检索结果-内蒙古大学图书馆

Enhancing Wildlife Monitoring Through Infrared Imaging: A Deep Learning Approach for Improved Animal Region Segmentation and Species Identification

引用

Journal of Harbin institute of Technology(New series) 2024年

作者： V Janani C Shanthi Department of Computer Science， School of Computing Sciences， Vels Institute of Science Technology and Advanced Studies

Infrared imaging has proven to be a powerful tool in various fields， with particular significance in wildlife monitoring. In this paper， we present an extension of our previous research， focusing on advancing the segmentation of animal regions in enhanced infrared images and expanding the scope to include species identification. Our proposed methodology builds upon the success of the R-CNN (Region-based Convolutional Neural Network) object detection to improve the accuracy and robustness of animal region segmentation， while simultaneously extending our model’s capabilities to identify and classify the species within those regions. By fine-tuning the R-CNN model on a larger dataset that includes annotated infrared images and species labels， we enhance its capacity to not only accurately segment animal regions but also classify them into specific species categories. To assess the performance of our extended model， we employ a comprehensive set of evaluation metrics， including pixel-based metrics like Intersection over Union (IoU)， as well as species classification accuracy. Our results demonstrate significant improvements in both region segmentation accuracy and species identification compared to our previous work and existing methods. This research showcases the potential of deep learning techniques， combined with transfer learning， to advance wildlife monitoring applications using infrared imaging.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Bayesian Deep Learning for Dynamic Power System State Prediction Considering Renewable Energy Uncertainty

引用

Journal of Modern Power Systems and Clean Energy 2022年第4期10卷 913-922页

作者： Shiyao Zhang James J.Q.Yu Academy for Advanced Interdisciplinary Studies Southern University of Science and TechnologyShenzhen 518055China Department of Computer Science and Engineering Southern University of Science and TechnologyShenzhen 518055China

Modern power systems are incorporated with distributed energy sources to be environmental-friendly and ***,due to the uncertainties of the system integrated with renewable energy sources,effective strategies need to be adopted to stabilize the entire power ***,the system operators need accurate prediction tools to forecast the dynamic system states *** this paper,we propose a Bayesian deep learning approach to predict the dynamic system state in a general power ***,the input system dataset with multiple system features requires the data pre-processing ***,we obtain the dynamic state matrix of a general power system through the Newton-Raphson power flow ***,by incorporating the state matrix with the system features,we propose a Bayesian long short-term memory(BLSTM)network to predict the dynamic system state variables *** results show that the accurate prediction can be achieved at different scales of power systems through the proposed Bayesian deep learning approach.

关键词： Bayesian deep learning data analytics NewtonRaphson power flow renewable energy source system state

来源：评论

学校读者我要写书评

暂无评论

Optimizing Data-Driven Federated Learning in UAV Networks 30

Optimizing Data-Driven Federated Learning in UAV Networks

引用

30th IEEE International Conference on Parallel and Distributed Systems, ICPADS 2024

作者： Li, Datian Xiao, Mingjun Xu, Yin Wu, Jie University of Science and Technology of China School of Computer Science and Technology & Suzhou Institute for Advanced Research China Temple University Department of Computer and Information Sciences United States

ISBN: (纸本)9798331515966

Federated Learning (FL) is an emerging privacy-preserving distributed machine learning paradigm that enables numerous clients to collaboratively train a global model without transmitting private datasets to the FL server. Unlike most existing research, this paper introduces a Data-Driven FL system in Unmanned Aerial Vehicle (UAV) networks, named DDFL, which features an innovative three-layer architecture. In this architecture, UAVs, inherently lacking data, serve as mobile clients and are tasked with periodically collecting desired data from a group of Points of Interest (PoIs) for FL training. Meanwhile, a Base Station (BS) coordinates these UAVs to efficiently train a high-quality global model. Our objective is to determine a data collection strategy for each UAV to minimize the time required to meet the global model's loss requirement within the constraint of energy consumption. Through theoretical analysis, we establish a bound for the convergence speed of DDFL and quantify the impact of collecting data from different PoIs on the global model's loss function. Leveraging these analyses, we formulate the PoI selection problem as a novel two-stage Combinatorial Multi-Armed Bandit (CMAB) problem with multiple constraints. We then propose an Adaptive Two-stage CMAB-based algorithm, named FedATC, to jointly optimize the data collection route and UAV velocity. Extensive simulations demonstrate that FedATC significantly reduces the time required to achieve desired model quality compared to state-of-the-art algorithms. © 2024 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Crime Type and Occurrence Prediction Using Machine Learning 2

Crime Type and Occurrence Prediction Using Machine Learning

引用

2nd IEEE International Conference on Advances in Information Technology, ICAIT 2024

作者： Kavya, T.M. Lavanya, T.N. Ithihas, Maddi Shetty, Namitha R. Amrutha, H.N. Hebbar, Sushruth Department of Computer Science and Engineering Adichunchanagiri Institute of Technology Chikkamagaluru India Department of P.G Studies and Research in Computer Science Kuvempu University Karnataka Shivamogga India

ISBN: (纸本)9798350383867

Nowadays, crime has become a way to get people and people into trouble. Rising crime has fueled unrest at polling stations across the country. Understanding crime patterns is necessary to identify and respond to these criminal activities. This study conducts this crime analysis using crime data obtained from the Kaggle open source which has been used to predict most crimes recently. The main aim of the project is to predict what type of crime is most likely to occur and when and where the crime occurs. This project introduces some machine learning like KNN,SVM to detect various crimes and the accuracy achieved is higher than previous ones. © 2024 IEEE.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Infodemic Detection on Twitter During Pandemic Situation using LDA Topic Modelling 11

Infodemic Detection on Twitter During Pandemic Situation usi...

引用

11th International Conference on System Modeling and Advancement in Research Trends, SMART 2022

作者： Devi, N. Shyamala Sharmila, K. Vels Institute of Science Technology and Advanced Studies Department of Computer Science Chennai India

ISBN: (纸本)9781665487320

Online social media propagates lot of misinformation during pandemic situation which trigger the human community in to a fear. Hence to avoid the propagation of infodemic during pandemic time. We proposed approach unsupervised infodemic identification using latent dirichlet allocation(LDA). This novel approach initially preprocess the unstructured data from twitter to structured data using NLP preprocessing technique then the data is converted to numerical vector using TF-IDF model. Finally the data is topic modeled for identifying the outlier. © 2022 IEEE.

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

The Role of Relocation Policies in Urban Segregation Dynamics

The Role of Relocation Policies in Urban Segregation Dynamic...

引用

Workshops of the EDBT/ICDT 2024 Joint Conference, EDBT/ICDT-WS 2024

作者： Mauro, Giovanni Pappalardo, Luca Institute of Information Science and Technologies"Alessandro Faedo" - ISTI-CNR Pisa Italy Department of Computer Science University of Pisa Italy IMT School for Advanced Studies Lucca Italy

This study addresses a gap in the existing literature on the Schelling segregation model by conducting a comprehensive qualitative assessment of various relocation policies. We introduce novel Schelling models driven by different relocation policies and analyse their impact on the convergence time and final segregation levels. Our findings demonstrate that all policies result in segregation levels within bounds established by policies where agents relocate to maximize their happiness. Notably, a policy ensuring the minimum improvement in agent segregation significantly reduces the model's convergence time. These results underscore the potential influence of relocation policies, such as those employed by online recommenders in real estate platforms, on societal segregation dynamics. The study provides valuable insights into potential strategies for mitigating and decelerating segregation through tailored recommendations. Copyright © 2024 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).

关键词： Autonomous agents

来源：评论

学校读者我要写书评

暂无评论

Benchmarking Large Language Models for Persian: A Preliminary Study Focusing on ChatGPT 30

Benchmarking Large Language Models for Persian: A Preliminar...

引用

Joint 30th International Conference on Computational Linguistics and 14th International Conference on Language Resources and Evaluation, LREC-COLING 2024

作者： Abaskohi, Amirhossein Baruni, Sara Masoudi, Mostafa Abbasi, Nesa Babalou, Mohammad Hadi Edalat, Ali Kamahi, Sepehr Sani, Samin Mahdizadeh Naghavian, Nikoo Namazifard, Danial Sadeghi, Pouya Yaghoobzadeh, Yadollah Computer Science Department University of British Columbia Vancouver Canada School of Electrical and Computer Engineering College of Engineering University of Tehran Tehran Iran Tehran Institute for Advanced Studies Khatam University Tehran Iran

ISBN: (纸本)9782493814104

This paper explores the efficacy of large language models (LLMs) for Persian. While ChatGPT and consequent LLMs have shown remarkable performance in English, their efficiency for more low-resource languages remains an open question. We present the first comprehensive benchmarking study of LLMs across diverse Persian language tasks. Our primary focus is on GPT-3.5-turbo, but we also include GPT-4 and OpenChat-3.5 to provide a more holistic evaluation. Our assessment encompasses a diverse set of tasks categorized into classic, reasoning, and knowledge-based domains. To enable a thorough comparison, we evaluate LLMs against existing task-specific fine-tuned models. Given the limited availability of Persian datasets for reasoning tasks, we introduce two new benchmarks: one based on elementary school math questions and another derived from the entrance exams for 7th and 10th grades. Our findings reveal that while LLMs, especially GPT-4, excel in tasks requiring reasoning abilities and a broad understanding of general knowledge, they often lag behind smaller pretrained models fine-tuned specifically for particular tasks. Additionally, we observe improved performance when test sets are translated to English before inputting them into GPT-3.5. These results highlight the significant potential for enhancing LLM performance in the Persian language. This is particularly noteworthy due to the unique attributes of Persian, including its distinct alphabet and writing styles. We have made our codes, prompts, and data available here: https://***/Ipouyall/Benchmarking_ChatGPT_for_Persian. © 2024 ELRA Language Resource Association: CC BY-NC 4.0.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

Data Acquisition and Image Augmentation using Adversarial Networks

Data Acquisition and Image Augmentation using Adversarial Ne...

引用

2022 International Conference on Edge Computing and Applications, ICECAA 2022

作者： Lathamaheswari, U. Jebathagam, J. Department of Computer Science Vels Institute of Science Technology and Advanced Studies Chennai India

ISBN: (数字)9781665482325

ISBN: (纸本)9781665482325

It is possible to avoid challenges causedby overfitting, and the performance of machine learning algorithms can improve when there is a large amount of data. The improved training data diversity that is offered by data augmentation, which does not require the collection of fresh data, is beneficial to the algorithms that are used in machine learning. In this paper, we collect the data from the associated paddy leaf dataset, however, it is found that the collected data is insufficient for conducting the training and testing of a classifier. In order to increase the samples for training and testing to predict the leaf disease using a machine or a deep learning classifier, it is essential to increase the number of instances for efficient classification. In regards to this, the study uses Generative Adversarial Networks (GANs) to increase the images required for training and testing using image cropping, flipping, color transformation, rotation and noise injection on the collected datasets. The simulation is conducted in python for the generation of images from the input datasets and we store these augmented images for further processing. © 2022 IEEE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Fish Species Classification Using Convolutional Neural Networks

Fish Species Classification Using Convolutional Neural Netwo...

引用

International Conference on IoT, Intelligent Computing and Security, IICS 2021

作者： Fatima, Nishat Yadav, Vrinda Department of Computer Science and Engineering Centre for Advanced Studies Uttar Pradesh Lucknow226031 India

ISBN: (纸本)9789811981357

Human curiosity has recently stretched beyond the ground to include the sky and the water. Automated fish categorization is becoming more popular as a way to accurately identify fish species and traits in a standardized, non-invasive, and cost-effective manner. Deep learning, which uses a variety of neural network techniques to model human brains, can be facilitated by an advanced system with efficient computation. The findings of a convolutional neural network (CNN) used to identify fish species across a large-scale fish dataset are presented in this paper. A cross-layer pooling approach is proposed, which avoids the requirement for a significant quantity of training data by leveraging a convolutional neural network as a generalized feature detector (MobileNetV2) and was successful in achieving an accuracy of 99.89% © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Fish

来源：评论

学校读者我要写书评

暂无评论

Efficient Scheduling of Real-Time Messages over Heterogeneous in-Vehicle Network Domains 100

Efficient Scheduling of Real-Time Messages over Heterogeneou...

引用

100th IEEE Vehicular Technology Conference, VTC 2024-Fall

作者： Chowdary Kota, Ram Mohan Mayank, Jaishree Sarkar, Amab Indian Institute of Information Technology Design and Manufacturing Department of Computer Science and Engineering Kancheepuram Chennai India Indian Institute of Technology Kharagpur Advanced Technology Development Centre Department of Computer Science and Engineering India

ISBN: (纸本)9798331517786

Time Sensitive Ethernet is quickly emerging to be the preferred choice as the backbone network for in-vehicle communication, due to its high bandwidth, reliability, scalability, backward compatibility, and support for diverse traffic types. However, individual real-time control subsystems which may need to communicate via this backbone network, may be driven by non-Ethernet protocols like CAN, FlexRay, etc. Beyond ad hoc approaches, this highlights the need for a systematic message scheduling mechanism that enables seamless real-time message transmission over heterogeneous network domains. In this context, this work proposes a formal SMT (Satisfiability Modulo Theories) formulation for statically scheduling a set of persistent, periodic real-time messages. While the sources and destinations of these messages are in different CAN-network domains, the message transmissions must be conducted via an intermediate Time Sensitive Networking (TSN;IEEE802.1Q) backbone. Although, this formal strategy achieves high resource utilization while guaranteeing end-to-end timeliness, it is computationally exponential in nature with overheads becoming prohibitively expensive even for moderate problem sizes. Hence, we propose a lower overhead scheduling strategy called CAN-THER which can deliver efficient and quick solutions even for large problem sizes. The proposed scheduling strategies have been analyzed and compared using extensive simulation based experiments. Results reveal that CAN-THER is able to achieve performance that is up to 90% of the SMT formulation, while producing solutions at speeds that are approximately 104 times faster for problems having up to 12 CAN flows and 15 TSN switches. © 2024 IEEE.

关键词： Internet protocols

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：