arXiv - CS - Computational Engineering, Finance, and Science最新文献_第6页

Single Atom Convolutional Matching Pursuit: Theoretical Framework and Application to Lamb Waves based Structural Health Monitoring 单原子卷积匹配追求：基于λ波的结构健康监测的理论框架与应用

arXiv - CS - Computational Engineering, Finance, and Science Pub Date : 2024-08-16 DOI: arxiv-2408.08929

Sebastian Rodriguez, Marc Rébillat, Shweta Paunikar, Pierre Margerit, Eric Monteiro, Francisco Chinesta, Nazih Mechbal

{"title":"Single Atom Convolutional Matching Pursuit: Theoretical Framework and Application to Lamb Waves based Structural Health Monitoring","authors":"Sebastian Rodriguez, Marc Rébillat, Shweta Paunikar, Pierre Margerit, Eric Monteiro, Francisco Chinesta, Nazih Mechbal","doi":"arxiv-2408.08929","DOIUrl":"https://doi.org/arxiv-2408.08929","url":null,"abstract":"Structural Health Monitoring (SHM) aims to monitor in real time the health\u0000state of engineering structures. For thin structures, Lamb Waves (LW) are very\u0000efficient for SHM purposes. A bonded piezoelectric transducer (PZT) emits LW in\u0000the structure in the form of a short tone burst. This initial wave packet (IWP)\u0000propagates in the structure and interacts with its boundaries and\u0000discontinuities and with eventual damages generating additional wave packets.\u0000The main issues with LW based SHM are that at least two LW modes are\u0000simultaneously excited and that those modes are dispersive. Matching Pursuit\u0000Method (MPM), which consists of approximating a signal as a sum of different\u0000delayed and scaled atoms taken from an a priori known learning dictionary,\u0000seems very appealing in such a context, however is limited to nondispersive\u0000signals and relies on a priori known dictionary. An improved version of MPM\u0000called the Single Atom Convolutional Matching Pursuit method (SACMPM), which\u0000addresses the dispersion phenomena by decomposing a measured signal as delayed\u0000and dispersed atoms and limits the learning dictionary to only one atom, is\u0000proposed here. Its performances are illustrated when dealing with numerical and\u0000experimental signals as well as its usage for damage detection. Although the\u0000signal approximation method proposed in this paper finds an original\u0000application in the context of SHM, this method remains completely general and\u0000can be easily applied to any signal processing problem.","PeriodicalId":501309,"journal":{"name":"arXiv - CS - Computational Engineering, Finance, and Science","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142211289","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Assessing and Enhancing Large Language Models in Rare Disease Question-answering 评估和增强罕见病问答中的大型语言模型

arXiv - CS - Computational Engineering, Finance, and Science Pub Date : 2024-08-15 DOI: arxiv-2408.08422

Guanchu Wang, Junhao Ran, Ruixiang Tang, Chia-Yuan Chang, Chia-Yuan Chang, Yu-Neng Chuang, Zirui Liu, Vladimir Braverman, Zhandong Liu, Xia Hu

{"title":"Assessing and Enhancing Large Language Models in Rare Disease Question-answering","authors":"Guanchu Wang, Junhao Ran, Ruixiang Tang, Chia-Yuan Chang, Chia-Yuan Chang, Yu-Neng Chuang, Zirui Liu, Vladimir Braverman, Zhandong Liu, Xia Hu","doi":"arxiv-2408.08422","DOIUrl":"https://doi.org/arxiv-2408.08422","url":null,"abstract":"Despite the impressive capabilities of Large Language Models (LLMs) in\u0000general medical domains, questions remain about their performance in diagnosing\u0000rare diseases. To answer this question, we aim to assess the diagnostic\u0000performance of LLMs in rare diseases, and explore methods to enhance their\u0000effectiveness in this area. In this work, we introduce a rare disease\u0000question-answering (ReDis-QA) dataset to evaluate the performance of LLMs in\u0000diagnosing rare diseases. Specifically, we collected 1360 high-quality\u0000question-answer pairs within the ReDis-QA dataset, covering 205 rare diseases.\u0000Additionally, we annotated meta-data for each question, facilitating the\u0000extraction of subsets specific to any given disease and its property. Based on\u0000the ReDis-QA dataset, we benchmarked several open-source LLMs, revealing that\u0000diagnosing rare diseases remains a significant challenge for these models. To facilitate retrieval augmentation generation for rare disease diagnosis,\u0000we collect the first rare diseases corpus (ReCOP), sourced from the National\u0000Organization for Rare Disorders (NORD) database. Specifically, we split the\u0000report of each rare disease into multiple chunks, each representing a different\u0000property of the disease, including their overview, symptoms, causes, effects,\u0000related disorders, diagnosis, and standard therapies. This structure ensures\u0000that the information within each chunk aligns consistently with a question.\u0000Experiment results demonstrate that ReCOP can effectively improve the accuracy\u0000of LLMs on the ReDis-QA dataset by an average of 8%. Moreover, it significantly\u0000guides LLMs to generate trustworthy answers and explanations that can be traced\u0000back to existing literature.","PeriodicalId":501309,"journal":{"name":"arXiv - CS - Computational Engineering, Finance, and Science","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142211291","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Crystalline Material Discovery in the Era of Artificial Intelligence 人工智能时代的晶体材料发现

arXiv - CS - Computational Engineering, Finance, and Science Pub Date : 2024-08-15 DOI: arxiv-2408.08044

Zhenzhong Wang, Haowei Hua, Wanyu Lin, Ming Yang, Kay Chen Tan

{"title":"Crystalline Material Discovery in the Era of Artificial Intelligence","authors":"Zhenzhong Wang, Haowei Hua, Wanyu Lin, Ming Yang, Kay Chen Tan","doi":"arxiv-2408.08044","DOIUrl":"https://doi.org/arxiv-2408.08044","url":null,"abstract":"Crystalline materials, with their symmetrical and periodic structures,\u0000possess a diverse array of properties and have been widely used in various\u0000fields, e.g., sustainable development. To discover crystalline materials,\u0000traditional experimental and computational approaches are often time-consuming\u0000and expensive. In these years, thanks to the explosive amount of crystalline\u0000materials data, great interest has been given to data-driven materials\u0000discovery. Particularly, recent advancements have exploited the expressive\u0000representation ability of deep learning to model the highly complex atomic\u0000systems within crystalline materials, opening up new avenues for fast and\u0000accurate materials discovery. These works typically focus on four types of\u0000tasks, including physicochemical property prediction, crystalline material\u0000synthesis, aiding characterization, and force field development; these tasks\u0000are essential for scientific research and development in crystalline materials\u0000science. Despite the remarkable progress, there is still a lack of systematic\u0000research to summarize their correlations, distinctions, and limitations. To\u0000fill this gap, we systematically investigated the progress made in deep\u0000learning-based material discovery in recent years. We first introduce several\u0000data representations of the crystalline materials. Based on the\u0000representations, we summarize various fundamental deep learning models and\u0000their tailored usages in material discovery tasks. We also point out the\u0000remaining challenges and propose several future directions. The main goal of\u0000this review is to offer comprehensive and valuable insights and foster progress\u0000in the intersection of artificial intelligence and material science.","PeriodicalId":501309,"journal":{"name":"arXiv - CS - Computational Engineering, Finance, and Science","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142211292","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Efficient low rank model order reduction of vibroacoustic problems under stochastic loads 随机载荷下振动声学问题的高效低阶模型阶次缩减

arXiv - CS - Computational Engineering, Finance, and Science Pub Date : 2024-08-15 DOI: arxiv-2408.08402

Yannik Hüpel, Ulrich Römer, Matthias Bollhöfer, Sabine Langer

{"title":"Efficient low rank model order reduction of vibroacoustic problems under stochastic loads","authors":"Yannik Hüpel, Ulrich Römer, Matthias Bollhöfer, Sabine Langer","doi":"arxiv-2408.08402","DOIUrl":"https://doi.org/arxiv-2408.08402","url":null,"abstract":"This contribution combines a low-rank matrix approximation through Singular\u0000Value Decomposition (SVD) with second-order Krylov subspace-based Model Order\u0000Reduction (MOR), in order to efficiently propagate input uncertainties through\u0000a given vibroacoustic model. The vibroacoustic model consists of a plate\u0000coupled to a fluid into which the plate radiates sound due to a turbulent\u0000boundary layer excitation. This excitation is subject to uncertainties due to\u0000the stochastic nature of the turbulence and the computational cost of\u0000simulating the coupled problem with stochastic forcing is very high. The\u0000proposed method approximates the output uncertainties in an efficient way, by\u0000reducing the evaluation cost of the model in terms of DOFs and samples by using\u0000the factors of the SVD low-rank approximation directly as input for the MOR\u0000algorithm. Here, the covariance matrix of the vector of unknowns can\u0000efficiently be approximated with only a fraction of the original number of\u0000evaluations. Therefore, the approach is a promising step to further reducing\u0000the computational effort of large-scale vibroacoustic evaluations.","PeriodicalId":501309,"journal":{"name":"arXiv - CS - Computational Engineering, Finance, and Science","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142211297","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Multilayer Network of Cardiovascular Diseases and Depression via Multipartite Projection 通过多方投影研究心血管疾病和抑郁症的多层网络

arXiv - CS - Computational Engineering, Finance, and Science Pub Date : 2024-08-14 DOI: arxiv-2408.07562

Jie Li, Cillian Hourican, Pashupati P. Mishra, Binisha H. Mishra, Mika Kähönen, Olli T. Raitakari, Reijo Laaksonen, Mika Ala-Korpela, Liisa Keltikangas-Järvinen, Markus Juonala, Terho Lehtimäki, Jos A. Bosch, Rick Quax

{"title":"Multilayer Network of Cardiovascular Diseases and Depression via Multipartite Projection","authors":"Jie Li, Cillian Hourican, Pashupati P. Mishra, Binisha H. Mishra, Mika Kähönen, Olli T. Raitakari, Reijo Laaksonen, Mika Ala-Korpela, Liisa Keltikangas-Järvinen, Markus Juonala, Terho Lehtimäki, Jos A. Bosch, Rick Quax","doi":"arxiv-2408.07562","DOIUrl":"https://doi.org/arxiv-2408.07562","url":null,"abstract":"There is a significant comorbidity between cardiovascular diseases (CVD) and\u0000depression that is highly predictive of poor clinical outcome. Yet, its\u0000underlying biological pathways remain challenging to decipher, presumably due\u0000to its non-linear associations across multiple mechanisms. Mutual information\u0000provides a framework to analyze such intricacies. In this study, we proposed a\u0000multipartite projection method based on mutual information correlations to\u0000construct multilayer disease networks. We applied the method to a\u0000cross-sectional dataset from a wave of the Young Finns Study. This dataset\u0000assesses CVD and depression, along with related risk factors and two omics of\u0000biomarkers: metabolites and lipids. Instead of directly correlating CVD-related\u0000phenotypes and depressive symptoms, we extended the notion of bipartite\u0000networks to create a multipartite network that connects these phenotype and\u0000symptom variables to intermediate biological variables. Projecting from these\u0000intermediate variables results in a weighted multilayer network, where each\u0000link between CVD and depression variables is marked by its `layer' (i.e.,\u0000metabolome or lipidome). Using this projection method, we identified potential\u0000mediating biomarkers that connect CVD to depression. These biomarkers thus may\u0000play significant roles in the biological pathways of CVD-depression\u0000comorbidity. Additionally, the projected network highlights sex and BMI as the\u0000most important risk factors, or confounders, associated with the comorbidity.\u0000Our method can generalize to any number of omics layers and disease phenotypes,\u0000offering a truly system-level overview of biological pathways contributing to\u0000comorbidity.","PeriodicalId":501309,"journal":{"name":"arXiv - CS - Computational Engineering, Finance, and Science","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142211294","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

M2L Translation Operators for Kernel Independent Fast Multipole Methods on Modern Architectures 现代架构上内核独立快速多极子方法的 M2L 转换算子

arXiv - CS - Computational Engineering, Finance, and Science Pub Date : 2024-08-14 DOI: arxiv-2408.07436

Srinath Kailasa, Timo Betcke, Sarah El Kazdadi

{"title":"M2L Translation Operators for Kernel Independent Fast Multipole Methods on Modern Architectures","authors":"Srinath Kailasa, Timo Betcke, Sarah El Kazdadi","doi":"arxiv-2408.07436","DOIUrl":"https://doi.org/arxiv-2408.07436","url":null,"abstract":"Current and future trends in computer hardware, in which the disparity\u0000between available flops and memory bandwidth continues to grow, favour\u0000algorithm implementations which minimise data movement even at the cost of more\u0000flops. In this study we review the requirements for high performance\u0000implementations of the kernel independent Fast Multipole Method (kiFMM), a\u0000variant of the crucial FMM algorithm for the rapid evaluation of N-body\u0000potential problems. Performant implementations of the kiFMM typically rely on\u0000Fast Fourier Transforms for the crucial M2L (Multipole-to-Local) operation.\u0000However, in recent years for other FMM variants such as the black-box FMM also\u0000BLAS based M2L translation operators have become popular that rely on direct\u0000matrix compression techniques. In this paper we present algorithmic\u0000improvements for BLAS based M2L translation operator and benchmark them against\u0000FFT based M2L translation operators. In order to allow a fair comparison we\u0000have implemented our own high-performance kiFMM algorithm in Rust that performs\u0000competitively against other implementations, and allows us to flexibly switch\u0000between BLAS and FFT based translation operators.","PeriodicalId":501309,"journal":{"name":"arXiv - CS - Computational Engineering, Finance, and Science","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142211341","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

SSAAM: Sentiment Signal-based Asset Allocation Method with Causality Information SSAAM：基于情绪信号的资产配置方法（含因果关系信息

arXiv - CS - Computational Engineering, Finance, and Science Pub Date : 2024-08-13 DOI: arxiv-2408.06585

Rei Taguchi, Hiroki Sakaji, Kiyoshi Izumi

引用次数: 0

An improved point-to-surface contact algorithm with penalty method for peridynamics 改进的点对面接触算法与周动力学惩罚法

arXiv - CS - Computational Engineering, Finance, and Science Pub Date : 2024-08-13 DOI: arxiv-2408.06556

Haoran Zhang, Lisheng Liu, Xin Lai, Jun Li

{"title":"An improved point-to-surface contact algorithm with penalty method for peridynamics","authors":"Haoran Zhang, Lisheng Liu, Xin Lai, Jun Li","doi":"arxiv-2408.06556","DOIUrl":"https://doi.org/arxiv-2408.06556","url":null,"abstract":"It is significantly challenging to obtain accurate contact forces in\u0000peridynamics (PD) simulations due to the difficulty of surface particles\u0000identification, particularly for complex geometries. Here, an improved\u0000point-to-surface contact model is proposed for PD with high accuracy. First,\u0000the outer surface is identified using the eigenvalue method and then we\u0000construct a Verlet list to identify potential contact particle pairs\u0000efficiently. Subsequently, a point-to-surface contact search algorithm is\u0000utilized to determine precise contact locations with the penalty function\u0000method calculating the contact force. Finally, the accuracy of this\u0000point-to-surface contact model is validated through several representative\u0000contact examples. The results demonstrate that the point-to-surface contact\u0000model model can predict contact forces and deformations with high accuracy,\u0000aligning well with the classical Hertz contact theory solutions. This work\u0000presents a contact model for PD that automatically recognizes external surface\u0000particles and accurately calculates the contact force, which provides guidance\u0000for the study of multi-body contact as well as complex contact situations.","PeriodicalId":501309,"journal":{"name":"arXiv - CS - Computational Engineering, Finance, and Science","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142211344","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

High-order projection-based upwind method for simulation of transitional turbulent flows 基于高阶投影的上风法模拟过渡湍流

arXiv - CS - Computational Engineering, Finance, and Science Pub Date : 2024-08-13 DOI: arxiv-2408.06698

Philip L. Lederer, Xaver Mooslechner, Joachim Schöberl

引用次数: 0

Graph Neural Network Approach to Predict the Effects of Road Capacity Reduction Policies: A Case Study for Paris, France 预测道路通行能力削减政策效果的图神经网络方法：法国巴黎案例研究

arXiv - CS - Computational Engineering, Finance, and Science Pub Date : 2024-08-13 DOI: arxiv-2408.06762

Elena Natterer, Roman Engelhardt, Sebastian Hörl, Klaus Bogenberger

{"title":"Graph Neural Network Approach to Predict the Effects of Road Capacity Reduction Policies: A Case Study for Paris, France","authors":"Elena Natterer, Roman Engelhardt, Sebastian Hörl, Klaus Bogenberger","doi":"arxiv-2408.06762","DOIUrl":"https://doi.org/arxiv-2408.06762","url":null,"abstract":"Rapid urbanization and growing urban populations worldwide present\u0000significant challenges for cities, including increased traffic congestion and\u0000air pollution. Effective strategies are needed to manage traffic volumes and\u0000reduce emissions. In practice, traditional traffic flow simulations are used to\u0000test those strategies. However, high computational intensity usually limits\u0000their applicability in investigating a magnitude of different scenarios to\u0000evaluate best policies. This paper introduces an innovative approach to assess\u0000the effects of traffic policies using Graph Neural Networks (GNN). By\u0000incorporating complex transport network structures directly into the neural\u0000network, this approach could enable rapid testing of various policies without\u0000the delays associated with traditional simulations. We provide a proof of\u0000concept that GNNs can learn and predict changes in car volume resulting from\u0000capacity reduction policies. We train a GNN model based on a training set\u0000generated with a MATSim simulation for Paris, France. We analyze the model's\u0000performance across different road types and scenarios, finding that the GNN is\u0000generally able to learn the effects on edge-based traffic volume induced by\u0000policies. The model is especially successful in predicting changes on major\u0000streets. Nevertheless, the evaluation also showed that the current model has\u0000problems in predicting impacts of spatially small policies and changes in\u0000traffic volume in regions where no policy is applied due to spillovers and/or\u0000relocation of traffic.","PeriodicalId":501309,"journal":{"name":"arXiv - CS - Computational Engineering, Finance, and Science","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142211296","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0