检索结果-内蒙古大学图书馆

Understanding metric-related pitfalls in image analysis validation

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Reinke, Annika Tizabi, Minu D. Baumgartner, Michael Eisenmann, Matthias Heckmann-Nötzel, Doreen Kavur, A. Emre Rädsch, Tim Sudre, Carole H. Acion, Laura Antonelli, Michela Arbel, Tal Bakas, Spyridon Benis, Arriel Blaschko, Matthew B. Buettner, Florian Cardoso, M. Jorge Cheplygina, Veronika Chen, Jianxu Christodoulou, Evangelia Cimini, Beth A. Collins, Gary S. Farahani, Keyvan Ferrer, Luciana Galdran, Adrian van Ginneken, Bram Glocker, Ben Godau, Patrick Haase, Robert Hashimoto, Daniel A. Hoffman, Michael M. Huisman, Merel Isensee, Fabian Jannin, Pierre Kahn, Charles E. Kainmueller, Dagmar Kainz, Bernhard Karargyris, Alexandros Karthikesalingam, Alan Kenngott, Hannes Kleesiek, Jens Kofler, Florian Kooi, Thijs Kopp-Schneider, Annette Kozubek, Michal Kreshuk, Anna Kurc, Tahsin Landman, Bennett A. Litjens, Geert Madani, Amin Maier-Hein, Klaus Martel, Anne L. Mattson, Peter Meijering, Erik Menze, Bjoern Moons, Karel G.M. Müller, Henning Nichyporuk, Brennan Nickel, Felix Petersen, Jens Rafelski, Susanne M. Rajpoot, Nasir Reyes, Mauricio Riegler, Michael A. Rieke, Nicola Saez-Rodriguez, Julio Sánchez, Clara I. Shetty, Shravya Summers, Ronald M. Taha, Abdel A. Tiulpin, Aleksei Tsaftaris, Sotirios A. van Calster, Ben Varoquaux, Gaël Yaniv, Ziv R. Jäger, Paul F. Maier-Hein, Lena Faculty of Mathematics and Computer Science Heidelberg University Heidelberg Germany Heidelberg Division of Intelligent Medical Systems Germany NCT Heidelberg A Partnership Between DKFZ University Medical Center Heidelberg Germany Heidelberg Division of Medical Image Computing Germany Heidelberg Division of Intelligent Medical Systems Germany MRC Unit for Lifelong Health and Ageing UCL Centre for Medical Image Computing Department of Computer Science University College London London United Kingdom School of Biomedical Engineering and Imaging Science King’s College London London United Kingdom Instituto de Cálculo CONICET – Universidad de Buenos Aires Buenos Aires Argentina Centre for Medical Image Computing University College London London United Kingdom McGill University Montreal Canada Division of Computational Pathology Dept of Pathology & Laboratory Medicine Indiana University School of Medicine IU Health Information and Translational Sciences Building Indianapolis United States University of Pennsylvania Richards Medical Research Laboratories FL7 PhiladelphiaPA United States Department of Digital Medical Technologies Holon Institute of Technology Holon Israel European Federation for Medical Informatics Le Mont-sur-Lausanne Switzerland Center for Processing Speech and Images Department of Electrical Engineering KU Leuven Leuven Belgium partner site Frankfurt/Mainz a partnership between DKFZ and UCT Frankfurt Marburg Germany Heidelberg Germany Goethe University Frankfurt Department of Medicine Germany Goethe University Frankfurt Department of Informatics Germany and Frankfurt Cancer Insititute Germany Department of Computer Science IT University of Copenhagen Copenhagen Denmark Leibniz-Institut für Analytische Wissenschaften – ISAS – e.V. Dortmund Germany Imaging Platform Broad Institute of MIT and Harvard CambridgeMA United States Centre for Statistics in Medicine University of Oxford Oxford United Kingdom Center for Biomedical In

Validation metrics are key for tracking scientific progress and bridging the current chasm between artificial intelligence (AI) research and its translation into practice. However, increasing evidence shows that particularly in image analysis, metrics are often chosen inadequately. While taking into account the individual strengths, weaknesses, and limitations of validation metrics is a critical prerequisite to making educated choices, the relevant knowledge is currently scattered and poorly accessible to individual researchers. Based on a multi-stage Delphi process conducted by a multidisciplinary expert consortium as well as extensive community feedback, the present work provides the first reliable and comprehensive common point of access to information on pitfalls related to validation metrics in image analysis. While focused on biomedical image analysis, the addressed pitfalls generalize across application domains and are categorized according to a newly created, domain-agnostic taxonomy. The work serves to enhance global comprehension of a key topic in image analysis validation. © 2023, CC BY.

关键词： computer vision

Analyzing the Structure of Attention in a Transformer Language Model

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Vig, Jesse Belinkov, Yonatan Palo Alto Research Center Machine Learning and Data Science Group Interaction and Analytics Lab Palo AltoCA United States Harvard John A. Paulson School of Engineering and Applied Sciences MIT Computer Science and Artificial Intelligence Laboratory CambridgeMA United States

The Transformer is a fully attention-based alternative to recurrent networks that has achieved state-of-the-art results across a range of NLP tasks. In this paper, we analyze the structure of attention in a Transformer language model, the GPT-2 small pretrained model. We visualize attention for individual instances and analyze the interaction between attention and syntax over a large corpus. We find that attention targets different parts of speech at different layer depths within the model, and that attention aligns with dependency relations most strongly in the middle layers. We also find that the deepest layers of the model capture the most distant relationships. Finally, we extract exemplar sentences that reveal highly specific patterns targeted by particular attention heads. Copyright © 2019, The Authors. All rights reserved.

关键词： Recurrent neural networks

The second data release from the European Pulsar Timing Array: II. Customised pulsar noise models for spatially correlated gravitational waves

学校读者我要写书评

暂无评论

ASTRONOMY & ASTROPHYSICS 2023年 678卷

作者： [Anonymous] Institute of Astrophysics FORTH N. Plastira 100 70013 Heraklion Greece Max-Planck-Institut für Radioastronomie Auf dem Hügel 69 53121 Bonn Germany Department of Physics Indian Institute of Technology Roorkee Roorkee-247667 India Department of Electrical Engineering IIT Hyderabad Kandi Telangana 502284 India Université Paris-Cité CNRS Astroparticule et Cosmologie 75013 Paris France The Institute of Mathematical Sciences C. I. T. Campus Taramani Chennai 600113 India Homi Bhabha National Institute Training School Complex Anushakti Nagar Mumbai 400094 India Fakultät für Physik Universität Bielefeld Postfach 100131 33501 Bielefeld Germany ASTRON Netherlands Institute for Radio Astronomy Oude Hoogeveensedijk 4 7991 PD Dwingeloo The Netherlands Department of Physical Sciences Indian Institute of Science Education and Research Mohali Punjab 140306 India Laboratoire de Physique et Chimie de l’Environnement et de l’Espace Université d’Orléans / CNRS 45071 Orléans Cedex 02 France Observatoire Radioastronomique de Nançay Observatoire de Paris Université PSL Université d’Orléans CNRS 18330 Nançay France Dipartimento di Fisica “G. Occhialini” Università degli Studi di Milano-Bicocca Piazza della Scienza 3 20126 Milano Italy INFN Sezione di Milano-Bicocca Piazza della Scienza 3 20126 Milano Italy INAF – Osservatorio Astronomico di Brera via Brera 20 20121 Milano Italy Institute for Gravitational Wave Astronomy and School of Physics and Astronomy University of Birmingham Edgbaston Birmingham B15 2TT UK INAF – Osservatorio Astronomico di Cagliari via della Scienza 5 09047 Selargius (CA) Italy Hellenic Open University School of Science and Technology 26335 Patras Greece Kavli Institute for Astronomy and Astrophysics Peking University Beijing 100871 PR China Department of Astronomy and Astrophysics Tata Institute of Fundamental Research Homi Bhabha Road Navy Nagar Colaba Mumbai

Aims. The nanohertz gravitational wave background (GWB) is expected to be an aggregate signal of an ensemble of gravitational waves emitted predominantly by a large population of coalescing supermassive black hole binaries in the centres of merging galaxies. Pulsar timing arrays (PTAs), which are ensembles of extremely stable pulsars at approximately kiloparsec distances precisely monitored for decades, are the most precise experiments capable of detecting this background. However, the subtle imprints that the GWB induces on pulsar timing data are obscured by many sources of noise that occur on various timescales. These must be carefully modelled and mitigated to increase the sensitivity to the background ***. In this paper, we present a novel technique to estimate the optimal number of frequency coefficients for modelling achromatic and chromatic noise, while selecting the preferred set of noise models to use for each pulsar. We also incorporated a new model to fit for scattering variations in the Bayesian pulsar timing package temponest. These customised noise models enable a more robust characterisation of single-pulsar noise. We developed a software package based on tempo2 to create realistic simulations of European Pulsar Timing Array (EPTA) datasets that allowed us to test the efficacy of our noise modelling ***. Using these techniques, we present an in-depth analysis of the noise properties of 25 millisecond pulsars (MSPs) that form the second data release (DR2) of the EPTA and investigate the effect of incorporating low-frequency data from the Indian Pulsar Timing Array collaboration for a common sample of ten MSPs. We used two packages, enterprise and temponest, to estimate our noise models and compare them with those reported using EPTA DR1. We find that, while in some pulsars we can successfully disentangle chromatic from achromatic noise owing to the wider frequency coverage in DR2, in others the noise models evolve in a much mo

关键词： pulsars: general gravitational waves methods: statistical

Improving generalization via attribute selection on out-of-the-box data

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Xu, Xiaofeng Tsang, Ivor W. Liu, Chuancai School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing210094 China Centre for Artificial Intelligence University of Technology Sydney UltimoNSW2007 Australia Collaborative Innovation Center of IoT Technology and Intelligent Systems Minjiang University Fuzhou350108 China

Zero-shot learning (ZSL) aims to recognize unseen objects (test classes) given some other seen objects (training classes), by sharing information of attributes between different objects. Attributes are artificially annotated for objects and are treated equally in recent ZSL tasks. However, some inferior attributes with poor predictability or poor discriminability may have negative impact on the ZSL system performance. This paper first derives a generalization error bound for ZSL tasks. Our theoretical analysis verifies that selecting key attributes set can improve the generalization performance of the original ZSL model which uses all the attributes. Unfortunately, previous attribute selection methods are conducted based on the seen data, their selected attributes have poor generalization capability to the unseen data, which is unavailable in training stage for ZSL tasks. Inspired by learning from pseudo relevance feedback, this paper introduces the out-of-the-box data, which is pseudo data generated by an attribute-guided generative model, to mimic the unseen data. After that, we present an iterative attribute selection (IAS) strategy which iteratively selects key attributes based on the out-of-the-box data. Since the distribution of the generated out-of-the-box data is similar to the test data, the key attributes selected by IAS can be effectively generalized to test data. Extensive experiments demonstrate that IAS can significantly improve existing attribute-based ZSL methods and achieve state-of-the-art performance. Copyright © 2019, The Authors. All rights reserved.

关键词： Iterative methods

Addressing survey fatigue bias in longitudinal social contact studies to improve pandemic preparedness

学校读者我要写书评

暂无评论

Scientific Reports 2025年第1期15卷 1-17页

作者： Dan, Shozen Ling, Zhi Chen, Yu Tegegne, Joshua Jaeger, Veronika K. Karch, André Mishra, Swapnil Ratmann, Oliver Vollmer, Sebastian Kraemer, Moritz Duchene, David A. Faria, Nuno Sejdinovic, Dino Unwin, Juliette Miscouridou, Xenia Semenova, Elizaveta Flaxman, Seth Bhatt, Samir Department of Mathematics Imperial College London London United Kingdom National University of Singapore Saw Swee Hock School of Public Health Singapore Singapore University of Münster Institute of Epidemiology and Social Medicine Münster Germany National University of Singapore Institute of Data Science Singapore Singapore Technical University of Kaiserslautern Kaiserslautern Germany German Research Center for Artificial Intelligence Kaiserslautern Germany Department of Statistics University of Warwick Coventry United Kingdom Department of Biology Univesity of Oxford Oxford United Kingdom Department of Public Health University of Copenhagen Copenhagen Denmark MRC Centre for Global Infectious Disease Analysis Imperial College London London United Kingdom School of Computer and Mathematical Sciences The University of Adelaide Adelaide Australia School of Mathematics University of Bristol Bristol United Kingdom Department of Mathematics and Statistics University of Cyprus Nicosia Cyprus Department of Epidemiology and Biostatistics Imperial College London London United Kingdom Department of Computer Science University of Oxford Oxford United Kingdom Department of Infectious Disease Epidemiology Imperial College London London United Kingdom School of Public Health University of Copenhagen Copenhagen Denmark

Social contact surveys are an important tool to assess infection risks within populations, and the effect of non-pharmaceutical interventions on social behaviour during disease outbreaks, epidemics, and pandemics. Numerous longitudinal social contact surveys were conducted during the COVID-19 era, however data analysis is plagued by survey fatigue, a phenomenon whereby the average number of social contacts reported declines with the number of repeat participations and as participants’ engagement decreases over time. Using data from the German COVIMOD Study between April 2020 to December 2021, we demonstrate that survey fatigue varied considerably by sociodemographic factors and was consistently strongest among parents reporting children contacts (parental proxy reporting), students, middle-aged individuals, those in full-time employment and those self-employed. We find further that, when using data from first-time participants as gold standard, statistical models incorporating a simple logistic function to control for survey fatigue were associated with substantially improved estimation accuracy relative to models with no survey fatigue adjustments, and that no cap on the number of repeat participations was required. These results indicate that existing longitudinal contact survey data can be meaningfully interpreted under an easy-to-implement statistical approach addressing survey fatigue confounding, and that longitudinal designs including repeat participants are a viable option for future social contact survey designs. © The Author(s) 2025.

关键词：

Unsupervised deep domain adaptation for hyperspectral image classification

学校读者我要写书评

暂无评论

Unsupervised deep domain adaptation for hyperspectral image ...

IEEE International Symposium on Geoscience and Remote Sensing (IGARSS)

作者： Wei Li Wei Wei Lei Zhang Cong Wang Yanning Zhang School of Computer Science Northwestern Polytechnical University Xi'an Shanxi China Inception Institute of Artificial Intelligence (IIAI) UAE The National Engineering Laboratory for Integrated Aero-Space-Ground-Ocean Big Data Application Technology Xi'an Shanxi China

Deep neural networks have been proven to be a promising way for hyperspectral image (HSI) classification. Their success depends on a premise that source domain (i.e., training) and target domain (i.e., test) samples are identically distributed. However, due to various imaging environments, in practice obvious distribution discrepancy often exists between these two domains, which can dramatically reduce the capacity of the classifier trained in source domain generalizing to target domain. To mitigate this problem, we present a novel deep unsupervised domain adaptation framework for HSI classification, which can simultaneously align the distributions of two domains and learn a classifier in source domain. Firstly, we employ two auto-encoder networks to separately project the samples from two domains into two low-dimensional feature spaces. Then, a multi-level maximum mean discrepancy (MMD) loss is imposed on the feature space to reduce the distribution discrepancy between two domains. Given the resultant features, a classification subnet is further learned to classify the labeled samples in source domain. Since the classifier is trained based on the domain-invariant features, it can well generalize to the target domain. Experimental results on one benchmark cross-domain HSI datasets prove the superior performance of the proposed method.

关键词：

Tensor Decomposition for EEG Signals Retrieval

学校读者我要写书评

暂无评论

Tensor Decomposition for EEG Signals Retrieval

IEEE International Conference on Systems, Man and Cybernetics

作者： Zehong Cao Yu-Cheng Chang Mukesh Prasad M. Tanveer Chin-Teng Lin School of Technology Environments and Design College of Sciences and Engineering University of Tasmania TAS Australia Centre for Artificial Intelligence and School of Computer Science Faculty of Engineering and Information Technology University of Technology Sydney NSW Australia Discipline of Mathematics Indian Institute of Technology Indore India

Prior studies have proposed methods to recover multi-channel electroencephalography (EEG) signal ensembles from their partially sampled entries. These methods depend on spatial scenarios, yet few approaches aiming to a temporal reconstruction with lower loss. The goal of this study is to retrieve the temporal EEG signals independently which was overlooked in data pre-processing. We considered EEG signals are impinging on tensor-based approach, named nonlinear Canonical Polyadic Decomposition (CPD). In this study, we collected EEG signals during a resting-state task. Then, we defined that the source signals are original EEG signals and the generated tensor is perturbed by Gaussian noise with a signal-to-noise ratio of 0 dB. The sources are separated using a basic nonnegative CPD and the relative errors on the estimates of the factor matrices. Comparing the similarities between the source signals and their recovered versions, the results showed significantly high correlation over 95%. Our findings reveal the possibility of recoverable temporal signals in EEG applications.

关键词： Electroencephalography Sensors Correlation Compressed sensing Gaussian noise Australia

Learning to adapt invariance in memory for person re-identification

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Zhong, Zhun Zheng, Liang Luo, Zhiming Li, Shaozi Yang, Yi Department of Artificial Intelligence Xiamen University Xiamen361005 China Centre for Artificial Intelligence University of Technology Sydney UltimoNSW2007 Australia Research School of Computer Science Australian National University CanberraACT0200 Australia Center of Information and Communication Engineering Xiamen University Xiamen361005 China

This work considers the problem of unsupervised domain adaptation in person re-identification (re-ID), which aims to transfer knowledge from the source domain to the target domain. Existing methods are primary to reduce the inter-domain shift between the domains, which however usually overlook the relations among target samples. This paper investigates into the intra-domain variations of the target domain and proposes a novel adaptation framework w.r.t three types of underlying invariance, i.e., Exemplar-Invariance, Camera-Invariance, and Neighborhood-Invariance. Specifically, an exemplar memory is introduced to store features of samples, which can effectively and efficiently enforce the invariance constraints over the global dataset. We further present the Graph-based Positive Prediction (GPP) method to explore reliable neighbors for the target domain, which is built upon the memory and is trained on the source samples. Experiments demonstrate that 1) the three invariance properties are indispensable for effective domain adaptation, 2) the memory plays a key role in implementing invariance learning and improves the performance with limited extra computation cost, 3) GPP could facilitate the invariance learning and thus significantly improves the results, and 4) our approach produces new state-of-the-art adaptation accuracy on three re-ID large-scale benchmarks. Copyright © 2019, The Authors. All rights reserved.

关键词： Graphic methods

Adapting Stochastic Block Models to Power-Law Degree Distributions

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Qiao, Maoying Yu, Jun Bian, Wei Li, Qiang Tao, Dacheng School of Computer Science and Technology at Hangzhou Dianzi University Hangzhou China Centre for Artificial Intelligence in the Faculty of Engineering and Information Technology at the University of Technology Sydney School of Software in the Faculty of Engineering and Information Technology at the University of Technology Sydney Department of Computing at Hong Kong Polytechnic University UBTECH Sydney Artificial Intelligence Centre School of Information Technologies Faculty of Engineering and Information Technologies University of Sydney 6 Cleveland St DarlingtonNSW2008 Australia

Stochastic block models (SBMs) have been playing an important role in modeling clusters or community structures of network data. But, it is incapable of handling several complex features ubiquitously exhibited in real-world networks, one of which is the power-law degree characteristic. To this end, we propose a new variant of SBM, termed power-law degree SBM (PLD-SBM), by introducing degree decay variables to explicitly encode the varying degree distribution over all nodes. With an exponential prior, it is proved that PLD-SBM approximately preserves the scale-free feature in real networks. In addition, from the inference of variational E-Step, PLD-SBM is indeed to correct the bias inherited in SBM with the introduced degree decay factors. Furthermore, experiments conducted on both synthetic networks and two real-world datasets including Adolescent Health data and the political blogs network verify the effectiveness of the proposed model in terms of cluster prediction accuracies. Copyright © 2019, The Authors. All rights reserved.

关键词： Stochastic systems