检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Breden, Maxime Chu, Hugo Lamb, Jeroen S.W. Rasmussen, Martin CMAP CNRS École polytechnique Institut Polytechnique de Paris Palaiseau91120 France Department of Mathematics Imperial College London LondonSW7 2AZ United Kingdom International Research Center for Neurointelligence The University of Tokyo Tokyo113-0033 Japan Centre for Applied Mathematics and Bioinformatics Department of Mathematics and Natural Sciences Gulf University for Science and Technology Halwally 32093 Kuwait

We develop a powerful and general method to provide arbitrarily accurate rigorous upper and lower bounds for Lyapunov exponents of stochastic flows. Our approach is based on computer-assisted tools, the adjoint method and established results on the ergodicity of diffusion processes. We do not require any structural assumptions on the stochastic system and work under mild hypoellipticity conditions outside of perturbative regimes. Therefore, our method allows for the treatment of systems that were so far inaccessible from existing mathematical tools. We demonstrate our method to exhibit the chaotic nature of three non-Hamiltonian systems. Finally, we show that our approach is robust to continuation methods to produce bounds on Lyapunov exponents for large parameter *** Codes 37M25, 37H15, 65P30, 60J22, 65G20 © 2024, CC BY.

关键词： Lyapunov functions

来源：评论

学校读者我要写书评

暂无评论

Graph Neural Network Interatomic Potential Ensembles with Calibrated Aleatoric and Epistemic Uncertainty on Energy and Forces

arXiv

引用

arXiv 2023年

作者： Busk, Jonas Schmidt, Mikkel N. Winther, Ole Vegge, Tejs Jørgensen, Peter Bjørn Department of Energy Conversion and Storage Technical University of Denmark Kongens Lyngby Denmark Department of Applied Mathematics and Computer Science Technical University of Denmark Kongens Lyngby Denmark Center for Genomic Medicine Rigshospitalet Copenhagen University Hospital Denmark Bioinformatics Centre Department of Biology University of Copenhagen Denmark

Inexpensive machine learning potentials are increasingly being used to speed up structural optimization and molecular dynamics simulations of materials by iteratively predicting and applying interatomic forces. In these settings, it is crucial to detect when predictions are unreliable to avoid wrong or misleading results. Here, we present a complete framework for training and recalibrating graph neural network ensemble models to produce accurate predictions of energy and forces with calibrated uncertainty estimates. The proposed method considers both epistemic and aleatoric uncertainty and the total uncertainties are recalibrated post hoc using a nonlinear scaling function to achieve good calibration on previously unseen data, without loss of predictive accuracy. The method is demonstrated and evaluated on two challenging, publicly available datasets, ANI-1x (Smith et al. 1 ) and Transition1x (Schreiner et al. 2 ), both containing diverse conformations far from equilibrium. A detailed analysis of the predictive performance and uncertainty calibration is provided. In all experiments, the proposed method achieved low prediction error and good uncertainty calibration, with predicted uncertainty correlating with expected error, on energy and forces. To the best of our knowledge, the method presented in this paper is the first to consider a complete framework for obtaining calibrated epistemic and aleatoric uncertainty predictions on both energy and forces in ML potentials. © 2023, CC BY.

关键词： Calibration

来源：评论

学校读者我要写书评

暂无评论

Calibrated uncertainty for molecular property prediction using ensembles of message passing neural networks

arXiv

引用

arXiv 2021年

作者： Busk, Jonas Jørgensen, Peter Bjørn Bhowmik, Arghya Schmidt, Mikkel N. Winther, Ole Vegge, Tejs Department for Energy Conversion and Storage Technical University of Denmark Lyngby Denmark Department of Applied Mathematics and Computer Science Technical University of Denmark Lyngby Denmark Center for Genomic Medicine Rigshospitalet Copenhagen University Hospital Copenhagen Denmark Bioinformatics Centre Department of Biology University of Copenhagen Copenhagen Denmark

Data-driven methods based on machine learning have the potential to accelerate computational analysis of atomic structures. In this context, reliable uncertainty estimates are important for assessing confidence in predictions and enabling decision making. However, machine learning models can produce badly calibrated uncertainty estimates and it is therefore crucial to detect and handle uncertainty carefully. In this work we extend a message passing neural network designed specifically for predicting properties of molecules and materials with a calibrated probabilistic predictive distribution. The method presented in this paper differs from previous work by considering both aleatoric and epistemic uncertainty in a unified framework, and by recalibrating the predictive distribution on unseen data. Through computer experiments, we show that our approach results in accurate models for predicting molecular formation energies with well calibrated uncertainty in and out of the training data distribution on two public molecular benchmark datasets, QM9 and PC9. The proposed method provides a general framework for training and evaluating neural network ensemble models that are able to produce accurate predictions of properties of molecules with well calibrated uncertainty estimates. © 2021, CC BY.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Design Considerations for Hybrid Underwater Target Tracking: Integrating Particle Filters and Deep Learning for Enhanced Decision

Design Considerations for Hybrid Underwater Target Tracking:...

引用

Decision Aid Sciences and Application (DASA), International Conference on

作者： Abdullatif Baba Basil Alothman Maha Shabon Ziad Salem Computer Science and Engineering Department Kuwait College of Science and Technology (KCST) Kuwait Computer Engineering Department University of Turkish Aeronautical Association (UTAA) Ankara Türkiye Computer Science Department Center for Applied Mathematics and Bioinformatics (CAMB) Gulf University for Science and Technology (GUST) Hawally Kuwait

ISBN: (数字)9798350369106

ISBN: (纸本)9798350369113

This paper delves into the key characteristics of Autonomous underwater vehicle (AUV) design, highlighting considerations such as hull structure, hydrodynamics, propulsion systems, and sensor integration. The main contribution of this paper is the tracking of underwater dynamic targets, distinguishing between reactive and deliberative tracking levels by proposing a novel hybrid tracking approach, the Multiple Model Particle Filter (MMPF), that integrates deep learning to predict dynamic target trajectory. Finally, a few simulations are delivered to demonstrate the efficiency of the proposed approach in tracking underwater maneuvering targets under different trajectory scenarios.

关键词： Deep learning Autonomous underwater vehicles Target tracking Simultaneous localization and mapping Navigation Particle filters Trajectory Safety Security Vehicle dynamics

来源：评论

学校读者我要写书评

暂无评论

Particle Swarm Optimization-Based Deep Neural Network vs Whale Optimization Algorithm-Based Deep Convolutional Neural Networks for Critical Heat Flux Prediction

Particle Swarm Optimization-Based Deep Neural Network vs Wha...

引用

Decision Aid Sciences and Application (DASA), International Conference on

作者： Messaoud Djeddou Jehad Al Dallal Aouatef Hellal Ibrahim A. Hameed Xingang Zhao Hydraulic Department Oum El-Bouaghi University Ain Beida Algeria Center for Applied Mathematics and Bioinformatics Gulf University for Science and Technology Kuwait Department of ICT and Natural Sciences Norwegian University of Science and Technology (NTNU) Alesund Norway Department of Nuclear Engineering University of Tennessee Knoxville Knoxville USA

ISBN: (数字)9798350369106

ISBN: (纸本)9798350369113

Critical heat flux (CHF) denotes the thermal limit at which a phase change occurs during heating, that causes a significant decline in heat transfer efficiency and eventual localized overheating of the heating surface. This study applies two hybrid deep learning models: a Deep Convolutional Neural Network (CNN) optimized with the Whale Optimization Algorithm and a Deep Neural Network (DNN) optimized with Particle Swarm Optimization, to predict Critical Heat Flux (CHF). The results show that the latter model outperforms the former in terms of performance, simplicity, and computation efficiency in the CHF prediction task.

关键词： Heating systems Deep learning Training Computational modeling Predictive models Prediction algorithms Whale optimization algorithms Convolutional neural networks Particle swarm optimization Heat transfer

来源：评论

学校读者我要写书评

暂无评论

EXISTENCE AND UNIQUENESS OF QUASI-STATIONARY AND QUASI-ERGODIC MEASURES FOR ABSORBING MARKOV CHAINS: A BANACH LATTICE APPROACH

arXiv

引用

arXiv 2021年

作者： Castro, Matheus M. Lamb, Jeroen S.W. Olicón-Mendez, Guillermo Rasmussen, Martin Department of Mathematics Imperial College London LondonSW7 2AZ United Kingdom International Research Center for Neurointelligence The University of Tokyo Tokyo113-0033 Japan Centre for Applied Mathematics and Bioinformatics Department of Mathematics and Natural Sciences Gulf University for Science and Technology Hawally32093 Kuwait Institut für Mathematik Freie Universität Berlin Arnimallee 7 Berlin14195 Germany

We establish the existence and uniqueness of quasi-stationary and quasi-ergodic measures for almost surely absorbed discrete-time Markov chains under weak conditions. We obtain our results by exploiting Banach lattice... 详细信息

关键词： Markov processes

来源：评论

学校读者我要写书评

暂无评论

Hybrid Deep Convolutional Neural Networks Combined with Autoencoders and Augmented Data to Predict the Look-Up Table 2006

Hybrid Deep Convolutional Neural Networks Combined with Auto...

引用

Decision Aid Sciences and Application (DASA), International Conference on

作者： Messaoud Djeddou Jehad Al Dallal Aouatef Hellal Ibrahim A. Hameed Xingang Zhao Hydraulic Department Oum El-Bouaghi University Ain Beida Algeria Center for Applied Mathematics and Bioinformatics Gulf University for Science and Technology Kuwait Department of ICT and Natural Sciences Norwegian University of Science and Technology (NTNU) Ålesund Norway Department of Nuclear Engineering University of Tennessee Knoxville Knoxville USA

ISBN: (数字)9798350369106

ISBN: (纸本)9798350369113

Lookup tables derived from empirical research contain limited data for critical heat flux (CHF). Machine learning techniques can be employed to create CHF prediction models by training them with existing experimental-based CHF lookup tables. Consequently, these models can generate CHF look-up tables that provide expected CHF values for a wider variety of scenarios. This study presents a novel predictive method integrating auto-encoders with deep convolutional neural networks (DCNN). The purpose is to obtain precise and reliable predictions for CHF retrieval. The accuracy of these models was evaluated utilizing the coefficient of determination (R 2 ), Nash-Sutcliffe efficiency (NSE), mean absolute error (MAE), and normalized root-mean-squared error (NRMSE). The findings demonstrate that the suggested hybrid model reveals an important degree of accuracy.

关键词： Training Heating systems Analytical models Accuracy Prevention and mitigation Autoencoders Predictive models Table lookup Convolutional neural networks Reliability

来源：评论

学校读者我要写书评

暂无评论

Single-Cell Entropy to Quantify the Cellular Order Parameter from Single-Cell RNA-Seq Data

引用

Biophysical Reviews and Letters 2020年第1期15卷 35-49页

作者： Jingxin Liu You Song Jinzhi Lei School of Software Beihang University Beijing 100191 P. R. China http://***/0000-0002-2670-6760 Zhou Pei-Yuan Center for Applied Mathematics MOE Key Laboratory of Bioinformatics Tsinghua University Beijing 100084 P. R. China Current address: School of Mathematical Science Tiangong University Tianjin 300387 China. Email:

The cell is the basic functional and biological unit of life, and a complex system that contains a huge number of molecular components. How can we quantify the macroscopic state of a cell from the microscopic information of these molecular components? This is a fundamental question to increase the understanding of the human body. The recent maturation of single-cell RNA sequencing (scRNA-seq) technologies has allowed researchers to gain information on the transcriptomes of individual cells. Although considerable progress has been made in terms of cell-type clustering over the past few years, there is no strong consensus about how to define a cell state from scRNA-seq data. Here, we present single-cell entropy (scEntropy) as an order parameter for cellular transcriptome profiles from scRNA-seq data. scEntropy is a straightforward parameter with which to define the intrinsic transcriptional state of a cell that can provide a quantity to measure the developmental process and to distinguish different cell types. The proposed scEntropy followed by Gaussian mixture model (scEGMM) provides a coherent method of cell-type classification that is simple, includes no parameters or clustering and is comparable to existing machine learning-based methods in benchmarking studies. The results of cell-type classification based on scEGMM are robust and easy to biologically interpret.

关键词： Single-cell RNA sequencing entropy order parameter cell-type classification

来源：评论

学校读者我要写书评

暂无评论

Blood collection tube and RNA purification method recommendations for extracellular RNA transcriptome profiling

引用

Nature Communications 2025年第1期16卷 1-17页

作者： Jasper Anckaert Francisco Avila Cobos Anneleen Decock Philippe Decruyenaere Jill Deleu Jilke De Wilde Thibaut D’huyvetter Celine Everaert Hetty H. Helsmoortel Eva Hulstaert Pieter Mestdagh Annelien Morlion Justine Nuytens Franco Poma-Soto Kathleen Schoofs Eveline Vanden Eynde Jo Vandesompele Tom Van Maerken Ruben Van Paemel Kimberly Verniers Jasper Verwilt Nurten Yigit Katleen De Preter Olivier De Wever Bert Dhondt An Hendrix Annouck Philippron Carolina Fierro Nele Nijs Thomas Piofczyk Jan Koster Scott Kuersten Gary P. Schroth Tim R. Mercer Olivier Thas OncoRNALab Department of Biomolecular Medicine Ghent University Ghent Belgium Cancer Research Institute Ghent (CRIG) Ghent Belgium TOBI lab Center for Medical Genetics Department of Biomolecular Medicine Ghent University Ghent Belgium Department of Hematology Ghent University Hospital Ghent Belgium Department of Dermatology Ghent University Hospital Ghent Belgium Biogazelle Zwijnaarde Belgium Department of Laboratory Medicine AZ Groeninge Kortrijk Belgium Department of Pediatrics Ghent University Hospital Ghent Belgium Laboratory of Experimental Cancer Research Ghent University Ghent Belgium Department of Urology Ghent University Hospital Ghent Belgium Lab for Experimental Surgery Ghent University Ghent Belgium Department of Gastrointestinal Surgery Ghent University Hospital Ghent Belgium Center for Experimental and Molecular Medicine (CEMM) Amsterdam University Medical Centers University of Amsterdam Amsterdam The Netherlands Illumina San Diego California USA Australian Institute of Bioengineering and Nanotechnology University of Queensland Brisbane QLD Australia Kinghorn Centre for Clinical Genomics Garvan Institute of Medical Research Sydney NSW Australia Data Science Institute (DSI) Interuniversity Institute for Biostatistics and Statistical Bioinformatics (I-BioStat) Hasselt University Hasselt Belgium Department of Applied Mathematics Computer Science and Statistics Ghent University Gent Belgium National Institute of Applied Statistics Research Australia (NIASRA) University of Wollongong Wollongong Australia

Blood-based extracellular RNA (cell-free RNA; exRNA) biomarkers require validated sample collection, processing, and quantification procedures. No study to date has systematically tested pre-analytical variables affecting transcriptome-wide exRNA analysis. By evaluating their impact on deep transcriptome profiling of microRNAs and mRNAs in blood plasma or serum, we compared ten blood collection tubes, three blood processing time intervals, and eight RNA purification methods. In addition, we assessed interactions among a selected pre-analytical variable set, resulting in 456 extracellular transcriptomes. Blood preservation tubes failed to stabilize exRNA and RNA purification methods differed significantly in performance, causing variations in concentration, detected gene numbers, replicability and observed transcriptome complexity. Critical interactions between tubes, purification methods and time intervals were identified. We provide 11 analytical performance metrics for exRNA quantification methods and put forward recommendations for both users and manufacturers of RNA purification methods and blood collection tubes, collectively, essential groundwork for exRNA-based precision medicine applications.

关键词：

来源：评论

学校读者我要写书评

暂无评论

The neglect of nonnative orthopterans as potential invaders:A call for awareness

引用

Insect Science 2024年第4期31卷 994-1000页

作者： Anna K.Kulessa Paride Balzani Ismael Soto Antonín Kouba David Renault Ali Serhan Tarkan Phillip J.Haubrock Ruhr University Bochum Faculty of Biology and BiotechnologyBochumGermany Faculty of Biology University of Duisburg-EssenUniversitätsstraße 5EssenGermany Faculty of Fisheries and Protection of Waters South Bohemian Research Center of Aquaculture and Biodiversity of HydrocenosesUniversity of South Bohemia in CˇeskéBudeˇjoviceVodnˇanyCzech Republic University of Rennes CNRSECOBIO(EcosystèmesBiodiversitéEvolution)UMRRennesFrance Faculty of Fisheries Department of Basic SciencesMu˘gla SıtkıKoçman UniversityMu˘glaTurkey Department of Life and Environmental Sciences Bournemouth UniversityPooleDorsetUK Faculty of Biology and Environmental Protection Department of Ecology and Vertebrate ZoologyUniversity ofŁodzŁodzPoland Department of River Ecology and Conservation Senckenberg Research Institute and Natural History Museum FrankfurtGelnhausenGermany Center for Applied Mathematics and Bioinformatics Gulf University for Science and TechnologyMubarak Al-AbdullahKuwait

Despite the potential ecological and economic impacts of invasive species,there is a dearth of data on the presence,impacts,and management implications of potentially invasive Orthoptera *** lack of research and inconsistent data,including risk screenings and impact assessments,is especially evident in ***,assessing the status,distribution,and potential threats of nonnative Orthoptera in Europe remains challenging,impeding the development of effective management *** address this gap,we call for increased efforts to collect and curate data on non-native and possibly invasive Orthoptera in *** efforts will improve our understanding of this order’s invasion dynamics,facilitate the identification of priority areas for conservation,and support the development of effective management policies and preventive measures.

关键词： costs cricket damages grasshopper impact research

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：