoDCS: An IoT-Edge Orchestrated Online Deep Compressed Sensing Framework

OrcoDCS: An IoT-Edge Orchestrated Online Deep Compressed Sen...

引用

International Conference on Distributed Computing Systems Workshop

作者： Cheng-Wei Ching Chirag Gupta Zi Huang Liting Hu Dept. of Computer Science and Engineering University of California Santa Cruz USA Dept. of Computer Science Virginia Tech USA

Compressed data aggregation (CDA) over wireless sensor networks (WSNs) is task-specific and subject to environmental changes. However, the existing compressed data aggregation (CDA) frameworks (e.g., compressed sensing-based data aggregation, deep learning(DL)-based data aggregation) do not possess the flexibility and adaptivity required to handle distinct sensing tasks and environmental changes. Additionally, they do not consider the performance of follow-up IoT data-driven deep learning (DL)-based applications. To address these shortcomings, we propose OrcoDCS, an IoT-Edge orchestrated online deep compressed sensing framework that offers high flexibility and adaptability to distinct IoT device groups and their sensing tasks, as well as high performance for follow-up applications. The novelty of our work is the design and deployment of IoT-Edge orchestrated online training framework over WSNs by leveraging an specially-designed asymmetric autoencoder, which can largely reduce the encoding overhead and improve the reconstruction performance and robustness. We show analytically and empirically that OrcoDCS outperforms the state-of-the-art DCDA on training time, significantly improves flexibility and adaptability when distinct reconstruction tasks are given, and achieves higher performance for follow-up applications.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Case Study-Based Approach of Quantum Machine Learning in Cybersecurity: Quantum Support Vector Machine for Malware Classification and Protection

Case Study-Based Approach of Quantum Machine Learning in Cyb...

引用

IEEE Annual International computer software and Applications Conference (COMPSAC)

作者： Mst Shapna Akter Hossain Shahriar Sheikh Iqbal Ahamed Kishor Datta Gupta Muhammad Rahman Atef Mohamed Mohammad Rahman Akond Rahman Fan Wu Dept. of Computer Science Kennesaw State University Kennesaw USA Dept. of Information Technology Kennesaw State University Kennesaw USA Dept. of Computer Science Marquette University Wisconsin USA Dept. of Computer Science Clark Atlanta University Georgia USA Dept. of Information Technology Clayton State University Morrow USA Dept. of Information Technology Georgia Southern University Georgia Dept. of Electrical and Computer Engineering Florida International University Florida USA Dept. of Computer Science and Software Engineering Auburn University Auburn USA Dept of Computer Science Tuskegee University Tuskegee USA

Quantum machine learning (QML) is an emerging field of research that leverages quantum computing to improve the classical machine learning approach to solve complex real-world problems. QML has the potential to address cybersecurity-related challenges. Considering the novelty and complex architecture of QML, resources are not yet explicitly available that can pave cybersecurity learners to instill efficient knowledge of this emerging technology. In this research, we design and develop QML-based ten learning modules covering various cybersecurity topics by adopting student centering case-study based learning approach. We apply one subtopic of QML on a cybersecurity topic comprised of pre-lab, lab, and post-lab activities towards providing learners with hands-on QML experiences in solving real-world security problems. In order to engage and motivate students in a learning environment that encourages all students to learn, pre-lab offers a brief introduction to both the QML subtopic and cybersecurity problem. In this paper, we utilize quantum support vector machine (QSVM) for malware classification and protection where we use open source Pennylane QML framework on the drebin215 dataset. We demonstrate our QSVM model and achieve an accuracy of 95% in malware classification and protection. We will develop all the modules and introduce them to the cybersecurity community in the coming days.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An Artificial Intelligence based Machine Learning Approach for Automatic Blood Glucose Level Identification of Diabetes Patients

An Artificial Intelligence based Machine Learning Approach f...

引用

2023 International Conference on Disruptive Technologies, ICDT 2023

作者： Maguluri, Lakshmana Phaneendra Santhosh, K. Meenakshi, K. Farooqui, Muhammad Ahtesham Parveen Sultana, H. Saravanan, V. Department of Computer Science and Engineering Andhra Pradesh Guntur India Kombolcha Institute of Technology Wollo University Department of Mechatronics Ethiopia Cmr Institute of Technology Department of Mathematics Bengaluru Karnataka India Sam Global University Dept. of Computer Science & Engineering Madhya Pradesh Raisen India Vellore Institute of Technology School of Computer Science and Engineering Tamil Nadu Vellore India Dambi Dollo University Department of Computer Science Dambi Dollo Ethiopia

ISBN: (纸本)9798350323887

In general, treatments and medications for diabetics are usually prescribed based on the level of glucose-level in the human-blood. Blood glucose-level was calculated by performing two separate tests: before meals and after meals. The practical functions present in this test enable physicians to carry out appropriate treatment modalities. This paper introduces an improved method of running a machine learning system. Its main task is to accurately analyze the given input data and calculate the correct point of blood-glucose in the blood of diabetics. It was designed to do so and then list the appropriate control methods and drugs and share that data with the user in a paperless digital manner. Thus it is enough for patients to go to laboratories and do tests. Only their results will be calculated and sent to them for treatment. © 2023 IEEE.

关键词： Glucose

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：