检索结果-内蒙古大学图书馆

OS03.6.A UNSUPERVISEd CLUSTERING OF MORPHOLOGY PATTERNS ON WHOLE SLIdE IMAGES GUIdE PROGNOSTIC STRATIFICATION OF GLIOBLASTOMA PATIENTS

引用

Neuro-Oncology 2023年第SUPPLEMENT_2期25卷 ii15-ii15页

作者： Baheti, B Innani, S Nasrallah, M P Bakas, S Center for Artificial Intelligence and Data Science for Integrated Diagnostics (AI2D) and Center for Biomedical Image Computing and Analytics (CBICA) University of Pennsylvania Philadelphia PA United States Department of Pathology and Laboratory Medicine Perelman School of Medicine University of Pennsylvania Philadelphia PA United States Department of Pathology and Laboratory Medicine Department of Radiology Perelman School of Medicine University of Pennsylvania Philadelphia PA United States

来源：评论

学校读者我要写书评

暂无评论

Genetic variants underlying the neuro-anatomical signatures of hypertension measured on structural MRI

引用

Alzheimer's & dementia 2023年第S10期19卷

作者： Sindhuja Tirumalai Govindarajan Elizabeth Mamourian Yuhan Cui Junhao Wen Guray Erus Ahmed Abdulkadir Randa Melhem R Nick Bryan Haochang Shou david A. Wolk Ilya M. Nasrallah Christos davatzikos Centre for Biomedical Image Computing and Analytics University of Pennsylvania Philadelphia PA USA Artificial Intelligence in Biomedical Imaging Laboratory (AIBIL) Center for and Data Science for Integrated Diagnostics (AI2D) Perelman School of Medicine University of Pennsylvania Philadelphia PA USA University of Pennsylvania Health System Philadeplphia PA USA Department of Neurology University of Pennsylvania School of Medicine Philadelphia PA USA

Background Hypertension (HTN) is associated with gray matter (GM) atrophy and increased white matter hyperintensity (WMH) burden, increasing their susceptibility to Alzheimer’s disease and related dementias. We developed a machine learning based model to quantify spatial patterns of abnormality recognizing HTN-related brain changes (SPARE-HTN) from structural magnetic resonance images (sMRI). We performed a genome wide association study (GWAS) to identify the genetic variants and associated biological traits that underlie the “expression” of HTN-related brain changes, i.e., the individualized SPARE-HTN index. Method SPARE-HTN model was trained on the large multi-study iSTAGING dataset, with regional GM volumes from T1-weighted images and lobar WMH volumes from T2-weighted FLaiR images used as imaging features. GWAS was carried out in a subset of N = 7490 (46% Female, Age = 65.5±7 years) hypertensive participants from UK-Biobank. A GWAS of SPARE-HTN scores was run using linear regression with Plink, correcting for the top 40 genetic principal components, age, sex, and intracranial volume. Significant SNPs were identified and mapped to functional genes using the Functional Mapping and Annotation (FUMA) platform; and candidate SNPs were mapped to the GWAS catalog. Result SPARE-HTN was associated with regional atrophy most prominent in perisylvian and temporal cortex and increased WMH volumes, particularly in frontal white matter. SPARE-HTN scores showed better discriminative power between normal and stage 1 participants (Cohen’s-d effect size, d = 0.2) and stage 2 (d = 0.71) hypertensive participants when compared to total WMH volume (d = 0.04 and 0.38, respectively). GWAS identified N = 321 independently significant SNPs, of which a subset N = 46 candidate SNPs were mapped to 6 genes (ICA1L, CARF, WdR12, NBEAL1, KRT8P15, CYP20A1) in Chromosome 2 (Figure 2). These genes were previously reported to be associated with several cardiovascular conditions including coronary

关键词：

来源：评论

学校读者我要写书评

暂无评论

Machine learning with data assimilation and uncertainty quantification for dynamical systems: a review

arXiv

引用

arXiv 2023年

作者： Cheng, Sibo Quilodrán-Casas, César Ouala, Said Farchi, Alban Liu, Che Tandeo, Pierre Fablet, Ronan Lucor, didier Iooss, Bertrand Brajard, Julien Xiao, dunhui Janjic, Tijana ding, Weiping Guo, Yike Carrassi, Alberto Bocquet, Marc Arcucci, Rossella Data Science Institute Department of Computing Imperial College London LondonSW7 2AZ United Kingdom Department of Earth Science and Engineering Imperial College London LondonSW7 2AZ United Kingdom Department of Computer Science and Engineering Hong Kong University of Science and Technology 999077 Hong Kong IMT Atlantique Lab-STICC UMR CNRS 6285 France and Odyssey Inria/IMT France RIKEN Center for Computational Science Kobe Japan CEREA École des Ponts and EDF R&D île-de-France France The Laboratoire Interdisciplinaire des Sciences du Numérique CNRS Paris-Saclay University OrsayF-91403 France 78401 Chatou France Institut de Mathématiques de Toulouse Toulouse31062 France SINCLAIR AI Lab Saclay France Bergen Norway School of Mathematical Sciences Tongji University Shanghai200092 China Mathematical Institute for Machine Learning and Data Science KU Eichstätt-Ingolstadt Bavaria Germany School of Information Science and Technology Nantong University Nantong226019 China Department of Physics and Astronomy Augusto Righi University of Bologna Bologna40124 Italy

data Assimilation (dA) and Uncertainty quantification (UQ) are extensively used in analysing and reducing error propagation in high-dimensional spatial-temporal dynamics. Typical applications span from computational fluid dynamics (CFd) to geoscience and climate systems. Recently, much effort has been given in combining dA, UQ and machine learning (ML) techniques. These research efforts seek to address some critical challenges in high-dimensional dynamical systems, including but not limited to dynamical system identification, reduced order surrogate modelling, error covariance specification and model error correction. A large number of developed techniques and methodologies exhibit a broad applicability across numerous domains, resulting in the necessity for a comprehensive guide. This paper provides the first overview of the state-of-the-art researches in this interdisciplinary field, covering a wide range of applications. This review aims at ML scientists who attempt to apply dA and UQ techniques to improve the accuracy and the interpretability of their models, but also at dA and UQ experts who intend to integrate cutting-edge ML approaches to their systems. Therefore, this article has a special focus on how ML methods can overcome the existing limits of dA and UQ, and vice versa. Some exciting perspectives of this rapidly developing research field are also discussed. © 2023, CC BY-NC-SA.

关键词： dynamical systems

来源：评论

学校读者我要写书评

暂无评论

A multivariate MRI model of Alzheimer’s disease risk is associated with clinical diagnosis, PET imaging, and plasma biomarkers in a mixed dementia sample

引用

Alzheimer's & dementia 2023年第S10期19卷

作者： Jeffrey S Phillips Sindhuja Tirumalai Govindarajan Gyujoon Hwang Guray Erus Katheryn A Q Cousins Sandhitsu R. das david A. Wolk david J. Irwin Murray Grossman Ilya M. Nasrallah Christos davatzikos Perelman School of Medicine University of Pennsylvania Philadelphia PA USA Centre for Biomedical Image Computing and Analytics University of Pennsylvania Philadelphia PA USA Artificial Intelligence in Biomedical Imaging Laboratory (AIBIL) Center for and Data Science for Integrated Diagnostics (AI2D) Perelman School of Medicine University of Pennsylvania Philadelphia PA USA Penn Alzheimer’s Disease Research Center University of Pennsylvania Philadelphia PA USA Frontotemporal Degeneration Center Department of Neurology Perelman School of Medicine University of Pennsylvania Philadelphia PA USA Department of Radiology University of Pennsylvania Philadelphia PA USA

Background We sought to validate two structural MRI-based brain health models: Spatial Pattern of Atrophy for REcognition of Alzheimer’s disease (SPARE-Ad) and SPARE-Brain Age Gap (SPARE-BAG), which estimates the discrepancy between biological brain age and chronological age. These models' generalization to non-amnestic Ad (naAd) syndromes and associations with positron emission tomography (PET) and plasma biomarkers remain underinvestigated. We hypothesized that SPARE-Ad scores in naAd would be lower (less abnormal) than in amnestic Ad (aAd) but elevated (more abnormal) relative to controls; and that SPARE-Ad but not SPARE-BAG scores would be associated with PET and plasma biomarkers of Ad. Method SPARE-Ad and SPARE-BAG scores were estimated for each of 1607 MRI scans from 277 cognitively normal adults (62.8% female; mean age = 69.5 years) and 902 clinical participants (54.2% female; mean age = 73.1 years), including 585 with amnestic mild cognitive impairment (aMCI) or aAd; 56 with naAd syndromes; 28 with vascular disease; 28 with Lewy body disorders (LBd); 18 with suspected frontotemporal lobar degeneration (FTLd); 56 with non-amnestic MCI of uncertain etiology; and 131 with other clinical diagnoses. We used mixed effects models adjusting for age, sex, and global cognition to test group differences in SPARE-Ad and SPARE-BAG. Additionally, we assessed associations with tau (n = 122) and amyloid-beta (n = 199) PET imaging and plasma-based phosphorylated tau-181 (p-tau181) and glial fibrillary acidic protein (GFAP) (n = 358). Result SPARE-Ad scores were elevated vs. controls in aMCI/aAd, naAd, vascular disease, FTLd, and naMCI (all p<0.003) but not in LBd or other diagnoses (p>0.1). NaAd patients had lower SPARE-Ad (p<0.02) but higher SPARE-BAG scores (p<0.02) than aMCI/aAd patients. SPARE-Ad scores were positively associated with mean cortical tau PET uptake, amyloid PET positivity, and plasma p-tau181 and GFAP (all p<0.001). In contrast, SPARE-BAG scores were not

关键词：

来源：评论

学校读者我要写书评

暂无评论

Cross-scale functional connectivity patterns of the aging brain learned from the multi-cohort iSTAGing study

引用

Alzheimer's & dementia 2023年第S17期19卷

作者： Zhen Zhou Hongming Li Chau B Tran Yuncong Ma dhivya Srinivasan Ahmed Abdulkadir Junhao Wen Guray Erus Elizabeth Mamourian Ilya M. Nasrallah Nick Bryan david A. Wolk Lori L Beason-Held Susan M. Resnick Haochang Shou Christos davatzikos Yong Fan University of Pennsylvania philadelphia PA USA Perelman School of Medicine University of Pennsylvania Philadelphia PA USA University of Pennsylvania Philadelphia PA USA Center for Biomedical Image Computing and Analytics University of Pennsylvania Philadelphia PA USA Laboratory for Research in Neuroimaging Department of Clinical Neurosciences Lausanne University Hospital (CHUV) and University of Lausanne Lausanne Switzerland Centre for Biomedical Image Computing and Analytics University of Pennsylvania Philadelphia PA USA Artificial Intelligence in Biomedical Imaging Laboratory (AIBIL) Center for and Data Science for Integrated Diagnostics (AI2D) Perelman School of Medicine University of Pennsylvania Philadelphia PA USA Department of Radiology University of Pennsylvania Philadelphia PA USA Artificial Intelligence in Biomedical Imaging Laboratory (AIBIL) Center for and Data Science for Integrated Diagnostics (AI2D) Perelman School of Medicine University of Pennsylvania Philadelphia PA USA Department of Radiology Perelman School of Medicine University of Pennsylvania Philadelphia PA USA University of Texas at Austin Austin TX USA Department of Pathology and Laboratory Medicine Alzheimer’s Disease Center Perelman School of Medicine University of Pennsylvania Philadelphia PA USA National Institute on Aging Baltimore MD USA Laboratory of Behavioral Neuroscience National Institute on Aging Baltimore MD USA Department of Biostatistics Epidemiology & Informatics University of Pennsylvania Philadelphia PA USA Correspondece

Background Brain functional connectivity (FC) measures derived from resting-state fMRI (rsfMRI) data have advanced our understanding of the brain organization. However, most existing studies investigate the brain functional networks (FNs) at a specific scale, ignoring potentially complementary FC information at different levels of spatial granularity as well as the cross-scale information. Given that the brain is a multi-scale, hierarchical system, FCs between FNs at different spatial scales may provide complementary information to those measured within-scales for characterizing the aging brain. Method We computed multiscale FNs to capture their FC measures across multiple scales and investigated the associations between cross-scale FCs and the brain aging based on rsfMRI data of a diverse cohort (n = 4186, aged 22 to 97 years) from seven sites of the iSTAGING study [1]. For each subject, we computed the FNs at seven scales, yielding seven sets of k (k = 17, 25, 50, 75, 100, 125 and 150) FNs, same as our previous study [2-3]. We estimated FC measures between the 542 FNs using Pearson’s correlation, yielding the cross-scale FC matrices that were harmonized across sites in the tangent space using ComBat-GAM [4] with sex and age as covariates. The harmonized FC measures were used as features to build an age prediction model using Ridge regression as illustrated in Figure 1. The brain age prediction model was compared with that built on harmonized FC measures between FNs of the same scales, with the performance estimated with a nested five-fold cross-validation. Result The cross-scale FC measures performed significantly better than the within-scale FC measures for the brain age prediction, as summarized in Table 1. Figure 2 shows the most informative FC measures for the brain age prediction. Specifically, FC measures computed both cross- and within-scales between the visual, somatomotor, and frontoparietal networks provided relatively more useful information for the bra

关键词：

来源：评论

学校读者我要写书评

暂无评论

discovering Alzheimer’s disease subtypes with imaging and genetic signatures via multi-view weakly-supervised deep clustering

引用

Alzheimer's & dementia 2023年第S17期19卷

作者： Zhijian Yang Junhao Wen Ahmed Abdulkadir Yuhan Cui Guray Erus Elizabeth Mamourian Randa Melhem dhivya Srinivasan Sindhuja Tirumalai Govindarajan Jiong Chen Mohamad Habes Colin L Masters Paul Maruff Jurgen Fripp Luigi Ferrucci Marilyn S. Albert Sterling C Johnson John C Morris Pamela LaMontagne daniel S. Marcus Tammie L.S. Benzinger david A. Wolk Li Shen Jingxuan Bao Susan M. Resnick duygu Tosun Haochang Shou Ilya M. Nasrallah Christos davatzikos the Alzheimer’s disease Neuroimaging Initiative, the Preclinical Ad Consortium, the Baltimore Longitudinal Study of Aging, and the iSTAGING study 1Artificial Intelligence in Biomedical Imaging Laboratory (AIBIL) Center for and Data Science for Integrated Diagnostics (AI2D) Perelman School of Medicine University of Pennsylvania Philadelphia PA USA Correspondece Artificial Intelligence in Biomedical Imaging Laboratory (AIBIL) Center for and Data Science for Integrated Diagnostics (AI2D) Perelman School of Medicine University of Pennsylvania Philadelphia PA USA University of Bern Bern Switzerland Artificial Intelligence in Biomedical Imaging Laboratory (AIBIL) Center for and Data Science for Integrated Diagnostics (AI2D) Perelman School of Medicine University of Pennsylvania Philadelphia PA USA Center for Biomedical Image Computing and Analytics University of Pennsylvania Philadelphia PA USA Centre for Biomedical Image Computing and Analytics University of Pennsylvania Philadelphia PA USA University of Texas Health San Antonio San Antonio TX USA Florey Institute of Neuroscience and Mental Health Parkville VIC Australia The Florey Institute of Neuroscience and Mental Health University of Melbourne Parkville VIC Australia CSIRO Health and Biosecurity Australian E-Health Research Centre Brisbane QLD Australia National Institute on Aging NIH Baltimore MD USA Department of Neurology Johns Hopkins University School of Medicine Baltimore MD USA Wisconsin Alzheimer’s Institute University of Wisconsin-Madison School of Medicine and Public Health Madison WI USA Knight Alzheimer Disease Research Center St. Louis MO USA Washington University School of Medicine Saint Louis MO USA Washington University School of Medicine St. Louis MO USA Department of Neurology University of Pennsylvania School of Medicine Philadelphia PA USA Institute for Biomedical Informatics University of Pennsylvania ATHENS PA USA University of Pennsylvania Philadelphia PA USA Laboratory of Behavioral Neuroscience National Institute on Aging Baltimore MD USA Department of Radiology and Biomedical Imaging University of Cal

Background The heterogeneity of neurodegenerative diseases, including Alzheimer’s disease (Ad), has hampered precision diagnosis and treatment. Machine learning methods enable the identification of genetically-explained disease subtypes with distinct brain phenotypes. Method Gene-SGAN is a novel, multi-view, weakly-supervised deep clustering method that jointly considers phenotypic and genetic data, thereby deriving subtypes of disease-related brain changes with genetic associations. We apply the method to derive subtypes related to Ad. Analyses used structural MRIs and whole-genome sequencing data from the Alzheimer’s disease Neuroimaging Initiative (1,533 participants, including controls and patients with Mild Cognitive Impairment (MCI) and Ad). Result Gene-SGAN identified four distinct subtypes within MCI/Ad: A1, normal anatomy; A2, focal medial temporal lobe (MTL) atrophy; A3, severe atrophy in neocortex and MTL; A4, severe cortical but relatively modest MTL atrophy (Figure 1). The four subtypes reveal significant differences in seven known Ad-related genetic variants (Figure 2). The identified subtype-associated SNPs not only resemble previous findings, but also suggest genetic protective effects contributing to the observed resilience (A1), potential inflammatory mechanisms underlying subtypes (A2&A3), and the effect of SNPs on atypical atrophy patterns of Ad (A4). Clinically, A1 and A3 participants showed the best and the worst cognitive performances (Figure 3). A3 participants had the most abnormal CSF Aβ (p<0.001 vs. A1&A4 and p = 0.039 vs. A2), and exhibited the highest white matter hyperintensity (WMH) volumes (p<0.001 vs. A1&A4 and p = 0.017 vs. A2). However, A2 participants showed significantly higher CSF phospho-tau than A3 participants (p = 0.019). A4 participants were the youngest group (p<0.001 vs. all other groups), suggesting the inclusion of early onset Ad participants, who have a higher prevalence of hippocampal-sparing variants such as posterio

关键词：

来源：评论

学校读者我要写书评

暂无评论

Efficient gravitational wave template bank generation with differentiable waveforms

引用

Physical Review d 2022年第12期106卷 122001-122001页

作者： Adam Coogan Thomas d. P. Edwards Horng Sheng Chia Richard N. George Katherine Freese Cody Messick Christian N. Setzer Christoph Weniger Aaron Zimmerman Ciela—Computation and Astrophysical Data Analysis Institute Montréal Quebec Canada Département de Physique Université de Montréal 1375 Avenue Thérèse-Lavoie-Roux Montréal Quebec H2V 0B3 Canada Mila—Quebec AI Institute 6666 St-Urbain No. 200 Montreal Quebec H2S 3H1 Canada Gravitation Astroparticle Physics Amsterdam (GRAPPA) University of Amsterdam Science Park 904 Amsterdam 1098 XH The Netherlands The Oskar Klein Centre Department of Physics Stockholm University AlbaNova SE-106 91 Stockholm Sweden Nordic Institute for Theoretical Physics (NORDITA) 106 91 Stockholm Sweden School of Natural Sciences Institute for Advanced Study Princeton New Jersey 08540 USA Center for Gravitational Physics University of Texas at Austin Austin Texas 78712 USA Department of Physics University of Texas at Austin Austin Texas 78712 USA LIGO Laboratory Massachusetts Institute of Technology Cambridge Massachusetts 02139 USA

The most sensitive search pipelines for gravitational waves from compact binary mergers use matched filters to extract signals from the noisy data stream coming from gravitational wave detectors. Matched-filter searches require banks of template waveforms covering the physical parameter space of the binary system. Unfortunately, template bank construction can be a time-consuming task. Here we present a new method for efficiently generating template banks that utilizes automatic differentiation to calculate the parameter space metric. Principally, we demonstrate that automatic differentiation enables accurate computation of the metric for waveforms currently used in search pipelines, whilst being computationally cheap. Additionally, by combining random template placement and a Monte Carlo method for evaluating the fraction of the parameter space that is currently covered, we show that search-ready template banks for frequency-domain waveforms can be rapidly generated. Finally, we argue that differentiable waveforms offer a pathway to accelerating stochastic placement algorithms. We implement all our methods into an easy-to-use python package based on the jax framework, diffbank, to allow the community to easily take advantage of differentiable waveforms for future searches.

关键词： Gravitational wave detection Laboratory studies of gravity

来源：评论

学校读者我要写书评

暂无评论

Novel genomic loci shape two neuroanatomical dimensions of Ad in non-demented populations

引用

Alzheimer's & dementia 2023年第S10期19卷

作者： Junhao Wen Zhijian Yang Ioanna Skampardoni Marilena de Pian Guray Erus Elizabeth Mamourian Yuhan Cui Bingxin Zhao Ilya M. Nasrallah Arthur W. Toga Haochang Shou Li Shen Christos davatzikos Artificial Intelligence in Biomedical Imaging Laboratory (AIBIL) Center for and Data Science for Integrated Diagnostics (AI2D) Perelman School of Medicine University of Pennsylvania Philadelphia PA USA Center for Biomedical Image Computing and Analytics University of Pennsylvania Philadelphia PA USA Centre for Biomedical Image Computing and Analytics University of Pennsylvania Philadelphia PA USA University of Pennsylvania philadelphia PA USA Upenn Philadelphia PA USA University of Pennsylvania Philadelphia PA USA Laboratory of Neuro Imaging (LONI) University of Southern California Los Angeles CA USA Department of Biostatistics Epidemiology & Informatics University of Pennsylvania Philadelphia PA USA University of Pennsylvania Philadelphia PA Greece

Background Artificial intelligence has been in burgeoning demand to be applied to magnetic imaging resonance (MRI) data to parse the neuroanatomical heterogeneity of Alzheimer’s disease (Ad) and/or mild cognitive impairment (MCI) (Wen et al., 2021a, 2021b; Yang et al., 2021). However, the genetic heterogeneity leading to this neuroanatomical heterogeneity, especially at the early asymptomatic stages, is not fully elucidated. Method A deep semi-supervised representation learning method, called Surreal-GAN (Yang et al., 2022), was applied to T1-weighted MRI from Ad/MCI patients from AdNI to define the two neuroanatomical dimensions of Ad at the symptomatic stages. The R1 dimension was characterized by widespread brain atrophy without an exclusive focus on the medial temporal lobe, whereas the R2 dimension showed focal medial temporal lobe atrophy (Wen et al., 2022). We then applied the trained model to 33,540 non-demented individuals from UK Biobank. Imputed genotyping data were analyzed in genome-wide association studies (GWAS). We performed multiple linear regressions controlling for confounding factors (e.g., age) via Plink (Purcell et al., 2007). We defined a genomic locus (represented by the top lead SNP) considering linkage disequilibrium. Finally, we manually queried all leading SNPs on GWAS Catalog (Buniello et al., 2019) to determine if the genomic loci were novel, provided they were not previously associated with any clinical traits. Result GWAS discovered 24 genomic loci, of which 7 were novel for R1 and R2, respectively (Fig. 1). For R1, 8 mapped genes, based on physical positions, eQTL, and chromatin interactions, were associated with Ad-related traits in the literature. For example, the KdM2B gene from the genomic locus (rs9795600, P-value = 1.13e-08) was previously associated with Ad (Gouveia et al., 2022). For R2, 13 mapped genes were associated with Ad-related traits. For example, the TESC gene from the genomic locus (rs77956314, P-value = 9.99e-29) w

关键词：

来源：评论

学校读者我要写书评

暂无评论

The tracking machine learning challenge: Accuracy phase

arXiv

引用

arXiv 2019年

作者： Amrouche, Sabrina Basara, Laurent Calafiura, Paolo Estrade, Victor Farrell, Steven Ferreira, diogo R. Finnie, Liam Finnie, Nicole Germain, Cécile Gligorov, Vladimir Vava Golling, Tobias Gorbunov, Sergey Gray, Heather Guyon, Isabelle Hushchyn, Mikhail Innocente, Vincenzo Kiehn, Moritz Moyse, Edward Puget, Jean-François Reina, Yuval Rousseau, david Salzburger, Andreas Ustyuzhanin, Andrey Vlimant, Jean-Roch Wind, Johan Sokrates Xylouris, Trian Yilmaz, Yetkin Département de Physique Nucléaire et Corpusculaire Université de Genève Geneva Switzerland LRI TAU Univ. Paris-Sud INRIA CNRS Université Paris-Saclay Gif-sur-Yvette France Physics Division Lawrence Berkeley National Laboratory University of California BerkeleyCA United States IST University of Lisbon Lisbon Portugal IBM Germany Research and Development Germany Bosch Center for Artificial Intelligence Germany LPNHE Sorbonne Université Paris Diderot Sorbonne Paris Cité CNRS IN2P3 Paris France Goethe University Frankfurt am Main Germany UPSud INRIA Université Paris-Saclay Orsay France ChaLearn BerkeleyCA United States National Research University Higher School of Economics Yandex School of Data Analysis Moscow Russia CERN Geneva Switzerland Department of Physics University of Massachusetts AmherstMA United States Data and AI R&D IBM France Lab Biot France Tel-Aviv Israel LAL Univ. Paris-Sud CNRS IN2P3 Université Paris-Saclay Orsay France California Institute of Technology PasadenaCA United States Norwegian University of Science and Technology Oslo Norway Frankfurt Germany

This paper reports the results of an experiment in high energy physics: using the power of the "crowd" to solve difficult experimental problems linked to tracking accurately the trajectory of particles in the Large Hadron Collider (LHC). This experiment took the form of a machine learning challenge organized in 2018: the Tracking Machine Learning Challenge (TrackML). Its results were discussed at the competition session at the Neural Information Processing Systems conference (NeurIPS 2018). Given 100.000 points, the participants had to connect them into about 10.000 arcs of circles, following the trajectory of particles issued from very high energy proton collisions. The competition was difficult with a dozen front-runners well ahead of a pack. The single competition score is shown to be accurate and effective in selecting the best algorithms from the domain point of view. The competition has exposed a diversity of approaches, with various roles for Machine Learning, a number of which are discussed in the document. © 2019, CC BY.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Why is the Winner the Best?

Why is the Winner the Best?

引用

Conference on Computer Vision and Pattern Recognition (CVPR)

作者： M. Eisenmann A. Reinke V. Weru M. d. Tizabi F. Isensee T. J. Adler S. Ali V. Andrearczyk M. Aubreville U. Baid S. Bakas N. Balu S. Bano J. Bernal S. Bodenstedt A. Casella V. Cheplygina M. daum M. de Bruijne A. depeursinge R. dorent J. Egger d. G. Ellis S. Engelhardt M. Ganz N. Ghatwary G. Girard P. Godau A. Gupta L. Hansen K. Harada M. Heinrich N. Heller A. Hering A. Huaulmé P. Jannin A. E. Kavur O. Kodym M. Kozubek J. Li H. Li J. Ma C. Martín-Isla B. Menze A. Noble V. Oreiller N. Padoy S. Pati K. Payette T. Rädsch J. Rafael-Patiño V. Singh Bawa S. Speidel C. H. Sudre K. Van Wijnen M. Wagner d. Wei A. Yamlahi M. H. Yap C. Yuan M. Zenk A. Zia d. Zimmerer d. Aydogan B. Bhattarai L. Bloch R. Brüngel J. Cho C. Choi Q. dou I. Ezhov C. M. Friedrich C. Fuller R. R. Gaire A. Galdran Á. García Faura M. Grammatikopoulou S. Hong M. Jahanifar I. Jang A. Kadkhodamohammadi I. Kang F. Kofler S. Kondo H. Kuijf M. Li M. Luu T. Martinčič P. Morais M. A. Naser B. Oliveira d. Owen S. Pang J. Park S. Park S. Płotka E. Puybareau N. Rajpoot K. Ryu N. Saeed A. Shephard P. Shi d. Štepec R. Subedi G. Tochon H. R. Torres H. Urien J. L. Vilaça K. A. Wahid H. Wang J. Wang L. Wang X. Wang B. Wiestler M. Wodzinski F. Xia J. Xie Z. Xiong S. Yang Y. Yang Z. Zhao K. Maier-Hein P. F. Jäger A. Kopp-Schneider L. Maier-Hein Division of Intelligent Medical Systems German Cancer Research Center (DKFZ) Heidelberg Germany Helmholtz Imaging German Cancer Research Center (DKFZ) Heidelberg Germany Faculty of Mathematics and Computer Science Heidelberg University Heidelberg Germany Division of Biostatistics German Cancer Research Center (DKFZ) Heidelberg Germany Division of Medical Image Computing German Cancer Research Center (DKFZ) Heidelberg Germany Faculty of Engineering and Physical Sciences School of Computing University of Leeds Leeds UK Institute of Informatics School of Management HES-SO Valais-Wallis University of Applied Sciences and Arts Western Switzerland Sierre Switzerland Department of Nuclear Medicine and Molecular Imaging Lausanne University Hospital Lausanne Switzerland Technische Hochschule Ingolstadt Ingolstadt Germany Center for Artificial Intelligence and Data Science for Integrated Diagnostics (AI2D) and Center for Biomedical Image Computing and Analytics (CBICA) University of Pennsylvania Philadelphia PA USA Department of Pathology and Laboratory Medicine Perelman School of Medicine University of Pennsylvania Philadelphia PA USA Department of Radiology Perelman School of Medicine University of Pennsylvania Philadelphia PA USA Department of Radiology University of Washington Seattle WA USA Department of Computer Science Wellcome/EPSRC Centre for Interventional and Surgical Sciences (WEISS) University College London London UK Universitat Autònoma de Barcelona & Computer Vision Center Barcelona Spain Division of Translational Surgical Oncology National Center for Tumor Diseases (NCT/UCC) Dresden Dresden Germany Department of Advanced Robotics Istituto Italiano di Tecnologia Italy Department of Electronics Information and Bioengineering Politecnico di Milano Milan Italy IT University of Copenhagen Copenhagen Denmark Department of General Visceral and Transplantation Surgery Heidelberg University Hospital Heidelberg Germany Department of Radiology and Nuc

International benchmarking competitions have become fundamental for the comparative performance assessment of image analysis methods. However, little attention has been given to investigating what can be learnt from these competitions. do they really generate scientific progress? What are common and successful participation strategies? What makes a solution superior to a competing method? To address this gap in the literature, we performed a multicenter study with all 80 competitions that were conducted in the scope of IEEE ISBI 2021 and MICCai 2021. Statistical analyses performed based on comprehensive descriptions of the submitted algorithms linked to their rank as well as the underlying participation strategies revealed common characteristics of winning solutions. These typically include the use of multi-task learning (63%) and/or multi-stage pipelines (61%), and a focus on augmentation (100%), image preprocessing (97%), data curation (79%), and post-processing (66%). The “typical” lead of a winning team is a computer scientist with a doctoral degree, five years of experience in biomedical image analysis, and four years of experience in deep learning. Two core general development strategies stood out for highly-ranked teams: the reflection of the metrics in the method design and the focus on analyzing and handling failure cases. According to the organizers, 43% of the winning algorithms exceeded the state of the art but only 11% completely solved the respective domain problem. The insights of our study could help researchers (1) improve algorithm development strategies when approaching new problems, and (2) focus on open research questions revealed by this work.

关键词：

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：