检索结果-内蒙古大学图书馆

Deep neural networks and PIDE discretizations

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Bohn, Bastian Griebel, Michael Kannan, Dinesh Fraunhofer Center for Machine Learning Schloss Birlinghoven Sankt Augustin53754 Germany Fraunhofer Institute for Algorithms and Scientific Computing SCAI Schloss Birlinghoven Sankt Augustin53754 Germany Institute for Numerical Simulation University of Bonn Friedrich-Hirzebruch-Allee 7 Bonn53115 Germany

In this paper, we propose neural networks that tackle the problems of stability and field-of-view of a Convolutional Neural Network (CNN). As an alternative to increasing the network's depth or width to improve performance, we propose integral-based spatially nonlocal operators which are related to global weighted Laplacian, fractional Laplacian and inverse fractional Laplacian operators that arise in several problems in the physical sciences. The forward propagation of such networks is inspired by partial integro-differential equations (PIDEs). We test the effectiveness of the proposed neural architectures on benchmark image classification datasets and semantic segmentation tasks in autonomous driving. Moreover, we investigate the extra computational costs of these dense operators and the stability of forward propagation of the proposed neural *** Codes 65D15, 65L07, 68T05, 68W25, 47G20, 47G30 © 2021, CC BY.

关键词： Deep neural networks

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Bulin, Johannes Hamaekers, Jan Department of Virtual Material Design Fraunhofer Institute for Algorithms and Scientific Computing Schloss Birlinghoven Sankt Augustin53757 Germany Fraunhofer Center for Machine Learning Schloss Birlinghoven Sankt Augustin53757 Germany

Determining whether two particle systems are similar is a common problem in particle simulations. When the comparison should be invariant under permutations, orthogonal transformations, and translations of the systems, special techniques are needed. We present an algorithm that can test particle systems of finite size for similarity and, if they are similar, can find the optimal alignment between them. Our approach is based on an invariant version of the root mean square deviation (RMSD) measure and is capable of finding the globally optimal solution in O(n3) operations where n is the number of three-dimensional particles. Copyright © 2021, The Authors. All rights reserved.

关键词： machine learning

Predictive analytics in quality assurance for assembly processes: lessons learned from a case study at an industry 4.0 demonstration cell

学校读者我要写书评

暂无评论

Procedia CIRP 2021年 104卷 641-646页

作者： Peter Burggräf Johannes Wagner Benjamin Heinbach Fabian Steinberg Alejandro R. Pérez M. Lennart Schmallenbach Jochen Garcke Daniela Steffes-lai Moritz Wolter Chair of International Production Engineering and Management (IPEM) Universität Siegen Paul-Bonatz-Straße 9-11 57076 Siegen Germany Fraunhofer Institute for Algorithms and Scientific Computing (SCAI) Schloss Birlinghoven 1 53757 Sankt Augustin Germany Fraunhofer Center for Machine Learning Schloss Birlinghoven 1 53757 Sankt Augustin Germany Institute for Numerical Simulation Universität Bonn Endenicher Allee 19b 53115 Bonn Germany Institute for Computer Science Universität Bonn Endenicher Allee 19a 53115 Bonn Germany

Quality assurance (QA) is an important task in manufacturing to assess whether products meet their specifications. However, QA might be expensive, time-consuming, or incomplete. This paper presents a solution for predictive analytics in QA based on machine sensor values during production while employing specialized machine-learning models for classification in a controlled environment. Furthermore, we present lessons learned while implementing this model, which helps to reduce complexity in further industrial applications. The paper’s outcome proves that the developed model was able to predict product quality, as well as to identify the correlation between machine-status and faulty product occurrence.

关键词： machine-learning Predictive Quality Production Quality Assurance Classification

SYNTHETIC DATA GENERATION FOR A LONGITUDINAL COHORT STUDY - EVALUATION, METHOD EXTENSION AND REPRODUCTION OF PUBLISHED DATA ANALYSIS RESULTS

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Kühnel, Lisa Schneider, Julian Perrar, Ines Adams, Tim Prasser, Fabian Nöthlings, Ute Fröhlich, Holger Fluck, Juliane Knowledge Management Zb Med Information Centre for Life Sciences Cologne50931 Germany Faculty of Technology Bielefeld University Bielefeld33615 Germany Institute of Nutritional and Food Sciences - Nutritional Epidemiology University of Bonn Bonn53115 Germany Department of Bioinformatics Fraunhofer Institute for Algorithms and Scientific Computing Scai Sankt Augustin53757 Germany Medical Informatics Group Berlin Institute of Health Charité - Universitätsmedizin Berlin Berlin10117 Germany Bonn-Aachen International Center for It University of Bonn Friedrich Hirzebruch-Allee 6 Bonn53115 Germany The Agricultural Faculty University of Bonn Bonn53115 Germany

Access to individual-level health data is essential for gaining new insights and advancing science. In particular, modern methods based on artificial intelligence rely on the availability of and access to large datasets. In the health sector, access to individual-level data is often challenging due to privacy concerns. A promising alternative is the generation of fully synthetic data, i.e. data generated through a randomised process that have similar statistical properties as the original data, but do not have a one-to-one correspondence with the original individual-level records. In this study, we use a state-of-the-art synthetic data generation method and perform in-depth quality analyses of the generated data for a specific use case in the field of nutrition. We demonstrate the need for careful analyses of synthetic data that go beyond descriptive statistics and provide valuable insights into how to realise the full potential of synthetic datasets. By extending the methods, but also by thoroughly analysing the effects of sampling from a trained model, we are able to largely reproduce significant real-world analysis results in the chosen use case. © 2023, CC BY.

关键词： Large dataset

Scalable hyperparameter optimization with lazy Gaussian processes

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Ram, Raju Müller, Sabine Pfreundt, Franz-Josef Gauger, Nicolas R. Keuper, Janis Competence Center High Performance Computing Fraunhofer ITWM Kaiserslautern Germany Fraunhofer Center Machine Learning Germany Scientific Computing Group TU Kaiserslautern Germany Institute for Machine Learning and Analytics Offenburg University Germany

Most machine learning methods require careful selection of hyper-parameters in order to train a high performing model with good generalization abilities. Hence, several automatic selection algorithms have been introduced to overcome tedious manual (try and error) tuning of these parameters. Due to its very high sample efficiency, Bayesian Optimization over a Gaussian Processes modeling of the parameter space has become the method of choice. Unfortunately, this approach suffers from a cubic compute complexity due to underlying Cholesky factorization, which makes it very hard to be scaled beyond a small number of sampling steps. In this paper, we present a novel, highly accurate approximation of the underlying Gaussian Process. Reducing its computational complexity from cubic to quadratic allows an efficient strong scaling of Bayesian Optimization while outperforming the previous approach regarding optimization accuracy. The first experiments show speedups of a factor of 162 in single node and further speed up by a factor of 5 in a parallel environment. Copyright © 2020, The Authors. All rights reserved.

关键词： Gaussian distribution

Combating mode collapse in GAN training: An empirical analysis using hessian eigenvalues

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Lopez, Ricard Durall Chatzimichailidis, Avraam Labus, Peter Keuper, Janis Fraunhofer ITWM Germany IWR University of Heidelberg Germany Chair for Scientific Computing TU Kaiserslautern Germany Fraunhofer Center Machine Learning Germany Institute for Machine Learning and Analytics Offenburg University Germany

Generative adversarial networks (GANs) provide state-of-the-art results in image generation. However, despite being so powerful, they still remain very challenging to train. This is in particular caused by their highly non-convex optimization space leading to a number of instabilities. Among them, mode collapse stands out as one of the most daunting ones. This undesirable event occurs when the model can only fit a few modes of the data distribution, while ignoring the majority of them. In this work, we combat mode collapse using second-order gradient information. To do so, we analyse the loss surface through its Hessian eigenvalues, and show that mode collapse is related to the convergence towards sharp minima. In particular, we observe how the eigenvalues of the G are directly correlated with the occurrence of mode collapse. Finally, motivated by these findings, we design a new optimization algorithm called nudged-Adam (NuGAN) that uses spectral information to overcome mode collapse, leading to empirically more stable convergence properties. © 2020, CC BY.

关键词： Eigenvalues and eigenfunctions

algorithms in the orchard: An embedding-based expert answering system for apple rust

学校读者我要写书评

暂无评论

Smart Agricultural Technology

Smart Agricultural Technology 2025年 12卷

作者： Astha Anand Jian Shen Armin Bernd Cremers Marc Jacobs Fraunhofer Institute for Algorithms and Scientific Computing SCAI Schloss Birlinghoven 1 Sankt Augustin 53757 North Rhine-Westphalia Germany RPU Rhenish Plant Power & Envirotech Co. Ltd Germany Bonn-Aachen International Center for Information Technology Friedrich-Hirzebruch-Allee 5 Bonn 53115 North Rhine-Westphalia Germany

As sustainable agricultural practices gain importance, the need for intelligent pest control decision-making has grown. This paper introduces SEEDS: Similarity-based Expert Embedding Decision System, a Retrieval-Augmented Generation (RAG) based agricultural question-answering (QA) system. It is built upon a domain-specific knowledge graph (KG), representing Cedar Apple Rust disease, its host and causative agents, plant defense molecules against apple rust infection, and various pesticides. Utilizing the OpenAI embedding model, the system generates embeddings for user queries and KG data, employing similarity metrics to rank KG entries, facilitating accurate and relevant pest control recommendations. SEEDS is a promising niche AI tool in plant protection, setting the stage for scalable, extensible QA frameworks in precision agriculture. The results signify not only a step forward in agricultural expert systems but also highlight the potential for expanding this approach to other crops and pests, marking a substantial advancement in the use of AI for agricultural pest control.

关键词： Embeddings Retrieval-augmented generation Knowledge graph Apple rust Large language model Generative artificial intelligence Agricultural pest control

Assessing the landscape of Alzheimer’s disease cohort data

学校读者我要写书评

暂无评论

Alzheimer's & Dementia 2020年第S4期16卷

作者： Colin Birkenbihl Yasamin Salimi Daniel Domingo-Fernández Holger Fröhlich Martin Hofmann-Apitius Fraunhofer Institute for Algorithms and Scientific Computing (SCAI) Sankt Augustin Germany Bonn-Aachen International Center for IT (b-it) Bonn Germany Fraunhofer Institute for Algorithms and Scientific Computing SCAI Sankt Augustin Germany

Background Data collected in cohort studies lay the groundwork for a plethora of Alzheimer’s disease (AD) research endeavours. While there exist numerous cohort datasets in our field, few dedicated efforts have focused on comparing these datasets to each other. In recent years, large research consortia, such as EMIF and ROADMAP, were formed to investigate and evaluate the AD data landscape. To the best of our knowledge, such previous approaches have solely investigated cohorts at the metadata level without considering the cohort dataset itself. In this work, we explored eight major AD cohort datasets with the aim of 1) characterizing their underlying data, 2) assessing the quantity and availability of data, 3) evaluating the interoperability across distinct cohort datasets, and 4) presenting our findings in an interactive web application. Thus, we allow researchers to investigate the AD data landscape and find the most suited datasets for their research. Method We collected data from multiple major AD cohort studies. Through intensive manual curation of these datasets, we identified which data modalities can be found in the shared data. Using descriptive statistics and visualizations, we highlighted differences and similarities between cohorts focusing on key demographic characteristics and AD biomarkers. Finally, we assessed and compared the longitudinal follow-up data for each study. Result We show that demographics can vary substantially across cohorts (Table 1). The overlap between datasets is limited with regard to assessed measurements (Figure 1). Additionally, distributions of the same variables often differ across cohorts due to differences in processing and normalization; thus, hampering dataset interoperability. By investigating longitudinal follow-up data, we demonstrate that although there is reasonable longitudinal coverage on cognitive assessments, important biomarkers were only collected for a small fraction of study participants over time (Figure 2).

关键词：

Predicting dementia in people with Parkinson’s disease

学校读者我要写书评

暂无评论

npj Parkinson's Disease 2025年第1期11卷 1-10页

作者： Aborageh, Mohamed Hähnel, Tom Martins Conde, Patricia Klucken, Jochen Fröhlich, Holger Department of Bioinformatics Fraunhofer Institute for Algorithms and Scientific Computing (SCAI) Augustin Sankt 53757 Germany Department of Neurology University Hospital and Faculty of Medicine Carl Gustav Carus TUD Dresden University of Technology Dresden Germany Luxembourg Centre for Systems Biomedicine (LCSB) University of Luxembourg Esch-sur-Alzette Luxembourg Centre Hospitalier de Luxembourg (CHL) Luxembourg Luxembourg Bonn-Aachen International Center for Information Technology (B-IT) Rheinische Friedrich-Wilhelms-Universität Bonn Bonn 53115 Germany

Parkinson’s disease (PD) exhibits a variety of symptoms, with approximately 25% of patients experiencing mild cognitive impairment and 45% developing dementia within ten years of diagnosis. Predicting this progression and identifying its causes remains challenging. Our study utilizes machine learning and multimodal data from the UK Biobank to explore the predictability of Parkinson’s dementia (PDD) post-diagnosis, further validated by data from the Parkinson’s Progression Markers Initiative (PPMI) cohort. Using Shapley Additive Explanation (SHAP) and Bayesian Network structure learning, we analyzed interactions among genetic predisposition, comorbidities, lifestyle, and environmental factors. We concluded that genetic predisposition is the dominant factor, with significant influence from comorbidities. Additionally, we employed Mendelian randomization (MR) to establish potential causal links between hypertension, type 2 diabetes, and PDD, suggesting that managing blood pressure and glucose levels in Parkinson’s patients may serve as a preventive strategy. This study identifies risk factors for PDD and proposes avenues for prevention. © The Author(s) 2025.

关键词：