检索结果-内蒙古大学图书馆

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Lowy, Andrew Liu, Daogao Asi, Hilal Wisconsin Institute for Discovery University of Wisconsin-Madison United States Department of Computer Science University of Washington United States Apple Machine Learning Research United States

We study private stochastic convex optimization (SCO) under user-level differential privacy (DP) constraints. In this setting, there are n users (e.g., cell phones), each possessing m data items (e.g., text messages), and we need to protect the privacy of each user's entire collection of data items. Existing algorithms for user-level DP SCO are impractical in many large-scale machine learning scenarios because: (i) they make restrictive assumptions on the smoothness parameter of the loss function and require the number of users to grow polynomially with the dimension of the parameter space;or (ii) they are prohibitively slow, requiring at least (mn)3/2 gradient computations for smooth losses and (mn)3 computations for non-smooth losses. To address these limitations, we provide novel user-level DP algorithms with state-of-the-art excess risk and runtime guarantees, without stringent assumptions. First, we develop a linear-time algorithm with state-of-the-art excess risk (for a non-trivial linear-time algorithm) under a mild smoothness assumption. Our second algorithm applies to arbitrary smooth losses and achieves optimal excess risk in ≈ (mn)9/8 gradient computations. Third, for non-smooth loss functions, we obtain optimal excess risk in n11/8m5/4 gradient computations. Moreover, our algorithms do not require the number of users to grow polynomially with the dimension. © 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

An Optimal Real-Time Routing and Clustering Protocol for UAV Networks Based on Security 4

An Optimal Real-Time Routing and Clustering Protocol for UAV...

引用

4th International Conference on Sentiment Analysis and Deep learning, ICSADL 2025

作者： Sophia, S. Leelipushpam Paulraj, Getzi Jeba Yuvarajan, D. Karunya Institute of Technology and Sciences Division of Artificial Intelligence and Machine Learning Coimbatore India Karunya Institute of Technology and Sciences Division of Computer Science and Engineering Coimbatore India

ISBN: (纸本)9798331523923

Unmanned Aerial Vehicle (UAV) networks play an important role in different application areas such as military surveillance, emergency services, and infrastructure. However, these networks face significant challenges such as data leakage, unauthorized access, limited processing capacity, and energy constraints. Available techniques such as Elliptic Curve Cryptography (ECC), hash-based techniques, and dynamic clustering algorithms have been proposed to enhance security and operational performance;however, they often encounter challenges concerning scalability, adaptability to real-time changes within the network, and resource-related constraints. This research presents a streamlined authentication framework with the intent to facilitate secure and effective functionalities in turbulent and fluctuating settings. The proposed protocol uses Dijkstra's algorithm for optimal real-time routing, Adaptive Optimization Clustering (AOC) for dynamic resource distribution, and K-means clustering for noise classification. The proposed methodology achieves substantial improvements in power efficiency, latency, and Packet Delivery Rate (PDR) due to reduced computational overhead as well as risks such as data interception and replay attacks. Simulation results conclude that the system suggested here would be secure, scalable, and energy-efficient to be used for applications in demanding scenarios, such as mission-critical scenarios. © 2025 IEEE.

关键词： Emergency services

来源：评论

学校读者我要写书评

暂无评论

Predicting Soil Types and Crop Suitability through Comprehensive Soil Analysis

Predicting Soil Types and Crop Suitability through Comprehen...

引用

2024 International Conference on Communication, Computing and Internet of Things, IC3IoT 2024

作者： Sivakumar, S. Dharani, U. Hephzibah, Dineep Srm Institute Of Science And Technology Department Of Computer Science And Engineering Tiruchirapalli India Srm Institute Of Science And Technology Dept. Of Comp. Sci. And Engineering Specialization In Artificial Intelligence And Machine Learning Tiruchirapalli India

ISBN: (纸本)9798350352689

The analysis of soil nutrients and water levels holds utmost significance in the realms of agriculture. Determining the soil type and identifying the crop suitable for that specific soil are critical factors in optimizing agricultural yield. These considerations play a pivotal role in precision agriculture, ensuring that the chosen crops align with the soil's characteristics. This strategic alignment enhances the efficiency of cultivation practices, promotes sustainable agricultural methods, and ultimately contributes to increased crop yields. In this research paper, we systematically identify key soil properties, including NPK concentrations, temperature, humidity, and soil pH. The collected dataset, encompassing these crucial soil attributes, is utilized to train machine learning models. The primary goal is to forecast the soil type utilizing the gathered data, and subsequently, undertake further analyses, to ascertain the relevant crop types that are most suitable for the identified soil characteristics. This comprehensive approach involves the integration of data-driven insights and machine learning techniques to enhance our understanding of soil composition and facilitate informed decision-making in agriculture. © 2024 IEEE.

关键词： Soil testing

来源：评论

学校读者我要写书评

暂无评论

A recurrent neural network architecture for android mobile data analysis for detecting malware infected data

引用

Soft Computing 2024年第21期28卷 12917-12928页

作者： Murugan, Prabhu Manimaran, A. Sundar, Ramesh Dakshinamoorthy, Prabakar Rajaram, Gnanajeyaraman Garg, Shruti Department of ECE Saveetha School of Engineering Simats Chennai602105 India Department of Artificial Intelligence and Machine Learning Saveetha School of Engineering Simats Chennai602105 India Department of Netwoking and Communication School of Computing SRM Institute of Science and Technology SRM Nagar Kattankulathur India Department of Data Science and Business System School of Computing SRM Institute of Science and Technology SRM Nagar Kattankulathur India Department of Applied Machine Learning Saveetha School of Engineering Saveetha Institute of Medical and Technological Sciences Tamilnadu Chennai India Birla Institute of Technology Mesra Ranchi India

One of the latest modern communication devices is a mobile device seriously affected by multiple malware. Malware is a virus software installed automatically by hackers on various computing devices. Malware corrupts the system software, kills *.exe files, and tries to access user-sensitive data from the device. Around the world, 80% of people use smartphones, and 70% use Android phones. All Android phones are installed with anti-virus software but not used properly. Thus, it creates a pathway for malware to enter Android mobile devices and deploy vulnerabilities. Several earlier methods have focused on using the malware detection method by analyzing static data, which cannot identify the dynamic behavior of the malware. Thus, this paper implemented one of the deep learning algorithms, RNN, for identifying and detecting malware activities in Android mobile devices statically and dynamically. An IoT network architecture is proposed for experimenting with the RNN algorithms verifying the performance. The results of the RNN method show that it is better than other models in malware detection in mobile devices. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024.

关键词： Android malware

来源：评论

学校读者我要写书评

暂无评论

African Vultures Based Feature Selection with Multi-modal Deep learning for Automatic Seizure Prediction 2

African Vultures Based Feature Selection with Multi-modal De...

引用

2nd International Conference on Distributed Computing and Optimization Techniques, ICDCOT 2024

作者： Nallur, Mounika Sandhya, M. Khan, Zabiha Mohan, B.R. Nayana, C.P. Rajashekhar, S.A. Murray State University Computer Information Systems United States East West Institute of Technology Department of Computer Science and Engineering Bengaluru India Nitte Meenakshi Institute of Technology Department of Artificial Intelligence & Machine Learning Bengaluru India

ISBN: (纸本)9798350382952

The superiority of life for people with epilepsy can be greatly improved with the assistance of accurate seizure prediction and early warning. An automatic prediction model is required to procedure the EEG signals and account for the leads optimization problematic, as opposed to the majority of hand-designed prediction approaches. In this research, we put forth a fully automated model for seizure prediction using Channel and Spatial attention (CASA). The first step in the feature extraction practice is to pre-process the raw EEG signals. Large amounts of computation can be saved by adding more features to the system, but finding the right ones can be tricky. The African vulture optimization algorithm's (AVOA) strong capacity to break out of local optima is what makes this procedure possible. CASA saved the raw EEG data's temporal and geographical details. Automatic optimization of EEG full-lead data was completed with channel attention (CA), leading to an increase in the accuracy of predictions. The aforementioned adaptive learning of feature parameters was accomplished via spatial attention (SA). When all else fails, a fully associated layer is used to make the seizure forecast. The suggested algorithm is tested on the Freiburg EEG database, and the results reveal that the AVOA-based system performs admirably when it comes to predicting seizures. © 2024 IEEE.

关键词： Electroencephalography

来源：评论

学校读者我要写书评

暂无评论

Innovative Solutions for Cervical Cancer: Deep learning Models for Accurate Stage Prediction 2

Innovative Solutions for Cervical Cancer: Deep Learning Mode...

引用

2nd IEEE International Conference on Contemporary Computing and Communications, InC4 2024

作者： Yadav, Roshan Kumar Kumar, Divyam Kadam, Kalyani Devi, Seeta Gangarde, Rupali Department of Artificial Intelligence and Machine Learning Pune India Department of Obstetrics and Gynecological Nursing Pune India Department of Computer Science and Engineering Pune India

ISBN: (纸本)9798350383652

Cervical cancer, caused by the skin infection of the Human papillomavirus (HPV), ranked as the fourth most prevalent cancer globally, has profound and sometimes fatal implications for women. With approximately 500,000 new cases reported annually, the disease significantly impacts women's overall well-being. Ongoing efforts by researchers focus on developing a more cost-effective solution than regular HPV testing to manage and control this widespread health concern. Our research included an innovative framework for identifying and classifying stages of cervical cancer by processing the images of the tumours. To optimize the performance of our model, we developed a unique approach for selecting the best image data points by calculating a score, which is defined as the product of focus, sharpness, and resolution. Additionally, our study distinguishes itself by integrating three diverse feature extraction techniques-Local Binary Pattern (LBP), ResNet, and VGGNet-an inclusive strategy not commonly observed in similar projects. We even added deep convolutional architectures and machine learning techniques, including Random Forest, SVM, Decision Tree, XGBoost, a basic Neural Network Classifier, and a Multi-layered Perceptron (MLP). This hybrid approach allowed for an intrinsic comparison, identifying the most effective strategy among these. In contrast to current state-of-the-art methods, our approach yielded around 75% accuracy and 81% AUC score, which is greater than the existing studies on such dataset. © 2024 IEEE.

关键词： Contrastive learning

来源：评论

学校读者我要写书评

暂无评论

UNDERSTANDING TRANSFERABLE REPRESENTATION learning AND ZERO-SHOT TRANSFER IN CLIP 12

UNDERSTANDING TRANSFERABLE REPRESENTATION LEARNING AND ZERO-...

引用

12th International Conference on learning Representations, ICLR 2024

作者： Chen, Zixiang Deng, Yihe Li, Yuanzhi Gu, Quanquan Department of Computer Science University of California Los Angeles United States Machine Learning Department Carnegie Mellon University Pittsburgh United States

Multi-modal learning has become increasingly popular due to its ability to leverage information from different data sources (e.g., text and images) to improve the model performance. Recently, CLIP has emerged as an effective approach that employs vision-language contrastive pretraining to learn joint image and text representations and exhibits remarkable performance in zero-shot learning and text-guided natural image generation. Despite the huge practical success of CLIP, its theoretical understanding remains elusive. In this paper, we formally study transferrable representation learning underlying CLIP and demonstrate how features from different modalities get aligned. We also analyze its zero-shot transfer performance on the downstream tasks. Inspired by our analysis, we propose a new CLIP-type approach, which achieves better performance than CLIP and other state-of-the-art methods on benchmark datasets. © 2024 12th International Conference on learning Representations, ICLR 2024. All rights reserved.

关键词： Zero-shot learning

来源：评论

学校读者我要写书评

暂无评论

Apriori Analysis of Deep learning Models on a Multi-class Data Set 4th

Apriori Analysis of Deep Learning Models on a Multi-class Da...

引用

4th International Conference on Smart IoT Systems: Innovations and Computing, SSIC 2023

作者： Kosuru, Sai Sathwik Padmanabha Reddy, Y.C.A. Akanksha, Perala Venkata Chaitanya, Kotagiri Parsi, Snehaja Chunduri, Madhu Babu Department of Computer Science and Engineering B V Raju Institute of Technology Telangana Medak Narsapur502313 India Department of Artificial Intelligence & Machine Learning Chaitanya Bharathi Institute of Technology Ganidept Hyderabad500075 India

ISBN: (纸本)9789819751990

Deep learning techniques have large scopes in terms of applications and resource requirements, hence calling for a deep analysis of their behavior. Many studies focused on the application part of these models and there is a relatively lesser discussion about the unit-wise behavior. A multi-class dataset was taken and three different models were trained where their performances of training processes are described in terms of runtime characteristics like learning rate, times, etc. An apriori analysis was conducted where the runtime characteristics are investigated for their similarity with pre-defined knowledge of their architectures, as opposed to the posterior analysis which particularly focuses on observations. The work also serves as a framework for defining the logistics of the model training processes and fine-tunes the parameters devised in future for the maximum accuracy of the model. The results were further converted into graphical representations for enhanced delivery of insights. The study is also novel where, as opposed to previous works, in which performances are only tested, the model is evaluated in a design-specific manner dealing with only architectures and the parameters defining them. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Contrastive learning

来源：评论

学校读者我要写书评

暂无评论

Hyperbolic Contrastive learning for Document Representations - A Multi-View Approach with Paragraph-Level Similarities 27

Hyperbolic Contrastive Learning for Document Representations...

引用

27th European Conference on Artificial Intelligence, ECAI 2024

作者： Nam, Jaeeun Chalkidis, Ilias Rezaei, Mina Department of Statistics LMU Munich Munich Germany Department of Computer Science University of Copenhagen Denmark Munich Center for Machine Learning Munich Germany

ISBN: (纸本)9781643685489

Self-supervised learning (SSL) has gained prominence due to the increasing availability of unlabeled data and advances in computational efficiency, leading to revolutionized natural language processing with pre-trained language models like BERT and *** learning, a core concept in SSL, aims to reduce data dimensionality while preserving meaningful *** SSL methods typically embed data in Euclidean ***, recent research has revealed that alternative geometries can hold even richer representations, unlocking more meaningful insights from the *** by this, we propose two novel methods for integrating Hilbert geometry into self-supervised learning for efficient document ***, we present a method directly incorporating Hilbert geometry into the standard Euclidean contrastive learning ***, we propose a multi-view hyperbolic contrastive learning framework contrasting both documents and *** findings demonstrate that contrasting only paragraphs, rather than entire documents, can lead to superior efficiency and effectiveness. © 2024 The Authors.

关键词： Contrastive learning

来源：评论

学校读者我要写书评

暂无评论

IMDL-BenCo: A Comprehensive Benchmark and Codebase for Image Manipulation Detection & Localization 38

IMDL-BenCo: A Comprehensive Benchmark and Codebase for Image...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Ma, Xiaochen Zhu, Xuekang Su, Lei Du, Bo Jiang, Zhuohang Tong, Bingkui Lei, Zeyu Yang, Xinyu Pun, Chi-Man Lv, Jiancheng Zhou, Jizhe College of Computer Science Sichuan University China Department of Computer and Information Science University of Macao China Engineering Research Center of Machine Learning and Industry Intelligence MOE China

A comprehensive benchmark is yet to be established in the Image Manipulation Detection & Localization (IMDL) field. The absence of such a benchmark leads to insufficient and misleading model evaluations, severely undermining the development of this field. However, the scarcity of open-sourced baseline models and inconsistent training and evaluation protocols make conducting rigorous experiments and faithful comparisons among IMDL models challenging. To address these challenges, we introduce IMDL-BenCo, the first comprehensive IMDL benchmark and modular codebase. IMDL-BenCo: i) decomposes the IMDL framework into standardized, reusable components and revises the model construction pipeline, improving coding efficiency and customization flexibility;ii) fully implements or incorporates training code for state-of-the-art models to establish a comprehensive IMDL benchmark;and iii) conducts deep analysis based on the established benchmark and codebase, offering new insights into IMDL model architecture, dataset characteristics, and evaluation standards. Specifically, IMDL-BenCo includes common processing algorithms, 8 state-of-the-art IMDL models (1 of which are reproduced from scratch), 2 sets of standard training and evaluation protocols, 15 GPU-accelerated evaluation metrics, and 3 kinds of robustness evaluation. This benchmark and codebase represent a significant leap forward in calibrating the current progress in the IMDL field and inspiring future breakthroughs. Code is available at: https://***/scu-zjz/IMDLBenCo. © 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

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：