检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Qian, Haosheng Fan, Yixing Zhang, Ruqing Guo, Jiafeng CAS Key Lab of Network Data Science and Technology Institute of Computing Technology Chinese Academy of Sciences Beijing100190 China

Retrieval-augmented generation (RAG) appears as a promising method to alleviate the "hallucination" problem in large language models (LLMs), since it can incorporate external traceable resources for response generation. The essence of RAG in combating the hallucination issue lies in accurately attributing claims in responses to the corresponding retrieved documents. However, most of existing works focus on improving the quality of generated responses from the LLM, while largely overlooked its ability to attribute sources accurately. In this study, we conduct a systematic analysis about the capabilities of LLMs in generating citations within response generation, and further introduce a novel method to enhance their citation generation abilities. Specifically, we evaluate both the correctness and citation quality for seven widely-used LLMs on two benchmark datasets. Meanwhile, we introduce new citation evaluation metrics to eliminate the over-penalization of unnecessary and excessive citations in existing metrics. Furthermore, we propose a Generate-then-Refine method that completes relevant citations and removes irrelevant ones without altering the response text. The results on WebGLM-QA, ASQA and ELI5 datasets show that our method substantially improves the quality of citations in responses generated by LLMs. Copyright © 2024, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

CurricularVPR: Curricular Contrastive Loss for Visual Place Recognition

CurricularVPR: Curricular Contrastive Loss for Visual Place ...

引用

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

作者： Dongshuo Zhang Nanhua Chen Meiqing Wu Siew-Kei Lam College of Computing and Data Science Nanyang Technological University Singapore School of Aerospace Engineering Beijing Institute of Technology Beijing China

ISBN: (数字)9798350377705

ISBN: (纸本)9798350377712

Visual Place Recognition (VPR) techniques commonly utilize Contrastive Losses (CL) to train models that generate compact and discriminative global descriptors for images. These models often result in poor performance due to one of the following reasons during training: 1) loss functions that focus primarily on easier samples, 2) reliance on time-consuming hard sample mining methods to identify informative supervisory samples, which hinders effective learning from large-scale datasets. To enhance both learning efficiency and effectiveness, we propose a Curricular Contrastive Loss (CCL) and use graded similarity labels as a measure of sample difficulty. Inspired by human learning that begin with easier concepts and progressively tackle more challenging ones, our CCL dynamically emphasizes easier samples during the initial training stages to achieve rapid convergence. The learning gradually focuses on harder samples in later training stages to bolster robustness of the models under challenging conditions. Our proposed method has been extensively evaluated on popular datasets, and the results demonstrate its superior performance compared to the CL and Generalized CL functions.

关键词： Training Accuracy Big data Loss measurement Robustness Intelligent robots Visual place recognition Convergence

来源：评论

学校读者我要写书评

暂无评论

Fast waveform generation for gravitational waves using evolutionary algorithms

引用

Physical Review D 2024年第4期110卷 043035页

作者： Quirijn Meijer Sarah Caudill Institute for Gravitational and Subatomic Physics (GRASP) Department of Physics Science Park 105 1098XG Amsterdam The Netherlands Department of Physics Center for Scientific Computing and Data Science Research

Gravitational-wave analyses depend heavily on waveforms that model the evolution of compact binary coalescences as seen by observing detectors. In many cases these waveforms are given by waveform approximants, models that approximate the amplitude and phase of the waveform at a set of frequencies. Because of their omnipresence, improving the speed at which approximants can generate waveforms is crucial to accelerating the overall analysis of gravitational-wave detections. An optimization algorithm is proposed that can select at which frequencies in the spectrum an approximant should compute the power of a waveform, and at which frequencies the power can be safely interpolated at a minor loss in accuracy. The algorithm used is an evolutionary algorithm modeled after the principle of natural selection, iterating frequency arrays that perform better at every iteration. As an application, the candidates proposed by the algorithm are used to reconstruct signal-to-noise ratios. It is shown that the imrphenomxphm approximant can be sped up by at least 30% at a loss of at most 2.87% on the drawn samples, measured by the accuracy of the reconstruction of signal-to-noise ratios. The behavior of the algorithm as well as lower bounds on both speedup and error are explored, leading to a proposed proof of concept candidate that obtains a speedup of 46% with a maximum error of 0.5% on a sample of the parameter space used.

关键词： General relativity Gravitational waves

来源：评论

学校读者我要写书评

暂无评论

Deep Learning for Channel Prediction in Non-Stationary Wireless Fading Environments

Deep Learning for Channel Prediction in Non-Stationary Wirel...

引用

2024 IEEE International Conference on Intelligent Signal Processing and Effective Communication Technologies, INSPECT 2024

作者： Upadhyay, Deepak Sharma, Kunj Bihari Gupta, Mridul Upadhyay, Abhay Venu, Nookala Graphic Era Hill University Dept. of Computer Science and Engineering Dehradun India Graphic Era deemed to be University Dept. of Computer Science and Engineering Dehradun India Graphic Era deemed to be University Dept. of Electronics and Communication Engineering Dehradun India Data Science and Quantum Computing ABV Indian Institute of Information Technology and Management Gwalior India

ISBN: (纸本)9798350379525

In non-stationary wireless fading channels deep learning models are effective in predicting channel states as indicated by this study. This work compared the LSTM and GRU models with autoregressive methods or Kalman filter, implying that deep learning is indeed able to capture temporal dependence between variables and was likely more adapted than linear method when dealing with many changes in dynamics. From the perspective of pre-processing data, keys like attention mechanism and transfer learning were decisive elements in helping them to improve prediction scores and make models more robust. These results showed that these enhanced procedures allow the models to generalise over a wide range of non-stationary situations, generating consistent performance in various diverse environments. This study indicates the promise of deep learning for enhancing wireless network operation, which could lead to greater self-awareness and utilization features in emerging future-generation wireless systems. © 2024 IEEE.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Revolutionizing Public Transport: RFID-Enabled Occupancy Monitoring and Frequency Analysis System

Revolutionizing Public Transport: RFID-Enabled Occupancy Mon...

引用

International Conference on Communication, computing and Internet of Things (IC3IoT)

作者： J. Uma Mageswaran G. Indumathi A.V. Kalpana G. Elangovan S.D. Lalitha S. Nalini Mayank Kumar Sohanda Dept. of Computer Science Engineering Amrita School of Computing Vengal Campus Tiruvallur Dist Tamil Nadu SRM Institute of Science and Technology Ramapuram Department of Data Science and Business Systems SRM Institute of Science and Technology Kattankulathur Department of Data Science and Business Systems SRM Institute of Science and Technology Department of Computer Science and Engineering R.M.K Engineering College Department of Computing Technologies SRM Institute of Science and Technolog Kattankulathur

ISBN: (数字)9798350352689

ISBN: (纸本)9798350352696

Public bus stops in India are becoming more crowded due to the country's fast population expansion. People wait a long time for buses to come, then suddenly congregate around them when they do, packing the buses with people and pushing them up into the footboards, which causes accidents. An additional concern of overcrowding is theft. All of this is the result of numerous bus stops not having adequate information on when busses would arrive. The proposed work paves a very important role for the passengers. The passengers will be provided with the information of the arrival of next bus, seat occupancy and the total passengers inside the bus. The information about the seat occupancy, in-passengers and out-passengers is done with the help of Radio Frequency Identification (RFID) tags. The arrival of the next bus is determined by calculating the duration of the buses that are near using RFID readers placed in different bus stations. The count and the frequency will be interfaced in the web application. The passengers can check the application to know the occupancy and today's society where people can catch the right bus at the station and at the right time. This will solve the problems like waiting in bus stops and wasting time and going in a crowded bus.

关键词： Sociology Telecommunication computing Internet of Things Statistics RFID tags Monitoring Accidents

来源：评论

学校读者我要写书评

暂无评论

TransUNet for Image Forgery Localization 3

TransUNet for Image Forgery Localization

引用

3rd International Conference on Computer Graphics, Image, and Virtualization, ICCGIV 2023

作者： Li, Kedi Li, Lin Li, Qingyan School of Science Jimei University Xiamen361021 China Computer Engineering College Jimei University Xiamen361021 China Digital Fujian Big Data Modeling and Intelligent Computing Institute Jimei University Xiamen361021 China

ISBN: (纸本)9781510671720

Image tampering can easily be used in illegal activities such as false propaganda, fake news and falsifying evidence in court which may have a negative impact on society. Therefore, we need to constantly update and improve the image tampering detection technology. TransUNet is an efficient model for medical image segmentation. This paper modified and migrated TransUNet which to image forgery localization, and conducted experiments on standard datasets to demonstrate the effectiveness of TransUNet. The findings shows that the proposed framework is superior to several existing forged area localization techniques. © 2023 SPIE.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

Performance Evaluation of Optical Wireless Communication System in Fog Weather Conditions 2

Performance Evaluation of Optical Wireless Communication Sys...

引用

2nd International Conference on Electronics and Renewable Systems, ICEARS 2023

作者： Venusamy, Kanagaraj Kannadhasan, S. Prakash, E. Priyadarsini, K. Study World College of Engineering Department of Electronics and Communication Engineering Chennai Coimbatore India Rajalakshmi Engineering College Department of Mechatronics Engineering Chennai India School of Computing Srm Institute of Science & Technology Department of Data Science and Business Systems Chennai Kattankulathur India

ISBN: (纸本)9798350346640

Free space optical communication being an emerging technology, has Fog as a primary challenge on its way which makes fog modelling necessary to manage wireless networks. In this paper, an analytical study was made on the performance of five mathematical fog models employing the NRZ-OOK modulation scheme for transmission and a Si PIN photodiode as the photo receptor. This article focuses on numerous performance specifications including Signal to noise ratio [SNR], Bit error rate [BER] and channel capacity all against visibility under fog for propagations wavelengths of 850 nm and 1550 nm and simulations were carried out in MATLAB. The results indicate that the wavelength of 1550nm as more advantageous option over the others for FSO wireless links and an investigation on different models and their characteristics define their suitability for implementation. © 2023 IEEE.

关键词： Visibility

来源：评论

学校读者我要写书评

暂无评论

Analysis of the Influence of Green Artificial Intelligence on Agricultural and Biological Fields Using Bibliometrics

Analysis of the Influence of Green Artificial Intelligence o...

引用

International Conference on Artificial Intelligence and data sciences (AiDAS)

作者： Jonitha Anand Hadhrami Ab Ghani Nooraini Yusoff Kiran Kumar Thoti Faculty of Data Science & Computing Universiti Malaysia Kelantan Malaysia M.S. Ramaiah Institute of Management Bangalore India

ISBN: (数字)9798331528553

ISBN: (纸本)9798331528560

The research examines how Green Artificial Intelligence (AI) impacts agricultural and biological domains by analysing relevant literature using bibliometrics. Green AI, which prioritizes sustainability and environmental awareness, is proving to be an asset in tackling the issues in agriculture and biology. The investigation examines a wide range of academic publications, patents, and research papers to identify significant trends, research topics, and the influence of Green AI on agricultural and biological fields. The research shows an increasing amount of literature investigating the combination of AI technology with green principles to improve resource management, boost agricultural output, and reduce environmental effects. This study uses analysis of citation networks, co-authorship patterns, and theme grouping to get insights into the present status and future trends of research in the fields of Green AI, agriculture, and biology. The data was extracted from the database maintained by Scopus using the PRISM Model. 11,142 article data points were used for the analysis of Green AI. This bibliometric study enhances comprehension of the changing environment of Green AI applications in agricultural and biological areas, providing vital recommendations for academics, policymakers, and practitioners working towards sustainable development objectives.

关键词： Green products Bibliometrics Air pollution Market research Biology Agriculture Resource management Artificial intelligence Sustainable development Farming

来源：评论

学校读者我要写书评

暂无评论

MP-GAN: Cyber-Attack Detection and Localization for Cyber-Physical Systems with Multi-Process Generative Adversarial Networks*

MP-GAN: Cyber-Attack Detection and Localization for Cyber-Ph...

引用

2023 International Conference on Artificial Intelligence of Things and Systems, AIoTSys 2023

作者： Zhou, Yikui Wang, Jie Tang, Junnan Gou, Chao Jiang, Zigui Li, Dan Ng, See-Kiong Sun-Yat Sen University School of Software Engineering Zhuhai China Sun-Yat Sen University School of Intelligent Systems Engineering Shenzhen China National University of Singapore Institute of Data Science School of Computing Singapore

ISBN: (纸本)9798350312270

Cyber-Physical System (CPS) integrates sensing, computation, cybernetics, and networking to control a hybrid physical system consisting of different functional subsystems, making the production process more intelligent and controllable. However, cyber-attacks during its operation will lead to abnormal system behaviors or even system breakdowns. In recent years, data-driven anomaly detection methods have been adopted to judge whether the CPS system is under cyber-attacks based on rich sensor measurements to avoid further economic losses or safety issues. However, the multi-process essence of CPS has not been adequately addressed in existing works to locate the processes or points being attacked for in-time actions. In this work, we proposed a Multi-Process Generative Adversarial Network (MP-GAN) framework to detect anomalous CPS statuses and locate cyber-attacks. Specifically, the Long-Short-Term-Memory Recurrent Neural Networks (LSTM-RNN) was adopted as the base model to capture the temporal correlation of time series distributions, and the data was transferred between latent space and data space in a bi-directional manner, which regulated the generator to generate more realistic samples and thus better grasping the underline principles of the system. Moreover, parallel generators were employed to capture system performances at different physical processes, thus localizing the attacked CPS processes. Experiments on three CPS datasets, two collected from the Secure Water Treatment (SWaT) system and one collected from the water Distribution (WADI) system, showed that the proposed MP-GAN framework effectively reported anomalies caused by various cyber-attacks inserted in these complex multiprocess CPSs, which outperforms the state-of-the-art methods and can also locate most of the detected irregularities at the corresponding stages where the attacks were inserted. © 2023 IEEE.

关键词： Cyber Physical System

来源：评论

学校读者我要写书评

暂无评论

Energy Price Forecasting Through Novel Fuzzy Type-1 Membership Functions

引用

Computers, Materials & Continua 2022年第10期73卷 1799-1815页

作者： Muhammad Hamza Azam Mohd Hilmi Hasan Azlinda A Malik Saima Hassan Said Jadid Abdulkadir Centre of Research in Data Science Computerand Information Sciences DepartmentUniversiti Teknologi PETRONASSeri Iskandar32610PerakMalaysia Petroleum Engineering Department Universiti Teknologi PETRONASSeri Iskandar32610PerakMalaysia Institute of Computing Kohat University of Science and TechnologyKohat26000K.P.KPakistan

Electricity price forecasting is a subset of energy and power forecasting that focuses on projecting commercial electricity market present and future *** price forecasting have been a critical input to energy corporations’strategic decision-making systems over the last 15 *** strategies have been utilized for price forecasting in the past,however Artificial Intelligence Techniques(Fuzzy Logic and ANN)have proven to be more efficient than traditional techniques(Regression and Time Series).Fuzzy logic is an approach that uses membership functions(MF)and fuzzy inference model to forecast future electricity *** c-means(FCM)is one of the popular clustering approach for generating fuzzy membership ***,the fuzzy c-means algorithm is limited to producing only one type of MFs,Gaussian *** generation of various fuzzy membership functions is critical since it allows for more efficient and optimal problem *** a result,for the best and most improved results for electricity price forecasting,an approach to generate multiple type-1 fuzzy MFs using FCM algorithm is ***,the objective of this paper is to propose an approach for generating type-1 fuzzy triangular and trapezoidal MFs using FCM algorithm to overcome the limitations of the FCM *** approach is used to compute and improve forecasting accuracy for electricity prices,where Australian Energy Market Operator(AEMO)data is *** results show that the proposed approach of using FCM to generate type-1 fuzzy MFs is effective and can be adopted.

关键词： Fuzzy logic fuzzy C-means type-1 fuzzy membership function electricity price forecasting

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：