检索结果-内蒙古大学图书馆

5th International Conference on Artificial Intelligence and Computer Engineering, ICAICE 2024

作者： Qiu, Yaru Zhang, Jianfang Wu, Guoxia Sun, Yuanyuan School of Data Science Qingdao University of Science and Technology Qingdao China School of Information Science and Technology Qingdao University of Science and Technology Qingdao China

ISBN: (纸本)9798331528911

Video pose transformers (VPTs) have demonstrated remarkable performance in 3D human pose prediction. However, transformer-based architectures are often computationally intensive, leading to prolonged training times. To address this issue, we propose an efficient and plug-and-play module named the Adaptive Token Management (ATM), to prune and recover pose tokens, enabling high efficiency and accurate predictions within the VPTs framework. ATM consists of two key modules: the Dynamic Token Pruning (DTP) and the Lightweight Token Recovery (LTR). The DTP is presented for dynamic tokens pruning, where it determines the optimal number of representative tokens based on feature distribution and selects the representative tokens through a scoring mechanism. The LTR module, an optional lightweight module, is designed for tasks requiring the recovery of full-length tokens. It selects an appropriate full-length token as a learnable initialization token, utilizing attention mechanisms to restore spatio-temporal information. We conducted extensive experiments on the Human3.6M and MPI-INF-3DHP datasets, demonstrating that our approach can enhance the efficiency of mainstream VPTs while maintaining prediction accuracy. The results show that, when using MixSTE as the underlying framework on the Human3.6M dataset, ATM accelerated training time per epoch, reduced FLOPs by over 40%, and decreased prediction error by 0.2. © 2024 IEEE.

关键词： Distribution transformers

来源：评论

学校读者我要写书评

暂无评论

Automated Medicinal Plant Identification through Image Processing and Machine Learning 8

Automated Medicinal Plant Identification through Image Proce...

引用

8th IEEE International Conference on Computing, Communication, Control and Automation, ICCUBEA 2024

作者： Kadam, Yash Yache, Ajay Solunke, Pavankumar Tati, Ramakant Kachhoria, Renu Buchade, Amar Vishwakarma Institute of Information Technology Artificial Intelligence and Data Science Pune India

ISBN: (纸本)9798350391770

This research addresses the pressing challenge of accurately identifying medicinal plants in India's Ayurvedic Pharmaceutics, aiming to mitigate issues stemming from morphological variability and market adulteration. Traditional methods of plant identification are hindered by laborious processes and the potential for error, underscoring the need for an automated solution. Leveraging the power of advanced image processing and machine learning, our proposed software framework integrates the Xception model and early stopping fine-tuning to automate plant identification based on plant images. The methodology involves curating a comprehensive dataset of labelled plant images representing diverse medicinal plants, subsequently employing image preprocessing techniques to extract salient features like color, texture, and shape. The Xception model, renowned for its prowess in image recognition tasks, serves as the backbone architecture for both feature extraction and classification. To further optimize model performance and mitigate overfitting, early stopping fine-tuning techniques are applied. Through rigorous experimentation and validation, our research demonstrates the efficacy of this approach in accurately identifying medicinal plants, thereby fostering confidence and trust within the herbal medicine supply chain. By seamlessly integrating modern technological advancements with traditional botanical knowledge, our framework not only streamlines the identification process but also contributes to the sustainable utilization of herbal medicine resources. Ultimately, this research represents a pivotal step towards establishing a robust and reliable system for identifying medicinal plants, facilitating informed decision-making, and promoting the long-term viability of herbal medicine practices in India and beyond. © 2024 IEEE.

关键词： Medicinal chemistry

来源：评论

学校读者我要写书评

暂无评论

WheatNGP: Neural Radiation Field Optimization for Wheat Flocks Under Boundless Free Trajectories 5

WheatNGP: Neural Radiation Field Optimization for Wheat Floc...

引用

5th International Conference on Artificial Intelligence and Computer Engineering, ICAICE 2024

作者： Tian, Liming Qiu, Yaru Wu, Guoxia Sun, Yuanyuan School of Data Science Qingdao University of Science and Technology Qingdao China School of Information Science and Technology Qingdao University of Science and Technology Qingdao China

ISBN: (纸本)9798331528911

Accurate NeRF synthesis can help plant breeders to effectively analyze and identify plant traits to improve crop yield. Most of the existing mesh-based NeRF new perspective synthesis methods are targeted at bounded scenes, such as Instant-NGP, Plenoxels, DVGO, and tensorf, but wheat is usually grown in outdoor field environments, which belongs to non-boundary scenes. At present, the widely used method is to rely on space warping, but it can only deal with forward trajectory and 360° center surround trajectory, which is a special trajectory, and can't deal with arbitrary trajectory, but there are more free trajectories in the daily shooting process. In order to solve this problem, we propose a new warping function called Variable warping. In this study, we use cell phone to shoot wheat experimental field data, and use the improved Variable warping algorithm to synthesize the new perspective of wheat under free trajectory. The experimental results show that it allows us to reconstruct the new perspective of wheat with high quality under arbitrary camera trajectory. The experimental results show that allows us to perform high-quality reconstruction of wheat under arbitrary camera trajectories with new perspective synthesis. © 2024 IEEE.

关键词： Borderless component Free Trajectories Novel view synthesis of wheat

来源：评论

学校读者我要写书评

暂无评论

Knowledge Graph Reasoning Neural Network Based on Dynamic Programming 4

Knowledge Graph Reasoning Neural Network Based on Dynamic Pr...

引用

4th International Symposium on Computer technology and information science, ISCTIS 2024

作者： Mei, Tong Zhang, Jiashuo Xu, Zipeng Chu, Shuqi Hu, Zili Qingdao University of Science and Technology School of Information Science and Technology Qingdao China Qingdao University of Science and Technology School of Data Science Qingdao China

ISBN: (纸本)9798350354560

Knowledge graph is a semantic network that abstracts data into points and edges, which contains logical and structural information. In order to mine the potential facts, the reasoning algorithm of the knowledge graph came into being. However, current knowledge graph reasoning algorithms usually use path or graph neural networks for reasoning. The former will cause the problem of time and space complexity explosion, while the latter will cause the problem of lack of interpretability, so it cannot be well applied to real life. In order to solve these two problems, we proposed a knowledge graph reasoning network model based on dynamic programming, which uses a recursive method to record path information, reduces the time and space complexity from exponential level to linear level, and the reasoning path is preserved when passing messages, so it is interpretable while keeping time and space complexity low. In addition, we use PNA aggregation to replace the aggregation method of traditional graph neural network, making the model performance even better. Experiments have proven that the network model we proposed not only surpasses the state-of-the-art performance on the WN18RR and NELL-995 data sets, but also has a spatio-temporal complexity far lower than other reasoning algorithms. © 2024 IEEE.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Hybrid speech enhancement in modulation domain

引用

Multimedia Tools and Applications 2024年 1-31页

作者： P S, Praveen Kumar H S, Jayanna Technical Lead-Machine Learning Merit Data and Technology Tamil Nadu Chennai India Department of Information Science and Engineering Siddaganga Institute of Technology Karnataka Tumkur India

This paper presents a comprehensive study on speech enhancement (SE) techniques, particularly focusing on the utilization of the discrete cosine transform (DCT) in the modulation domain (MD) in combination with the minimum mean square error (MMSE) and Kalman filter (KF) algorithms. The research investigates the efficacy of these techniques in improving speech quality and intelligibility under various noisy conditions, such as babble, drilling, horn, train, and traffic noises, across different signal-to-noise ratio (SNR) levels. Additionally, the study evaluates the performance of recurrent neural network (RNN) algorithms alongside traditional SE approaches. The study highlights the advantages of employing DCT in the modulation domain over the traditional discrete Fourier transform (DFT) in SE applications. Experimental results demonstrate significant improvements in objective quality metrics, including Perceptual Evaluation of Speech Quality (PESQ), composite measures, and short-time objective intelligibility (STOI), when implementing SE algorithms using discrete cosine transform in modulation domain. The proposed hybrid speech enhancement techniques, leveraging minimum mean square error and Kalman filter in combination with discrete cosine transform in modulation domain, outperform individual speech enhancement algorithms, showcasing superior noise reduction capabilities. This paper contributes to the advancement of speech enhancement methodologies, particularly in real-world noisy environments, and underscores the effectiveness of discrete cosine transform-based approaches in enhancing speech quality and intelligibility. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

Implementing Hadoop and Moodle Integrated Student Support System 5

Implementing Hadoop and Moodle Integrated Student Support Sy...

引用

5th International Conference on Advanced information Technologies, ICAIT 2024

作者： Aung, Myint Myat Min Htet, Shin Set, Hein Myo Hpone Thant, Aung Htut, Sithu Htet, Kyaw Zin Nwe Myint Thein, Nwe University of Information Technology Data Distribution Management Lab Yangon Myanmar University of Information Technology Data Distribution Management Lab Faculty of Computer Science Yangon Myanmar

ISBN: (纸本)9798331510046

Learning management systems (LMS) generate large amounts of real-time data which has the potential to provide valuable insights. Hadoop MapReduce is a scalable framework that is designed to handle such data. This paper presents a system architecture that integrates Hadoop MapReduce into Moodle, a popular LMS, for analyzing academic data. The implementation of the architecture allows users to generate a variety of reports from student attendance logs through a Moodle plugin which utilizes the Hadoop cluster for processing data. Users such as lecturers can then utilize these reports to identify which students do and do not engage in which courses as well as monitor their eligibility for exams. The use of Hadoop MapReduce provides high performance computing (HPC) capabilities that offer distributed parallel processing and economical scalability by means of adding commodity nodes to the cluster. © 2024 IEEE.

关键词： MapReduce

来源：评论

学校读者我要写书评

暂无评论

A Digital Twin and Consensus Empowered Cooperative Control Framework for Platoon-Based Autonomous Driving

引用

清华大学学报自然科学版（英文版） 2025年第3期30卷 1096-1111页

作者： Jiayu Cao Supeng Leng Kai Xiong Xiaosha Chen School of Information and Communication Engineering University of Electronic Science and Technology of ChinaChengdu 611731China Shenzhen Institute for Advanced Study University of Electronic Science and Technology of ChinaShenzhen 518110China College of Computer Science and Technology and also with College of Data Science Taiyuan University of TechnologyTaiyuan 030024China

Platoon-based autonomous driving is indispensable for traffic automation,but it confronts substantial constraints in rugged terrains with unreliable links and scarce communication *** paper proposes a novel hierarchical Digital Twin(DT)and consensus empowered cooperative control framework for safe driving in harsh ***,leveraging intra-platoon information exchange,one platoon-level DT is constructed on the leader and multiple vehicle-level DTs are distributed among platoon *** leader first makes critical platoon-driving decisions based on the platoon-level ***,considering the impact of unreliable links on the platoon-level DT accuracy and the consequent risk of unsafe decision-making,a distributed consensus scheme is proposed to negotiate critical decisions *** successful negotiation,vehicles proceed to execute critical decisions,relying on their vehicle-level ***,a Space-Air-Ground-Integrated-Network(SAGIN)enabled information exchange is utilized to update the platoon-level DT for subsequent safe decision-making in scenarios with unreliable links,no roadside units,and obstructed ***,based on this framework,an adaptive platooning scheme is designed to minimize total delay and ensure driving *** results indicate that our proposed scheme improves driving safety by 21.1％and reduces total delay by 24.2％in harsh areas compared with existing approaches.

关键词： platoon autonomous driving distributed consensus digital twin space-air-ground integrated network

来源：评论

学校读者我要写书评

暂无评论

Modeling the Percentage of NEET in Indonesia with Spatial Cauchy Regression through the Bayesian Analysis Approach

引用

IAENG International Journal of Applied Mathematics 2024年第7期54卷 1288-1295页

作者： Rantini, Dwi Fakhruzzaman, Muhammad Noor Ningrum, Ratih Ardiati Othman, Fazidah Choir, Achmad Syahrul Ramadan, Arip Alya, Najma Attaqiya Putri, Elfira Rahma Pratama, Muhammad Alfian Engineering Department Data Science Technology Study Program Faculty of Advanced Technology and Multidiscipline Universitas Airlangga Indonesia Engineering Department Data Science Technology Study Program Faculty of Advanced Technology andMultidiscipline Universitas Airlangga Indonesia Department of Computer System and Technology Faculty of Computer Science and Information Technology University of Malaya Malaysia Statistical Computing StudyProgram Politeknik Statistika STIS Indonesia Information System Study Program School of Industrial and System Engineering Telkom University Indonesia

Indonesia has entered a period of demographic bonus. Human resources must be optimized. The number of children who do not in employment, education or training (NEET) in each province needs attention. Several factors that could contribute to the decline in the NEET percentage are literacy rates and the number of adolescents with computer skills. Increasing these two factors is believed to be able to reduce the percentage of NEET in every province in Indonesia. To find out the relationship between these two factors and how much influence they have on the percentage of NEET, this research is modeled by Cauchy regression and includes spatial effects. The results of the analysis show that the best model is when the spatial effect is modeled by Fernandez Steel Skew Normal conditionally autoregressive (FSSN CAR). This result is seen from the smallest value of the Watanabe Akaike information Criterion (WAIC) in this model, which is 190.5. The parameter estimated shows that the higher the literacy rate and the number of adolescents with computer skills, the lower the percentage of NEET in each province in Indonesia. The results of this research can be useful for the Indonesian government to increase the number of educational facilities related to these two factors. © (2024), (International Association of Engineers). All rights reserved.

关键词： Bayesian CAR Cauchy Education NEET

来源：评论

学校读者我要写书评

暂无评论

TIPS:Tailored information Extraction in Public Security Using Domain-Enhanced Large Language Model

引用

Computers, Materials & Continua 2025年第5期83卷 2555-2572页

作者： Yue Liu Qinglang Guo Chunyao Yang Yong Liao School of Cyber Science and Technology University of Science and Technology of ChinaHefei230026China National Engineering Research Center for Public Safety Risk Perception and Control by Big Data China Academy of Electronics and Information TechnologyBeijing100041China

Processing police incident data in public security involves complex natural language processing(NLP)tasks,including information *** data contains extensive entity information—such as people,locations,and events—while also involving reasoning tasks like personnel classification,relationship judgment,and implicit ***,utilizing models for extracting information from police incident data poses a significant challenge—data scarcity,which limits the effectiveness of traditional rule-based and machine-learning *** address these,we propose *** collaboration with public security experts,we used de-identified police incident data to create templates that enable large language models(LLMs)to populate data slots and generate simulated data,enhancing data density and *** then designed schemas to efficiently manage complex extraction and reasoning tasks,constructing a high-quality dataset and fine-tuning multiple open-source *** showed that the fine-tuned ChatGLM-4-9B model achieved an F1 score of 87.14%,nearly 30%higher than the base model,significantly reducing error *** corrections further improved performance by 9.39%.This study demonstrates that combining largescale pre-trained models with limited high-quality domain-specific data can greatly enhance information extraction in low-resource environments,offering a new approach for intelligent public security applications.

关键词： Public security information extraction large language model prompt engineering

来源：评论

学校读者我要写书评

暂无评论

Attack based on data: a novel perspective to attack sensitive points directly

引用

Cybersecurity 2024年第3期7卷 111-123页

作者： Yuyao Ge Zhongguo Yang Lizhe Chen Yiming Wang Chengyang Li School of Information Science and Technology North China University of TechnologyBeijingChina Bejing Key Laboratory on Integration and Analysis of Large-Scale Stream Data North China University of TechnologyBeijingChina School of Computer Science Peking UniversityBeijingChina

Adversarial attack for time-series classification model is widely explored and many attack methods are *** there is not a method of attack based on the data *** this paper,we innovatively proposed a black-box sparse attack method based on data *** method directly attack the sensitive points in the time-series data accord-ing to statistical features extract from the *** frst,we have validated the transferability of sensitive points among DNNs with different ***,we use the statistical features extract from the dataset and the sensi-tive rate of each point as the training set to train the predictive ***,predicting the sensitive rate of test set by predictive ***,perturbing according to the sensitive *** attack is limited by constraining the LO norm to achieve one-point *** conduct experiments on several datasets to validate the effectiveness of this method.

关键词： Black-box adversarial attack Time series classification data mining

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：