检索结果-内蒙古大学图书馆

SpectralDefense: Detecting Adversarial Attacks on CNNs in the Fourier Domain

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Harder, Paula Pfreundt, Franz-Josef Keuper, Margret Keuper, Janis Competence Center High Performance Computing Fraunhofer ITWM Kaiserslautern Germany Scientic Computing University of Kaiserslautern Kaiserlautern Germany Fraunhofer Center Machine Learning Germany Data and Web Science Group University of Mannheim Germany Offenburg University Germany

—Despite the success of convolutional neural networks (CNNs) in many computer vision and image analysis tasks, they remain vulnerable against so-called adversarial attacks: Small, crafted perturbations in the input images can lead to false predictions. A possible defense is to detect adversarial examples. In this work, we show how analysis in the Fourier domain of input images and feature maps can be used to distinguish benign test samples from adversarial images. We propose two novel detection methods: Our first method employs the magnitude spectrum of the input images to detect an adversarial attack. This simple and robust classifier can successfully detect adversarial perturbations of three commonly used attack methods. The second method builds upon the first and additionally extracts the phase of Fourier coefficients of feature-maps at different layers of the network. With this extension, we are able to improve adversarial detection rates compared to state-of-the-art detectors on five different attack methods. © 2021, CC BY.

关键词： Image classification

Robust and Fast Measure of Information via Low-rank Representation

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Dong, Yuxin Gong, Tieliang Yu, Shujian Chen, Hong Li, Chen School of Computer Science and Technology Xi’an Jiaotong University Xi’an710049 China Shaanxi Provincial Key Laboratory of Big Data Knowledge Engineering Ministry of Education Xi’an710049 China Machine Learning Group UiT - The Arctic University of Norway Norway College of Science Huazhong Agriculture University Wuhan430070 China Engineering Research Center of Intelligent Technology for Agriculture Ministry of Education Wuhan430070 China

The matrix-based Rényi’s entropy allows us to directly quantify information measures from given data, without explicit estimation of the underlying probability distribution. This intriguing property makes it widely applied in statistical inference and machine learning tasks. However, this information theoretical quantity is not robust against noise in the data, and is computationally prohibitive in large-scale applications. To address these issues, we propose a novel measure of information, termed low-rank matrix-based Rényi’s entropy, based on low-rank representations of infinitely divisible kernel matrices. The proposed entropy functional inherits the specialty of of the original definition to directly quantify information from data, but enjoys additional advantages including robustness and effective calculation. Specifically, our low-rank variant is more sensitive to informative perturbations induced by changes in underlying distributions, while being insensitive to uninformative ones caused by noises. Moreover, low-rank Rényi’s entropy can be efficiently approximated by random projection and Lanczos iteration techniques, reducing the overall complexity from O(n3) to O(n2s) or even O(ns2), where n is the number of data samples and s n. We conduct large-scale experiments to evaluate the effectiveness of this new information measure, demonstrating superior results compared to matrix-based Rényi’s entropy in terms of both performance and computational efficiency. Copyright © 2022, The Authors. All rights reserved.

关键词： Entropy

Occam's razor for AI: Coarse-graining Hammett Inspired Product Ansatz in Chemical Space

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Bragato, Marco Von Rudorff, Guido Falk Von Lilienfeld, O. Anatole University of Vienna Faculty of Physics Kolingasse 1416 WienAT1090 Austria University Kassel Department of Chemistry Heinrich-Plett-Str.40 Kassel34132 Germany Heinrich-Plett-Straße 40 Kassel34132 Germany Vector Institute for Artificial Intelligence TorontoONM5S 1M1 Canada Departments of Chemistry Materials Science and Engineering and Physics University of Toronto St. George Campus TorontoON Canada Machine Learning Group Technische Universität Berlin Germany Berlin Institute for the Foundations of Learning and Data Berlin10587 Germany

data-hungry machine learning methods have become a new standard to efficiently navigate chemical compound space for molecular and materials design and discovery. Due to the severe scarcity and cost of high-quality experimental or synthetic simulated training data, however, data-acquisition costs can be considerable. Relying on reasonably accurate approximate baseline labels with low computational complexity represents one of the most effective strategies to curb data-needs, e.g. through Δ-, transfer-, or multi-fidelity learning. We introduce baseline labels in the form of a generic coarse-graining Hammett Inspired Product (HIP) Ansatz which generalizes the empirical Hammett equation towards arbitrary systems and properties. Numerical evidence for the applicability of HIP includes solvation free energies of molecules, formation energies of quaternary elpasolite crystals, carbon adsorption energies on heterogeneous catalytic surfaces, HOMO-LUMO gaps of metallorganic complexes, and activation energies for SN2 reactions. After calibration on the same training sets, HIP yields an effective baseline for improved Δ-quantum machine learning models with superior data-efficiency when compared to previously introduced specialised domain-specific models. © 2023, CC0.

关键词： data acquisition

Adaptive 3D Localization of 2D Freehand Ultrasound Brain Images

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Yeung, Pak-Hei Aliasi, Moska Haak, Monique Xie, Weidi Namburete, Ana I.L. Oxford Machine Learning in NeuroImaging Lab Department of Computer Science University of Oxford Oxford United Kingdom Department of Engineering Science Institute of Biomedical Engineering University of Oxford Oxford United Kingdom Division of Fetal Medicine Department of Obstetrics Leiden University Medical Center Leiden2333 ZA Netherlands Shanghai Jiao Tong University Shanghai China Visual Geometry Group Department of Engineering Science University of Oxford Oxford United Kingdom

Two-dimensional (2D) freehand ultrasound is the mainstay in prenatal care and fetal growth monitoring. The task of matching corresponding cross-sectional planes in the 3D anatomy for a given 2D ultrasound brain scan is essential in freehand scanning, but challenging. We propose AdLocUI, a framework that Adaptively Localizes 2D Ultrasound Images in the 3D anatomical atlas without using any external tracking sensor. We first train a convolutional neural network with 2D slices sampled from co-aligned 3D ultrasound volumes to predict their locations in the 3D anatomical atlas. Next, we fine-tune it with 2D freehand ultrasound images using a novel unsupervised cycle consistency, which utilizes the fact that the overall displacement of a sequence of images in the 3D anatomical atlas is equal to the displacement from the first image to the last in that sequence. We demonstrate that AdLocUI can adapt to three different ultrasound datasets, acquired with different machines and protocols, and achieves significantly better localization accuracy than the baselines. AdLocUI can be used for sensorless 2D freehand ultrasound guidance by the bedside. The source code is available at https://***/pakheiyeung/AdLocUI. © 2022, CC BY.

关键词： Ultrasonics

Towards self-driving laboratories: The central role of density functional theory in the AI age

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Huang, Bing von Rudorff, Guido Falk Anatole von Lilienfeld, O. University of Vienna Faculty of Physics Kolingasse 14-16 WienAT1090 Austria University Kassel Department of Chemistry Heinrich-Plett-Str.40 Kassel34132 Germany Heinrich-Plett-Straße 40 Kassel34132 Germany Vector Institute for Artificial Intelligence TorontoONM5S 1M1 Canada Department of Chemistry Materials Science and Engineering and Physics University of Toronto St. George Campus TorontoON Canada Machine Learning Group Technische Universität Berlin Berlin Institute for the Foundations of Learning and Data Berlin10587 Germany

Density functional theory (DFT) plays a pivotal role for the chemical and materials science due to its relatively high predictive power, applicability, versatility and computational efficiency. We review recent progress in machine learning model developments which has relied heavily on density functional theory for synthetic data generation and for the design of model architectures. The general relevance of these developments is placed in some broader context for the chemical and materials sciences. Resulting in DFT based machine learning models with high efficiency, accuracy, scalability, and transferability (EAST), recent progress indicates probable ways for the routine use of successful experimental planning software within self-driving laboratories. © 2023, CC0.

关键词： Density functional theory

Latent space conditioning on generative adversarial networks

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Lopez, Ricard Durall Ho, Kalun Pfreundt, Franz-Josef Keuper, Janis Fraunhofer ITWM Germany IWR University of Heidelberg Germany Fraunhofer Center Machine Learning Germany Data and Web Science Group University of Mannheim Germany Institute for Machine Learning and Analytics Offenburg University Germany

Generative adversarial networks are the state of the art approach towards learned synthetic image generation. Although early successes were mostly unsupervised, bit by bit, this trend has been superseded by approaches based on labelled data. These supervised methods allow a much finer-grained control of the output image, offering more flexibility and stability. Nevertheless, the main drawback of such models is the necessity of annotated data. In this work, we introduce an novel framework that benefits from two popular learning techniques, adversarial training and representation learning, and takes a step towards unsupervised conditional GANs. In particular, our approach exploits the structure of a latent space (learned by the representation learning) and employs it to condition the generative model. In this way, we break the traditional dependency between condition and label, substituting the latter by unsupervised features coming from the latent space. Finally, we show that this new technique is able to produce samples on demand keeping the quality of its supervised counterpart. © 2020, CC BY.

关键词： Generative adversarial networks

Accurate machine Learned Quantum-Mechanical Force Fields for Biomolecular Simulations

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Unke, Oliver T. Stöhr, Martin Ganscha, Stefan Unterthiner, Thomas Maennel, Hartmut Kashubin, Sergii Ahlin, Daniel Gastegger, Michael Sandonas, Leonardo Medrano Tkatchenko, Alexandre Müller, Klaus-Robert Google Research Brain Team Machine Learning Group Technische Universität Berlin Berlin10587 Germany Technische Universität Berlin Berlin10623 Germany Department of Physics and Materials Science University of Luxembourg Luxembourg CityL-1511 Luxembourg BASLEARN TU Berlin Berlin10587 Germany BASF Joint Lab for Machine Learning Technische Universität Berlin Berlin10587 Germany Department of Artificial Intelligence Korea University Anam-dong Seongbuk-gu Seoul02841 Korea Republic of Max Planck Institute for Informatics Stuhlsatzenhausweg Saarbrücken66123 Germany BIFOLD Berlin Institute for the Foundations of Learning and Data Berlin Germany

Molecular dynamics (MD) simulations allow atomistic insights into chemical and biological processes. Accurate MD simulations require computationally demanding quantum-mechanical calculations, being practically limited to short timescales and few atoms. For larger systems, efficient, but much less reliable empirical force fields are used. Recently, machine learned force fields (MLFFs) emerged as an alternative means to execute MD simulations, offering similar accuracy as ab initio methods at orders-of-magnitude speedup. Until now, MLFFs mainly capture short-range interactions in small molecules or periodic materials, due to the increased complexity of constructing models and obtaining reliable reference data for large molecules, where long-ranged many-body effects become important. This work proposes a general approach to constructing accurate MLFFs for large-scale molecular simulations (GEMS) by training on "bottom-up" and "top-down" molecular fragments of varying size, from which the relevant physicochemical interactions can be learned. GEMS is applied to study the dynamics of alanine-based peptides and the 46-residue protein crambin in aqueous solution, allowing nanosecond-scale MD simulations of >25k atoms at essentially ab initio quality. Our findings suggest that structural motifs in peptides and proteins are more flexible than previously thought, indicating that simulations at ab initio accuracy might be necessary to understand dynamic biomolecular processes such as protein (mis)folding, drug–protein binding, or allosteric regulation. Copyright © 2022, The Authors. All rights reserved.

关键词： Molecular dynamics

Lung Ultrasound for the Detection of Pulmonary Tuberculosis Using Expert- and AI-Guided Interpretation: A Prospective Cohort Study

学校读者我要写书评

暂无评论

SSRN

SSRN 2025年

作者： Suttels, Véronique Brokowski, Trevor Wachinou, Ablo Prudence Wolleb, Julia Hada, Aboudou Rasisou Du Toit, Jacques Daniel Fiogbé, Arnauld Attannon Guendehou, Brice Alovokpinhou, Frederic Sefou, Fadyl Makpemikpa, Ginette Bessat, Cécile Roux, Alexia Garcia, Elena Brahier, Thomas Opota, Onya Doenz, Jonathan Vignoud, Julien Agodokpessi, Gildas Affolabi, Dissou Hartley, Mary-Anne Boillat-Blanco, Noémie Department of Medicine Infectious Diseases Lausanne University Hospital University of Lausanne Lausanne Switzerland Yale School of Medicine Department of Biomedical Informatics & Data Science New HavenCT06510 United States Cotonou Benin Faculty of Health Sciences University of the Witwatersrand Johannesburg South Africa Emergency Department Lausanne University Hospital University of Lausanne Lausanne1011 Switzerland Institute of Microbiology University of Lausanne University Hospital Centre Lausanne Switzerland Lausanne1015 Switzerland National Reference Laboratory for Mycobacteriology Cotonou Benin Yale University United States Faculty of Health Sciences Benin Intelligent Global Health Machine Learning and Optimization Laboratory

Background: Point-of-care lung ultrasound (LUS) is a promising tool for portable sputum-free tuberculosis (TB) triage. We investigate the diagnostic performance of LUS to detect TB using expert and artificial intelligence (AI) guided interpretation. We introduce ULTR-AI (Ultrasound-led TB recognition using AI), a suite of deep learning (DL) models designed to automate TB prediction from LUS images. Methods: In this prospective cohort study in a tertiary center in urban Benin, a standardized 14-point sliding scan LUS protocol was performed for symptomatic patients by a trained operator. LUS images were reviewed by two blinded and independent readers. Same-day single sputum Xpert MTB/RIF Ultra® was the microbiological reference standard. The suite comprises three AI models, ULTR-AI, ULTR-AI[signs] and ULTR-AI[max]. ULTR-AI predicts TB directly from images using DL, ULTR-AI[signs] first generates human-recognizable pathological signs before TB risk prediction in a machine learning (ML) model. Finally, ULTR-AI[max] takes the maximal TB risk score predicted by these two models. Findings: Out of 760 patients screened, 504 were analyzed. Median age was 40 (IQR 30-52), 196 (39%) were female, 78 (15%) were people with HIV and 66 (13%) had previous TB. Overall, 192 (38%) had bacteriologically confirmed TB. Human expert interpretation of lung ultrasound achieved a sensitivity of 0·90 (95%CI: 0·89, 0·92), specificity of 0·61 (95%CI: 0·54, 0·67) [AUROC 0·84, 95% CI: 0.82-0.85]. AI-guided interpretation with ULTRA-AI[max] reached a sensitivity of 0·91 (95% CI, 0·90–0·96) and a specificity of 0·85 (95% CI, 0·74–0·88) [AUROC 0·93, 95% 0·92-0·95]. Interpretation: In this cohort, AI-guided LUS meets the WHO requirements for a sputum-free TB triage test, enabling point-of-care testing with minimal infrastructure. ULTR-AI could further help decentralize TB diagnostics in LMICs, improving timely detection. Validation in diverse populations is crucial to confirm clinical utility. © 2025,

关键词： Lung cancer

Touchstone Benchmark: Are We on the Right Way for Evaluating AI Algorithms for Medical Segmentation?

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Bassi, Pedro R.A.S. Li, Wenxuan Tang, Yucheng Isensee, Fabian Wang, Zifu Chen, Jieneng Chou, Yu-Cheng Roy, Saikat Kirchhoff, Yannick Rokuss, Maximilian Huang, Ziyan Ye, Jin He, Junjun Wald, Tassilo Ulrich, Constantin Baumgartner, Michael Maier-Hein, Klaus H. Jaeger, Paul Ye, Yiwen Xie, Yutong Zhang, Jianpeng Chen, Ziyang Xia, Yong Xing, Zhaohu Zhu, Lei Sadegheih, Yousef Bozorgpour, Afshin Kumari, Pratibha Azad, Reza Merhof, Dorit Shi, Pengcheng Ma, Ting Du, Yuxin Bai, Fan Huang, Tiejun Zhao, Bo Wang, Haonan Li, Xiaomeng Gu, Hanxue Dong, Haoyu Yang, Jichen Mazurowski, Maciej A. Gupta, Saumya Wu, Linshan Zhuang, Jiaxin Chen, Hao Roth, Holger Xu, Daguang Blaschko, Matthew B. Decherchi, Sergio Cavalli, Andrea Yuille, Alan L. Zhou, Zongwei Department of Computer Science Johns Hopkins University United States Department of Pharmacy and Biotechnology University of Bologna Italy Center for Biomolecular Nanotechnologies Istituto Italiano di Tecnologia Italy NVIDIA United States Germany Germany ESAT-PSI KU Leuven Belgium Faculty of Mathematics and Computer Science Heidelberg University Germany HIDSS4Health - Helmholtz Information and Data Science School for Health Germany Shanghai Jiao Tong University China Shanghai Artificial Intelligence Laboratory China Pattern Analysis and Learning Group Department of Radiation Oncology Heidelberg University Hospital Germany DKFZ Germany School of Computer Science and Engineering Northwestern Polytechnical University China Australian Institute for Machine Learning The University of Adelaide Australia College of Computer Science and Technology Zhejiang University China Hong Kong University of Science and Technology Guangzhou China Hong Kong University of Science and Technology Hong Kong Faculty of Informatics and Data Science University of Regensburg Germany Faculty of Electrical Engineering and Information Technology RWTH Aachen University Germany Fraunhofer Institute for Digital Medicine MEVIS Germany Electronic & Information Engineering School Harbin Institute of Technology Shenzhen China China The Chinese University of Hong Kong Hong Kong Peking University China Department of Electrical and Computer Engineering Duke University United States Stony Brook University United States Department of Computer Science and Engineering Department of Chemical and Biological Engineering Division of Life Science Hong Kong University of Science and Technology Hong Kong Data Science and Computation Facility Fondazione Istituto Italiano di Tecnologia Italy Ecole Polytechnique Fédérale de Lausanne Switzerland

How can we test AI performance? This question seems trivial, but it isn’t. Standard benchmarks often have problems such as in-distribution and small-size test sets, oversimplified metrics, unfair comparisons, and short-term outcome pressure. As a consequence, good performance on standard benchmarks does not guarantee success in real-world scenarios. To address these problems, we present Touchstone, a large-scale collaborative segmentation benchmark of 9 types of abdominal organs. This benchmark is based on 5,195 training CT scans from 76 hospitals around the world and 5,903 testing CT scans from 11 additional hospitals. This diverse test set enhances the statistical significance of benchmark results and rigorously evaluates AI algorithms across out-of-distribution scenarios. We invited 14 inventors of 19 AI algorithms to train their algorithms, while our team, as a third party, independently evaluated these algorithms. In addition, we also evaluated pre-existing AI frameworks—which, differing from algorithms, are more flexible and can support different algorithms—including MONAI from NVIDIA, nnU-Net from DKFZ, and numerous other open-source frameworks. We are committed to expanding this benchmark to encourage more innovation of AI algorithms for the medical domain. Copyright © 2024, The Authors. All rights reserved.

关键词： Open systems