检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Li, Xiaomin Gao, Mingye Zhang, Zhiwei Yue, Chang Hu, Hong John A. Paulson School of Engineering and Applied Sciences Harvard University United States Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology United States College of Information Sciences and Technology The Pennsylvania State University United States School of Electrical and Computer Engineering Princeton University United States McKelvey School of Engineering Washington University in St. Louis United States

The quality of training data significantly impacts the performance of large language models (LLMs). There are increasing studies using LLMs to rate and select data based on several human-crafted metrics (rules). However, these conventional rule-based approaches often depend too heavily on human heuristics, lack effective metrics for assessing rules, and exhibit limited adaptability to new tasks. In our study, we introduce an innovative rule-based framework that utilizes the orthogonality of score vectors associated with rules as a novel metric for rule evaluations. Our approach includes an automated pipeline that first uses LLMs to generate a diverse set of rules, encompassing various rating dimensions to evaluate data quality. Then it rates a batch of data based on these rules and uses the determinantal point process (DPP) from random matrix theory to select the most orthogonal score vectors, thereby identifying a set of independent rules. These rules are subsequently used to evaluate all data, selecting samples with the highest average scores for downstream tasks such as LLM training. We verify the effectiveness of our method through two experimental setups: 1) comparisons with ground truth ratings and 2) benchmarking LLMs trained with the chosen data. Our comprehensive experiments cover a range of scenarios, including general pre-training and domain-specific fine-tuning in areas such as IMDB, Medical, math, and Code. The outcomes demonstrate that our DPP-based rule rating method consistently outperforms other approaches, including rule-free rating, uniform sampling, importance resampling, and QuRating, in terms of both rating precision and model performance. Copyright © 2024, The Authors. All rights reserved.

关键词： Importance sampling

来源：评论

学校读者我要写书评

暂无评论

MSKI-Net: Towards modality-specific knowledge interaction for glioma survival prediction

MSKI-Net: Towards modality-specific knowledge interaction fo...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Ran Su Jin Xiao Hui Cui Ping Xuan Xikang Feng Leyi Wei Qiangguo Jin School of Computer Software College of Intelligence and Computing Tianjin University Tianjin China Department of Computer Science and Information Technology La Trobe University Melbourne Australia Department of Computer Science School of Engineering Shantou University China School of Software Northwestern Polytechnical University China AIDD Faculty of Applied Science Macao Polytechnic University Macao SAR China School of Information Xiamen University Xiamen China Yangtze River Delta Research Institute of Northwestern Polytechnical University Taicang China

ISBN: (数字)9798350386226

ISBN: (纸本)9798350386233

Gliomas hold a prominent position in neurooncology due to their high malignancy and poor survival rates. Accurately predicting the prognosis and survival risk of glioma patients is crucial for clinical treatment. Recent advances in survival prediction methods emphasize the importance of integrating complementary information from diverse modalities while neglecting the significant modality gap between pathological images and genomic data. To address this issue, we propose a modality-specific knowledge interaction network (MSKI-Net), which integrates whole slide images (WSI), RNA-Seq gene expression data, and copy number variation (CNV) data for glioma survival analysis. The MSKI-Net consists of a modality-specific feature enhancement (MSFE) module, a modality-interactive cross-attention (MICA) module, and a modality-specific knowledge-guided representation learning (MSKR) module. The three modules collaborate by complementing modality-specific features with modality-agnostic knowledge to improve the learning capability of MSKI-Net. Furthermore, we construct a dataset named TCGA mm , which combines WSI, RNA-Seq, and CNV data from The Cancer Genome Atlas (TCGA) to address the issue of data scarcity. Extensive experiments demonstrate that MSKI-Net achieves superior performance in predicting the survival risk of glioma cancer.

关键词： Knowledge engineering Representation learning Pathology Genomics Prediction methods Data models Gene expression Bioinformatics Prognostics and health management Cancer

来源：评论

学校读者我要写书评

暂无评论

UAV-LiDAR and Terrestrial Laser Scanning for Automatic Extraction of Forest Inventory Parameters

UAV-LiDAR and Terrestrial Laser Scanning for Automatic Extra...

引用

Virtual International Conference on Soft Computing, Optimization Theory and Applications, SCOTA 2021

作者： Meghraoui, Khadija Lfalah, Hamza Sebari, Imane Kellouch, Souhail Fadil, Sanaa Ait El Kadi, Kenza Bensiali, Saloua Research Unit of Geospatial Technologies for A Smart Decision Hassan II Institute of Agronomy and Veterinary Medicine Rabat Morocco Cartography–Photogrammetry Department School of Geomatics and Surveying Engineering Hassan II Institute of Agronomy and Veterinary Medicine Rabat Morocco AXIGEO Marrakech4040 Morocco Department of Water and Forests Rabat Morocco Department of Applied Statistics and Computer Science Hassan II Institute of Agronomy and Veterinary Medicine Rabat Morocco

ISBN: (纸本)9789811964053

The determination of the Dendrometric parameters of forest stands has a silvicultural and ecological interest for the forester, in particular for the evaluation of the dynamics of growth and productivity, and the evaluation of indicators of good ecological status. Currently, UAV-LiDAR (Unmanned Aerial Vehicle-Light Detection and Ranging) has become the new trend for measurement professionals, offering very high-resolution data collection at considerably lower survey costs. In addition, this technology has started to prove its utility in forest inventory applications namely to extract dendrometric parameters, where direct and conventional measurements are sometimes difficult. As for the TLS (Terrestrial Laser Scanning) technology, it has made it possible to obtain several abundant and refined structural information under the forest canopy. In the context of extraction of forest inventory parameters, the precision of extracting tree height for example using TLS alone, is insufficient. Hence the contribution of the combination of ALS (Aerial Laser Scanning) with TLS data to fill any information gaps that may exist. The main goal of this study is to present an approach to the automatic extraction of dendrometric parameters from UAV-LiDAR and TLS data. The proposed methodology is based on performing a TLS survey at a plot level and an ALS scan of the entire area. Our methodology is essentially made up of two steps: automatic crown delineation and automatic extraction of dendrometric parameters (position, Diameter at breast height, height, stem curve, concave and convex hull). For the first step, we compared the segmentation of the point cloud by the Watershed algorithm and by the SEGMA pipeline. Whereas the extraction of the dendrometric parameters was carried out using a set of algorithms namely RHT (Random Hough Transform) and LSR (Least Square Regression). The study focused on UAV-ALS and TLS datasets from different regions and with different densities (the Mediterran

关键词： Antennas

来源：评论

学校读者我要写书评

暂无评论

Modelling smart energy consumption with hybrid demand management in off-grid electrical system considering techno-economic indices

Science and Technology for Energy Transition (STET)

引用

science and Technology for Energy Transition (STET) 2024年第1期79卷

作者： Rui-Ming, Fang Liao, Qiangqiang Askar, Shavan Pallathadka, Harikumar Sapaev, I.B. Baldaniya, Lalji Chahar, Mamata Saini, Suman Kapila, Ish Jewahery, Hassan Farhan, Mohsen Aued Mohmmedi, M. Information Systems Engineering Department Erbil Technical Engineering College Erbil Polytechnic University Erbil Iraq Manipur International University Manipur Imphal India Department Physics and Chemistry Tashkent Institute of Irrigation and Agricultural Mechanization Engineers Institute National Research University Tashkent Uzbekistan Scientific Researcher University of Tashkent for Applied Sciences Str. Gavhar 1 Tashkent100149 Uzbekistan Scientific Researcher Western Caspian University Baku Azerbaijan Marwadi University Research Centre Department of Pharmacy Faculty of Health Science Marwadi University Gujarat Rajkot360003 India Deparment of Chemistry NIMS Institute of Engineering & Technology NIMS University Rajasthan Jaipur India Department of Applied Sciences Chandigarh Engineering College Chandigarh Group of Colleges Mohali Punjab Jhanjeri140307 India Centre for Research Imapct & Outcome Chitkara University Institute of Engineering and Technology Chitkara University Punjab Rajpura140401 India Department of Computer Techniques Engineering College of Technical Engineering The Islamic University Najaf Iraq Department of Computer Techniques Engineering College of Technical Engineering The Islamic University of Al Diwaniyah Al Diwaniyah Iraq Department of Computer Techniques Engineering College of Technical Engineering The Islamic University of Babylon Babylon Iraq Department of Management Al-Nisour University College Nisour Seq. Karkh Baghdad Iraq Department of Electrical Engineering Islamic Azad University Branch of Central Tehran Tehran Iran

This study proposes day-ahead power scheduling for electrical systems in off-grid mode, emphasizing consumer involvement. Bi-Demand Side Management (DSM) approaches like strategic conversion and demand shifting are proposed for consumer involvement. Multiple objectives are modelled to voltage profile improvement and reduce the operation energy cost. The non-dominated solutions of the voltage of buses and operation energy cost are generated by enhanced epsilon-constraint technique, simultaneously. The General Algebraic Modeling System (GAMS) software is proposed for solving optimization problems. A combination of decision-making methods like weight sum and fuzzy procedures are implemented for finding optimal solution non-dominated solutions. The proposed method's effectiveness is confirmed through numerical simulations carried out on several case studies that utilize the 33-bus electrical system. The findings illustrate the substantial effectiveness of demand-side participation in improving power dispatch and the optimal rate of multiple objectives. By using DSM, operation cost is reduced by 21.58% and the voltage index is improved by 13.36% than the lack of implementing DSM. © The Author(s), published by EDP sciences, 2024.

关键词： Demand response

来源：评论

学校读者我要写书评

暂无评论

Predicting Continuity of Online Conversations on Reddit

SSRN

引用

SSRN 2022年

作者： Kim, Jinhyeon Han, Jinyoung Choi, Daejin Department of Electronic Engineering Incheon National University Incheon Korea Republic of Department of Applied Artificial Intelligence Sungkyunkwan University Seoul Korea Republic of Department of Computer Science and Engineering Incheon National University Incheon Korea Republic of

Online conversation has been facilitating human society, which has encouraged people not only to share news, products, or daily events, but also to develop ideas. A popular type of online conversation is threaded conversation, where a person initiates a conversation with a new topic by uploading a post, then others reply to the post or the comments made by other participants in a recursive manner. Even though the growing importance of understanding and predicting the development process of threaded conversations has attracted research community to investigate continuity of online threaded conversations from a perspective of individual posts or comments, little work has focused on predicting the continuity or persistence of online conversations. In this paper, we propose a deep learning model to predict continuity of threaded conversations - whether there will be a newly-arrived comment or not. Cooperating with popular pre-trained text embedding models and graph neural network models, the proposed model captures text, structural, and temporal characteristics of the threaded conversation by an observable time for the final prediction. We evaluate the proposed model with two different types of the threaded conversations, self-introductory Q&A and discussion, which demonstrates that the proposed model can accurately predict the continuity of the conversations regardless of types of threaded conversations. We believe that the proposed methodology and the results can provide the potential insights to advertisers, opinion leaders, or platform designers who want to understand and predict the evolutionary process of threaded conversations. © 2022, The Authors. All rights reserved.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

在天文学中基于机器学习的流体力学问题加速求解

在天文学中基于机器学习的流体力学问题加速求解

引用

第十三届全国流体力学学术会议

作者：毛顺元董若冰王炜琦 Kwang Moo Yi 陆路王思凡 Paris Perdikaris Department of Physics and Astronomy University of Victoria Department of Computer Science University of British Columbia Department of Statistics and Data Science Yale University Graduate Group in Applied Mathematics and Computational Science University of Pennsylvania Department ofMechanical Engineering and Applied Mechanics University of Pennsylvania

探索行星形成的过程不仅揭示了地球及其生命的起源,还为寻找地外生命提供了关键线索。行星在原行星盘中形成。原行星盘是由气体和尘埃构成的扁平盘状系统,围绕在新生恒星的周围,其跨度可达数千亿公里。近年来,通过分析行星对原行星盘的... 详细信息

探索行星形成的过程不仅揭示了地球及其生命的起源,还为寻找地外生命提供了关键线索。行星在原行星盘中形成。原行星盘是由气体和尘埃构成的扁平盘状系统,围绕在新生恒星的周围,其跨度可达数千亿公里。近年来,通过分析行星对原行星盘的引力扰动,并比较观测和数值模拟,天文学家得以推断正在盘中形成的行星的质量和轨道。然而,这种方法依赖于复杂的流体力学数值模拟,因计算成本高昂而难以普及。为了克服这一挑战,我们开发了一款基于机器学习的工具。该工具应用数据驱动的深度算子网络(DeepONet),将系统的参数映射到流体力学偏微分方程的解,从而无需昂贵的数值模拟即可实现两项功能:1)快速求解正问题,即从行星参数预测盘中物质分布和速度场;2)高效求解反问题,即从观测到的扰动图像中推断行星参数。我们展示了该工具在求解速度、精确性和鲁棒性方面的显著优势,并讨论了其在理论研究和观测中的潜在应用。我们还探索了物理信息神经网络(PINN)在这一问题的应用,特别是针对原初行星盘中旋转吸积流的时间依赖性二维可压缩Navier-Stokes方程。行星引力扰动引入显著的时间依赖性空间特征,使问题复杂化。通过对物理定律而非标注数据的训练,PINN在捕捉系统演化的空间高频特征方面展现潜力。标准PINN在此任务中存在收敛困难、预测误差随时间累积、长时间解偏离物理定律以及捕捉空间高频特征能力不足的问题。为此,我们的研究引入了自适应激活函数、时间推进策略和网络权重继承方法,以提升PINN性能。改进后的PINN模型在与标准PINN的比较中表现出显著更低的预测误差,成功再现了大部分空间高频特征,无需标注数据。这项研究为将PINN应用于类似的时间依赖性流体力学问题提供了重要参考。

关键词：天文行星形成深度学习深度算子网络物理信息神经网络

来源：评论

学校读者我要写书评

暂无评论

Thermal comparison of mono-facial and bi-facial photovoltaic cells considering the effect of TPT layer absorptivity

引用

International Journal of Thermofluids 2023年 18卷

作者： Radwan, Ali Mahmoud, Montaser Olabi, Abdul-Ghani Rezk, Ahmed Maghrabie, Hussein M Abdelkareem, Mohammad Ali Sustainable Energy & Power Systems Research Centre RISE University of Sharjah Sharjah P.O. Box 27272 United Arab Emirates Mechanical Power Engineering Department Faculty of Engineering Mansoura University Mansoura35516 Egypt Department of Engineering School of Mathematics Computer Science and Engineering City University of London London United Kingdom Mechanical Engineering and Design Aston University School of Engineering and Applied Science Birmingham Aston TriangleB4 7ET United Kingdom College of Engineering and Physical Science Aston University Birmingham B4 7ET United Kingdom Department of Mechanical Engineering Faculty of Engineering South Valley University Qena83521 Egypt Chemical Engineering Department Minia University Elminia Egypt

With the widespread utilization of solar photovoltaics (PV), it is becoming increasingly important to understand its performance using various configurations to harvest solar energy at the most suitable efficiency, specifically in hot climates. Therefore, this paper compares mono-facial and bi-facial PV cells under the high-temperature desert climate of Sharjah (United Arab Emirates). The optimally tilted and vertical east-west configurations were compared to evaluate the annual energy production for these scenarios. In addition, new parameter, the effect of the rear side layer absorptivity of the mono-facial PV cell, was investigated. The comparative study considered solar radiation, ambient temperature, wind speed, and albedo. The results showed that considering the effect of absorptivity achieves better performance predictions. However, the effect of the absorptivity did not exhibit a significant influence compared to the case with neglected backsheet absorptivity. As such, the PV cell temperature increased by 0.4 °C and 1.4 °C at an albedo of 0.2 and 0.7, considering the actual absorptivity value of 0.2. On the other hand, it is worth mentioning that improper designs result in considerable temperature rises as the absorptivity value increases. Under the investigated conditions and based on the yearlong comparison, the tilted bi-facial performed better than the vertical bi-facial;as the albedo increased from 0.2 to 0.7, the power gain ratio of the tilted bi-facial increased from 7.18% to 20.88% and that of vertical bi-facial from -19.4% to 12.65%. © 2023 The Author(s)

关键词： Solar energy

来源：评论

学校读者我要写书评

暂无评论

Radiogenomic classification for MGMT promoter methylation status using multi-omics fused feature space for least invasive diagnosis through mpMRI scans (vol 13, 3291, 2023)

引用

SCIENTIFIC REPORTS 2023年第1期13卷 1-29页

作者： Qureshi, Shahzad Ahmad Hussain, Lal Ibrar, Usama Alabdulkreem, Eatedal Nour, Mohamed K. Alqahtani, Mohammed S. Nafie, Faisal Mohammed Mohamed, Abdullah Mohammed, Gouse Pasha Duong, Tim Q. Department of Computer and Information Sciences Pakistan Institute of Engineering and Applied Sciences Islamabad Pakistan Department of Computer Science and IT Neelum Campus The University of Azad Jammu and Kashmir Muzaffarabad Azad Kashmir Pakistan Department of Computer Science and IT King Abdullah Campus The University of Azad Jammu and Kashmir Muzaffarabad Azad Kashmir Pakistan Department of Radiology Albert Einstein College of Medicine and Montefiore Medical Center 111 East 210th Street Bronx NY 10467 USA Farooq Hospital Lahore Pakistan Department of Computer Sciences College of Computer and Information Sciences Princess Nourah Bint Abdulrahman University P.O. Box 84428 Riyadh 11671 Saudi Arabia Department of Computer Sciences College of Computing and Information System Umm Al-Qura University Mecca Saudi Arabia Radiological Sciences Department College of Applied Medical Sciences King Khalid University Abha 61421 Saudi Arabia Department of Computer Science College of Science and Humanities at Alghat Majmaah University Al-Majmaah 11952 Saudi Arabia Research Centre Future University in Egypt New Cairo 11845 Egypt Department of Computer and Self Development Preparatory Year Deanship Prince Sattam Bin Abdulaziz University AlKharj Saudi Arabia

Accurate radiogenomic classification of brain tumors is important to improve the standard of diagnosis, prognosis, and treatment planning for patients with glioblastoma. In this study, we propose a novel two-stage MGMT Promoter Methylation Prediction (MGMT-PMP) system that extracts latent features fused with radiomic features predicting the genetic subtype of glioblastoma. A novel fine-tuned deep learning architecture, namely Deep Learning Radiomic Feature Extraction (DLRFE) module, is proposed for latent feature extraction that fuses the quantitative knowledge to the spatial distribution and the size of tumorous structure through radiomic features: (GLCM, HOG, and LBP). The application of the novice rejection algorithm has been found significantly effective in selecting and isolating the negative training instances out of the original dataset. The fused feature vectors are then used for training and testing byk-NN and SVM classifiers. The 2021 RSNA Brain Tumor challenge dataset (BraTS-2021) consists of four structural mpMRIs, viz. fluid-attenuated inversion-recovery, T1-weighted, T1-weighted contrast enhancement, and T2-weighted. We evaluated the classification performance, for the very first time in published form, in terms of measures like accuracy, F1-score, and Matthews correlation coefficient. The Jackknife tenfold cross-validation was used for training and testing BraTS-2021 dataset validation. The highest classification performance is (96.84 ± 0.09)%, (96.08 ± 0.10)%, and (97.44 ± 0.14)% as accuracy, sensitivity, and specificity respectively to detect MGMT methylation status for patients suffering from glioblastoma. Deep learning feature extraction with radiogenomic features, fusing imaging phenotypes and molecular structure, using rejection algorithm has been found to perform outclass capable of detecting MGMT methylation status of glioblastoma patients. The approach relates the genomic variation with radiomic features forming a bridge between two areas of

关键词：

来源：评论

学校读者我要写书评

暂无评论

MIND-WANDERING DETECTION MODEL WITH ELECTROENCEPHALOGRAM 18

MIND-WANDERING DETECTION MODEL WITH ELECTROENCEPHALOGRAM

引用

18th International Conference on Cognition and Exploratory Learning in Digital Age, CELDA 2021

作者： Rungsilp, Chutimon Piromsopa, Krerk Viriyopase, Atthaphon U-Yen, Kongpop Department of Computer Engineering Faculty of Engineering Chulalongkorn University Bangkok Thailand Research Group on Applied Digital Technology in Medicine Chulalongkorn University Bangkok Thailand Future Innovation Science and Technology Chulalongkorn University Bangkok Thailand

ISBN: (纸本)9789898704337

The study of mind-wandering is popular since it is linked to the emotional problems and working/learning performance. In terms of education, it impacts comprehension during learning which affects academic success. Therefore, we sought to develop a machine learning model for an embedded portable device that can categorize mind-wandering state to assist people in keeping track of their minds. We utilize a low-channel EEG to record the brain state and to build the predictive model because of its practicality and user-friendly. Most machine learning experiments in mind-wandering using EEG exhibit good individual-level performance. For the group-level technique, only a few research has developed a model. As a result, the goal of this research is to achieve a high-accuracy group-level model. Thus, Leave One Participant Out Cross Validation (LOPOCV) was used to assess the model correctness. This study shows that using a baseline normalization technique assists feature extraction and improves performance. The model was built using a support vector machine (SVM), and the best model achieved an accuracy value of 75.6 percent. © 2021 Virtual Simulation Innovation Workshop, SIW 2021. All rights reserved.

关键词： Electroencephalography

来源：评论

学校读者我要写书评

暂无评论

Enhancing Production Planning in the Internet of Vehicles: A Transformer-based Federated Reinforcement Learning Approach

Enhancing Production Planning in the Internet of Vehicles: A...

引用

IEEE Conference on Vehicular Technology (VTC)

作者： Jinhua Chen Zihan Zhao Keping Yu Shahid Mumtaz Joel J. P. C. Rodrigues Mohsen Guizani Takuro Sato Graduate School of Science and Engineering Hosei University Tokyo Japan Department of Applied Informatics Silesian University of Technology Gliwice Poland Departement of Computer Sciences Nottingham Trent University Nottingham UK Amazonas State University Manaus AM Brazil Machine Learning Department Mohamed Bin Zayed University of Artificial Intelligence (MBZUAI) UAE Research Institute for Science and Engineering Waseda University Tokyo Japan

ISBN: (数字)9798350387414

ISBN: (纸本)9798350387421

The Internet of Vehicles (10V) brings significant economic benefits to countries. However, large-scale smart vehicle production planning remains challenging in the 10V. Currently, heuristic algorithms and solvers commonly used for these problems often lack scalability and fall into local optima. Moreover, security concerns about wireless data transfer arising from multi-factory manufacturing processes are garnering attention. To address these issues, this paper introduces an algorithm, TRL, which is a Transformer-based Reinforcement Learning for vehicle production planning problems. Furthermore, we propose a Transformer-based Federated Reinforcement Learning algorithm, named TFRL, tailored for large-scale manufacturing and secure wireless communication. Experimental results showcase the high performance and security of TFRL. It schedules 1000 orders in about 14 seconds and avoids exchanging plaintext during the interaction. Compared to Non-dominated Sorting Genetic Algorithm II(NSGA-II), the TFRL enhances computational speed by 95.12% and reduces constraint violation scores by 93.18%.

关键词： Wireless communication Vehicular and wireless technologies Schedules Production planning Reinforcement learning Transformers Manufacturing

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：