检索结果-内蒙古大学图书馆

Confidence intervals uncovered: Are we ready for real-world medical imaging AI?

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Christodoulou, Evangelia Reinke, Annika Houhou, Rola Kalinowski, Piotr Erkan, Selen Sudre, Carole H. Burgos, Ninon Boutaj, Sofiène Loizillon, Sophie Solal, Maëlys Rieke, Nicola Cheplygina, Veronika Antonelli, Michela Mayer, Leon D. Tizabi, Minu D. Jorge Cardoso, M. Simpson, Amber Jäger, Paul F. Kopp-Schneider, Annette Varoquaux, Gaël Colliot, Olivier Maier-Hein, Lena Heidelberg Div. Intelligent Medical Systems Germany AI Health Innovation Cluster Germany NCT Heidelberg a partnership between DKFZ Heidelberg University Hospital Germany DKFZ Heidelberg Helmholtz Imaging Germany HIDSS4Health - Helmholtz Information and Data Science School for Health Germany DKFZ Heidelberg Interactive Machine Learning Group Germany MRC Unit for Lifelong Health and Ageing UCL Centre for Medical Image Computing Department of Computer Science University College London United Kingdom School of Biomedical Engineering and Imaging Science King’s College London United Kingdom Sorbonne Université Institut du Cerveau - Paris Brain Institute - ICM CNRS Inria Inserm AP-HP Hôpital de la Pitié-Salpêtrière France NVIDIA Germany Department of Computer Science IT University of Copenhagen Denmark Centre for Medical Image Computing University College London United Kingdom School of Computing Queen’s University Canada Department of Biomedical and Molecular Sciences Queen’s University Canada Division of Biostatistics DKFZ Germany Parietal project team INRIA Saclay-Île de France France Faculty of Mathematics and Computer Science Heidelberg University Germany Medical Faculty Heidelberg University Germany

Medical imaging is spearheading the AI transformation of healthcare. Performance reporting is key to determine which methods should be translated into clinical practice. Frequently, broad conclusions are simply derived from mean performance values. In this paper, we argue that this common practice is often a misleading simplification as it ignores performance variability. Our contribution is threefold. (1) Analyzing all MICCAI segmentation papers (n = 221) published in 2023, we first observe that more than 50% of papers do not assess performance variability at all. Moreover, only one (0.5%) paper reported confidence intervals (CIs) for model performance. (2) To address the reporting bottleneck, we show that the unreported standard deviation (SD) in segmentation papers can be approximated by a second-order polynomial function of the mean Dice similarity coefficient (DSC). Based on external validation data from 56 previous MICCAI challenges, we demonstrate that this approximation can accurately reconstruct the CI of a method using information provided in publications. (3) Finally, we reconstructed 95% CIs around the mean DSC of MICCAI 2023 segmentation papers. The median CI width was 0.03 which is three times larger than the median performance gap between the first and second ranked method. For more than 60% of papers, the mean performance of the second-ranked method was within the CI of the first-ranked method. We conclude that current publications typically do not provide sufficient evidence to support which models could potentially be translated into clinical practice. © 2024, CC BY.

关键词： Medical imaging

Locally differentially private (contextual) bandits learning 20

学校读者我要写书评

暂无评论

Locally differentially private (contextual) bandits learning

Proceedings of the 34th International Conference on Neural Information Processing Systems

作者： Kai Zheng Tianle Cai Weiran Huang Zhenguo Li Liwei Wang Kwai Inc. School of Mathematical Sciences Peking University and Haihua Institute for Frontier Information Technology Huawei Noah's Ark Lab Key Laboratory of Machine Perception MOE School of EECS Peking University and Center for Data Science Peking University

ISBN: (纸本)9781713829546

We study locally differentially private (LDP) bandits learning in this paper. First, we propose simple black-box reduction frameworks that can solve a large family of context-free bandits learning problems with LDP guarantee. Based on our frameworks, we can improve previous best results for private bandits learning with one-point feedback, such as private Bandits Convex Optimization, and obtain the first result for Bandits Convex Optimization (BCO) with multi-point feedback under LDP. LDP guarantee and black-box nature make our frameworks more attractive in real applications compared with previous specifically designed and relatively weaker differentially private (DP) context-free bandits algorithms. Further, we extend our (ε, ∝)-LDP algorithm to Generalized Linear Bandits, which enjoys a sub-linear regret Õ(T3/4/ε) and is conjectured to be nearly optimal. Note that given the existing Ω(T) lower bound for DP contextual linear bandits [35], our result shows a fundamental difference between LDP and DP contextual bandits learning.

关键词：

Cosmological N-body simulations: a challenge for scalable generative models

学校读者我要写书评

暂无评论

Computational Astrophysics and Cosmology 2019年第1期6卷 1-17页

作者： Perraudin, Nathanaël Srivastava, Ankit Lucchi, Aurelien Kacprzak, Tomasz Hofmann, Thomas Réfrégier, Alexandre Swiss Data Science Center ETH Zurich Zurich Switzerland Institute for Machine Learning ETH Zurich Zurich Switzerland Institute for Particle Physics and Astrophysics ETH Zurich Zurich Switzerland

Deep generative models, such as Generative Adversarial Networks (GANs) or Variational Autoencoders (VAs) have been demonstrated to produce images of high visual quality. However, the existing hardware on which these models are trained severely limits the size of the images that can be generated. The rapid growth of high dimensional data in many fields of science therefore poses a significant challenge for generative models. In cosmology, the large-scale, three-dimensional matter distribution, modeled with N-body simulations, plays a crucial role in understanding the evolution of structures in the universe. As these simulations are computationally very expensive, GANs have recently generated interest as a possible method to emulate these datasets, but they have been, so far, mostly limited to two dimensional data. In this work, we introduce a new benchmark for the generation of three dimensional N-body simulations, in order to stimulate new ideas in the machine learning community and move closer to the practical use of generative models in cosmology. As a first benchmark result, we propose a scalable GAN approach for training a generator of N-body three-dimensional cubes. Our technique relies on two key building blocks, (i) splitting the generation of the high-dimensional data into smaller parts, and (ii) using a multi-scale approach that efficiently captures global image features that might otherwise be lost in the splitting process. We evaluate the performance of our model for the generation of N-body samples using various statistical measures commonly used in cosmology. Our results show that the proposed model produces samples of high visual quality, although the statistical analysis reveals that capturing rare features in the data poses significant problems for the generative models. We make the data, quality evaluation routines, and the proposed GAN architecture publicly available at https://***/nperraud/3DcosmoGAN.

关键词：

On the Utility Function of Experiments in Fundamental science

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Dorigo, Tommaso Doro, Michele Aehle, Max Gauger, Nicolas R. Awais, Muhammad Izbicki, Rafael Kieseler, Jan Masserano, Luca Nardi, Federico Vergara, Luis Recabarren Luleå University of Technology Laboratorievägen 14 Luleå97187 Sweden INFN - Sezione di Padova via F. Marzolo 8 Padova35131 Italy Universal Scientific Education and Research Network Italy Università di Padova Dipartimento di Fisica e Astronomia "G.Galilei" via F. Marzolo 8 Padova35131 Italy Gottlieb-Daimler-Strase Kaiserslautern67663 Germany Karlsruhe Institute for Technology Kaiserstrase 12 Karlsruhe76131 Germany Laboratoire de Physique de Clermont Auvergne 4 Avenue Blaise Pascal Aubière63170 France Centro di Ateneo di Studi e Attività Spaziali "Giuseppe Colombo" Via Venezia 15 PadovaI-35131 Italy Department of Statistics & Data Science Department of Machine Learning Carnegie Mellon University Pittsburgh United States Department of Statistics Federal University of São Carlos São Carlos Brazil

The majority of experiments in fundamental science today are designed to be multi-purpose: their aim is not simply to measure a single physical quantity or process, but rather to enable increased precision in the measurement of a number of different observable quantities of a natural system, to extend the search for new phenomena, or to exclude a larger phase space of candidate theories. Most of the time, a combination of the above goals is pursued;this breadth of scope adds a layer of complexity to the already demanding task of designing the measurement apparatus in an optimal way, by defining suitable geometries and choosing the most advantageous materials and appropriate detection technologies. The precise definition of a global optimality criterion may then require experimentalists to find a consensus on the relative scientific worth of those goals. In this work, we discuss the problem of formulating a utility function for multipurpose experiments, as an enabling step to employ artificial intelligence tools to explore the design space and assist humans in finding solutions at the Pareto front. Copyright © 2025, The Authors. All rights reserved.

关键词： Phase space methods

Explaining bayesian neural networks

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Bykov, Kirill Höhne, Marina M.-C. Creosteanu, Adelaida Müller, Klaus-Robert Machine Learning Group Technische Universität Berlin Marchstr. 23 Berlin10587 Germany Department of Artificial Intelligence Korea University Anam-dong Seongbuk-gu Seoul02841 Korea Republic of Max Planck Institute for Informatics Stuhlsatzenhausweg 4 Saarbrücken66123 Germany BIFOLD - Berlin Institute for the Foundations of Learning and Data Technische Universität Berlin Berlin Germany Google Research Brain team Berlin Germany Department of Computer Science TU Kaiserslautern Germany Heidelberg Germany Institute of Pathology Charite – Universitätsmedizin Berlin Berlin Germany Aignostics Berlin Germany RIKEN AIP 1-4-1 Nihonbashi Chuo-ku Tokyo Japan

—To make advanced learning machines such as Deep Neural Networks (DNNs) more transparent in decision making, explainable AI (XAI) aims to provide interpretations of DNNs’ predictions. These interpretations are usually given in the form of heatmaps, each one illustrating relevant patterns regarding the prediction for a given instance. Bayesian approaches such as Bayesian Neural Networks (BNNs) so far have a limited form of transparency (model transparency) already built-in through their prior weight distribution, but notably, they lack explanations of their predictions for given instances. In this work, we bring together these two perspectives of transparency into a holistic explanation framework for explaining BNNs. Within the Bayesian framework, the network weights follow a probability distribution. Hence, the standard (deterministic) prediction strategy of DNNs extends in BNNs to a predictive distribution, and thus the standard explanation extends to an explanation distribution. Exploiting this view, we uncover that BNNs implicitly employ multiple heterogeneous prediction strategies. While some of these are inherited from standard DNNs, others are revealed to us by considering the inherent uncertainty in BNNs. Our quantitative and qualitative experiments on toy/benchmark data and real-world data from pathology show that the proposed approach of explaining BNNs can lead to more effective and insightful explanations. Copyright © 2021, The Authors. All rights reserved.

关键词： Deep neural networks

Drone flight data reveal energy and greenhouse gas emissions savings for small package delivery

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Rodrigues, Thiago A. Patrikar, Jay Oliveira, Natalia L. Scott Matthews, H. Scherer, Sebastian Samaras, Constantine Department of Civil and Environmental Engineering Carnegie Mellon University 5000 Forbes Avenue PittsburghPA15213 United States Robotics Institute Carnegie Mellon University 5000 Forbes Avenue Pittsburgh15213 United States Department of Statistics and Data Science Carnegie Mellon University 5000 Forbes Avenue Pittsburgh15213 United States Machine Learning Department Carnegie Mellon University 5000 Forbes Avenue Pittsburgh15213 United States

The adoption of Uncrewed Aerial Vehicles (UAVs) for last-mile deliveries will affect the energy productivity of package delivery and require new methods to understand the associated energy consumption and greenhouse gas (GHG) emissions. Here we combine empirical testing of 187 quadcopter flights with first principles analysis to develop a usable energy model for drone package delivery. We develop a machine learning algorithm to assess energy use across three different flight regimes: takeoff, cruise, and landing. Our model shows that, in the US, a small electric quadcopter drone with a payload of 1 kg would consume approximately 0.05 MJ/km and result in 41 g of CO2e per package. The energy per package delivered by drones (0.19 MJ/package) can be up to 96% lower than conventional transportation modes. Our open model and generalizable coefficients can assist stakeholders in understanding and improving the energy use of drone package delivery. © 2021, CC BY-NC-SA.

关键词： Energy utilization

Universal patterns of intra-urban morphology: Defining a global typology of the urban fabric using unsupervised clustering

学校读者我要写书评

暂无评论

International Journal of Applied Earth Observation and Geoinformation 2025年 141卷

作者： Henri Debray Matthias Gassilloud Richard Lemoine-Rodríguez Michael Wurm Xiaoxiang Zhu Hannes Taubenböck Institute of Geography and Geology Department of Global Urbanization and Remote Sensing University of Würzburg 97074 Würzburg Germany German Remote Sensing Data Center (DFD) German Aerospace Center (DLR) 82234 Oberpfaffenhofen Germany Data Science in Earth Observation Technical University of Munich (TUM) 80333 Munich Germany Chair of Remote Sensing and Landscape Information Systems University of Freiburg 79106 Freiburg Germany Geolingual Studies Team University of Würzburg Am Hubland 97074 Würzburg Germany Munich Center for Machine Learning 80333 Munich Germany

The physical dimension of cities and its spatial patterns play a crucial role in shaping society and urban dynamics. Understanding the complexity of urban systems requires a detailed assessment of their physical structure. Urban geography has long focused on framing typologies to represent common patterns in the urban fabric using various methodologies. However, only recent advancements in computational methods and global land cover data have enabled to comprehensively identify typologies of urban patterns at the city scale through new unsupervised approaches. Nevertheless, typologies of finer-grained patterns at intra-urban scale have not yet been explored comprehensively at a global level. In this paper, building upon these advances, we explore the intra-urban patterns of more than 1500 cities across the globe. We rely on a Local Climate Zone land cover classification to represent the multidimensional variabilities of intra-urban morphology. Adapting a deep learning based unsupervised clustering approach, we find a typology of 138 intra-urban patterns. Analyzing the results of this data-driven approach, we prove that each pattern identified is unique, i.e. statistically different, in its composition and configuration. With this study summarizing the global diversity of the urban fabric, we reveal that any city of the world can be described as a specific assemblage of a fraction of these 138 universal patterns. These universal patterns reveal a predominance at a global scale of built-up forms of low density in the intra-urban fabric.

关键词：

BIGDML: Towards exact machine learning force fields for materials

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Sauceda, Huziel E. Gálvez-González, Luis E. Chmiela, Stefan Paz-Borbón, Lauro Oliver Müller, Klaus-Robert Tkatchenko, Alexandre Machine Learning Group Technische Universität Berlin Berlin10587 Germany BASLEARN TU Berlin BASF Joint Lab for Machine Learning Technische Universität Berlin Berlin10587 Germany División de Ciencias Exactas y Naturales Universidad de Sonora Blvd. Luis Encinas & Rosales Hermosillo Mexico BIFOLD – Berlin Institute for the Foundations of Learning and Data Germany Instituto de Física Universidad Nacional Autónoma de México Apartado Postal 20-364 CDMX01000 Mexico Google Research Brain team Berlin Germany Department of Artificial Intelligence Korea University Anam-dong Seongbuk-gu Seoul02841 Korea Republic of Max Planck Institute for Informatics Stuhlsatzenhausweg Saarbrücken66123 Germany Department of Physics and Materials Science University of Luxembourg LuxembourgL-1511 Luxembourg

machine-learning force fields (MLFF) should be accurate, computationally and data efficient, and applicable to molecules, materials, and interfaces thereof. Currently, MLFFs often introduce tradeoffs that restrict their practical applicability to small subsets of chemical space or require exhaustive datasets for training. Here, we introduce the Bravais-Inspired Gradient-Domain machine learning (BIGDML) approach and demonstrate its ability to construct reliable force fields using a training set with just 10-200 geometries for materials including pristine and defect-containing 2D and 3D semiconductors and metals, as well as chemisorbed and physisorbed atomic and molecular adsorbates on surfaces. The BIGDML model employs the full relevant symmetry group for a given material, does not assume artificial atom types or localization of atomic interactions and exhibits high data efficiency and state-of-the-art energy accuracies (errors substantially below 1 meV per atom) for an extended set of materials. Extensive path-integral molecular dynamics carried out with BIGDML models demonstrate the counterintuitive localization of benzene–graphene dynamics induced by nuclear quantum effects and allow to rationalize the Arrhenius behavior of hydrogen diffusion coefficient in a Pd crystal for a wide range of temperatures. Copyright © 2021, The Authors. All rights reserved.

关键词： machine learning

Adapting to noise distribution shifts in flow-based gravitational-wave inference

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Wildberger, Jonas Dax, Maximilian Green, Stephen R. Gair, Jonathan Pürrer, Michael Macke, Jakob H. Buonanno, Alessandra Schölkopf, Bernhard Max Planck Institute for Intelligent Systems Max-Planck-Ring 4 Tübingen72076 Germany School of Mathematical Sciences University of Nottingham University Park NottinghamNG7 2RD United Kingdom Am Mühlenberg 1 Potsdam14476 Germany Department of Physics East Hall University of Rhode Island KingstonRI02881 United States URI Research Computing Tyler Hall University of Rhode Island KingstonRI02881 United States Machine Learning in Science University of Tübingen Tübingen72076 Germany Department of Physics University of Maryland College ParkMD20742 United States

Deep learning techniques for gravitational-wave parameter estimation have emerged as a fast alternative to standard samplers—producing results of comparable accuracy. These approaches (e.g., Dingo) enable amortized inference by training a normalizing flow to represent the Bayesian posterior conditional on observed data. By conditioning also on the noise power spectral density (PSD) they can even account for changing detector characteristics. However, training such networks requires knowing in advance the distribution of PSDs expected to be observed, and therefore can only take place once all data to be analyzed have been gathered. Here, we develop a probabilistic model to forecast future PSDs, greatly increasing the temporal scope of Dingo networks. Using PSDs from the second LIGO-Virgo observing run (O2)—plus just a single PSD from the beginning of the third (O3)—we show that we can train a Dingo network to perform accurate inference throughout O3 (on 37 real events). We therefore expect this approach to be a key component to enable the use of deep learning techniques for low-latency analyses of gravitational waves. Copyright © 2022, The Authors. All rights reserved.

关键词： Gravity waves