检索结果-内蒙古大学图书馆

Denoising dictionary learning against adversarial perturbations

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Mitro, John Bridge, Derek Prestwich, Steven Insight Centre for Data Analytics Department of Computer Science University College Cork Ireland

We propose denoising dictionary learning (DDL), a simple yet effective technique as a protection measure against adversarial *** examined denoising dictionary learning on MNIST and CIFAR10 perturbed under two different perturbation techniques, fast gradient sign (FGSM) and jacobian saliency maps (JSMA). We evaluated it against five different deep neural networks (DNN) representing the building blocks of most recent architectures indicating a successive progression of model complexity of each *** show that each model tends to capture different representations based on their architecture. For each model we recorded its accuracy both on the perturbed test data previously misclassified with high confidence and on the denoised one after the reconstruction using dictionary learning. The reconstruction quality of each data point is assessed by means of PSNR (Peak Signal to Noise Ratio) and Structure Similarity Index (SSI). We show that after applying (DDL) the reconstruction of the original data point from a noisy sample results in a correct prediction with high confidence. Copyright © 2018, The Authors. All rights reserved.

关键词： Deep neural networks

Forecasting Thailand’s Precipitation with Cascading Model of CNN and GRU

学校读者我要写书评

暂无评论

Forecasting Thailand’s Precipitation with Cascading Model o...

International Conference on Information Technology and Electrical Engineering (ICITEE)

作者： Fuenglada Manokij Kanoksri Sarinnapakorn Peerapon Vateekul Department of Computer Engineering Faculty of Engineering Chulalongkorn University Big Data Analytics and IoT Center (CUBIC) Chulalongkorn University Bangkok Thailand Hydro-Informatics Institute (Public Organization) Ministry of Higher Education Science Research and Innovation Bangkok Thailand

Precipitation prediction is necessary to use in water management, especially in Thailand, it can be applied for various ways, such as flood warning, agriculture planning, etc. There are many prior attempts to forecast rainfall from the rain-gauge station. Some deployed traditional machine learning approaches: ARIMA, k-NN, etc. Recently, deep learning approach has shown promising results in this task. However, the accuracy is still limited since the raining period throughout the year in Thailand is very scarce, so most rainfall amount is zero. In this paper, we propose to cascade two deep learning networks to tackle this problem: one is a classification model to classify whether it rains or not, and the other is a regression model to predict rainfall amount. Our network is a combination of CNN and GRU, where CNN aims to capture relationship between various sensors and GRU aims to capture time-series information. Furthermore, we also perform multi-step forecasting by applying a rolling mechanism that uses the predicted rainfall and involved features to predict the next 6 steps. The experiment was conducted on hourly rainfall dataset for 6 years (2013-2018) provided from the public government sector in Thailand. We use RMSE as performance metric to evaluate three periods of rainfall: Overall, Rain, and non-rain periods and the results show that our cascading model is the winner with only 4.53% in term of RMSE which is the average percentage of difference from over all regions.

关键词： Deep learning Rain Predictive models data models Sensors Forecasting Task analysis

Inverse-dirichlet weighting enables reliable training of physics informed neural networks

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Maddu, Suryanarayana Sturm, Dominik Müller, Christian L. Sbalzarini, Ivo F. Technische Universität Dresden Faculty of Computer Science Dresden01069 Germany Max Planck Institute of Molecular Cell Biology and Genetics Dresden01307 Germany Center for Systems Biology Dresden Dresden01307 Germany Center for Computational Mathematics Flatiron Institute New YorkNY United States Department of Statistics LMU München Munich Germany Institute of Computational Biology Helmholtz Zentrum München Germany Center for Scalable Data Analytics and Artificial Intelligence ScaDS.AI Dresden/Leipzig Germany Cluster of Excellence Physics of Life TU Dresden Germany Helmholtz-Zentrum Dresden-Rossendorf DresdenD-01328 Germany GörlitzD-02826 Germany

We characterize and remedy a failure mode that may arise from multi-scale dynamics with scale imbalances during training of deep neural networks, such as Physics Informed Neural Networks (PINNs). PINNs are popular machine-learning templates that allow for seamless integration of physical equation models with data. Their training amounts to solving an optimization problem over a weighted sum of data-fidelity and equation-fidelity objectives. Conflicts between objectives can arise from scale imbalances, heteroscedasticity in the data, stiffness of the physical equation, or from catastrophic interference during sequential training. We explain the training pathology arising from this and propose a simple yet effective inverse-Dirichlet weighting strategy to alleviate the issue. We compare with Sobolev training of neural networks, providing the baseline of analytically ǫ-optimal training. We demonstrate the effectiveness of inverse-Dirichlet weighting in various applications, including a multi-scale model of active turbulence, where we show orders of magnitude improvement in accuracy and convergence over conventional PINN training. For inverse modeling using sequential training, we find that inverse-Dirichlet weighting protects a PINN against catastrophic forgetting. © 2021, CC BY-SA.

关键词： Inverse problems

Explainable Sentiment Analysis with Applications in Medicine

学校读者我要写书评

暂无评论

Explainable Sentiment Analysis with Applications in Medicine

2018 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2018

作者： Zucco, Chiara Liang, Huizhi Fatta, Giuseppe Di Cannataro, Mario Data Analytics Research Center Department of Medical and Surgical Sciences University 'Magna Græcia' Viale Europa Catanzaro88100 Italy Department of Computer Science University of Reading ReadingRG6 6AY United Kingdom

ISBN: (纸本)9781538654880

Sentiment Analysis can help to extract knowledge related to opinions and emotions from user generated text information. It can be applied in medical field for patients monitoring purposes. With the availability of large datasets, deep learning algorithms have become a state of the art also for sentiment analysis. However, deep models have the drawback of not being non human-interpretable, raising various problems related to model's interpretability. Very few work have been proposed to build models that explain their decision making process and actions. In this work, we review the current sentiment analysis approaches and existing explainable systems. Moreover, we present a critical review of explainable sentiment analysis models and discussed the insight of applying explainable sentiment analysis in the medical field. © 2018 IEEE.

关键词： Decision making

Automatic Extraction of Medication Mentions from Tweets-Overview of the BioCreative VII Shared Task 3 Competition

学校读者我要写书评

暂无评论

database : the journal of biological databases and curation 2023年第2023期2023卷 baac108页

作者： Davy Weissenbacher Karen O'Connor Siddharth Rawal Yu Zhang Richard Tzong-Han Tsai Timothy Miller Dongfang Xu Carol Anderson Bo Liu Qing Han Jinfeng Zhang Igor Kulev Berkay Köprü Raul Rodriguez-Esteban Elif Ozkirimli Ammer Ayach Roland Roller Stephen Piccolo Peijin Han V G Vinod Vydiswaran Ramya Tekumalla Juan M Banda Parsa Bagherzadeh Sabine Bergler João F Silva Tiago Almeida Paloma Martinez Renzo Rivera-Zavala Chen-Kai Wang Hong-Jie Dai Luis Alberto Robles Hernandez Graciela Gonzalez-Hernandez Department of Computational Biomedicine Cedars-Sinai Medical Center Los Angeles CA USA. DBEI The Perelman School of Medicine University of Pennsylvania Philadelphia PA USA. Department of Computer Science and Information Engineering National Central University No. 300 Zhongda Rd Zhongli District Taoyuan 320 Taiwan. IoX Center National Taiwan University Da'an District Section 4 Roosevelt Rd No. 1 Barry Lam Hall Taipei 106 Taiwan. Research Center for Humanities and Social Sciences Academia Sinica No. 128 Section 2 Academia Rd Nangang District Taipei 115 Taiwan. Computational Health Informatics Program Boston Children's Hospital Boston MA USA. Department of Pediatrics Harvard Medical School Boston MA USA. NVIDIA Santa Clara CA USA. Department of Statistics Florida State University Tallahassee FL USA. Data and Analytics Chapter F. Hoffmann-La Roche Ltd Switzerland. Pharmaceutical Research and Early Development Roche Innovation Center Basel Switzerland. Speech and Language Technology Lab DFKI Berlin Germany. Department of Biology Brigham Young University Provo UT USA. Department of Computational Medicine and Bioinformatics Medical School University of Michigan Ann Arbor MI USA. Department of Learning Health Sciences Medical School University of Michigan Ann Arbor MI USA. School of Information University of Michigan Ann Arbor MI USA. Department of Computer Science Georgia State University Atlanta GA USA. CLaC Labs Concordia University Montreal Canada. DETI Institute of Electronics and Informatics Engineering of Aveiro University of Aveiro Portugal. Department of Computation University of A Coruña Spain. Computer Science and Engineering Department Universidad Carlos III de Madrid Madrid Spain. Big Data Laboratory Chunghwa Telecom Laboratories Taoyuan Taiwan. Department of Computer Science National Yang Ming Chiao Tung University Hsinchu Taiwan. Department of Electrical Engineering College of Electrical Engineering and Computer Sci

This study presents the outcomes of the shared task competition BioCreative VII (Task 3) focusing on the extraction of medication names from a Twitter user's publicly available tweets (the user's 'timeline'). In general, detecting health-related tweets is notoriously challenging for natural language processing tools. The main challenge, aside from the informality of the language used, is that people tweet about any and all topics, and most of their tweets are not related to health. Thus, finding those tweets in a user's timeline that mention specific health-related concepts such as medications requires addressing extreme imbalance. Task 3 called for detecting tweets in a user's timeline that mentions a medication name and, for each detected mention, extracting its span. The organizers made available a corpus consisting of 182 049 tweets publicly posted by 212 Twitter users with all medication mentions manually annotated. The corpus exhibits the natural distribution of positive tweets, with only 442 tweets (0.2%) mentioning a medication. This task was an opportunity for participants to evaluate methods that are robust to class imbalance beyond the simple lexical match. A total of 65 teams registered, and 16 teams submitted a system run. This study summarizes the corpus created by the organizers and the approaches taken by the participating teams for this challenge. The corpus is freely available at https://***/tasks/biocreative-vii/track-3/. The methods and the results of the competing systems are analyzed with a focus on the approaches taken for learning from class-imbalanced data.

关键词：

LSCP: Locally Selective Combination in Parallel Outlier Ensembles

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Zhao, Yue Nasrullah, Zain Hryniewicki, Maciej K. Li, Zheng Department of Computer Science University of Toronto Data Analytics PricewaterhouseCoopers Canada Toronto Campus Northeastern University

In unsupervised outlier ensembles, the absence of ground truth makes the combination of base outlier detectors a challenging task. Specifically, existing parallel outlier ensembles lack a reliable way of selecting competent base detectors, affecting accuracy and stability, during model combination. In this paper, we propose a framework—called Locally Selective Combination in Parallel Outlier Ensembles (LSCP)—which addresses the issue by defining a local region around a test instance using the consensus of its nearest neighbors in randomly selected feature subspaces. The top-performing base detectors in this local region are selected and combined as the model’s final output. Four variants of the LSCP framework are compared with seven widely used parallel frameworks. Experimental results demonstrate that one of these variants, LSCP AOM, consistently outperforms baselines on the majority of twenty real-world datasets. Copyright © 2018, The Authors. All rights reserved.

关键词： Statistics

Big-data clustering: K-Means or k-indicators?

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Chen, Feiyu Yang, Yuchen Xu, Liwei Zhang, Taiping Zhang, Yin School of Big Data and Software Engineering Chongqing University Chongqing China Department of Computational and Applied Mathematics Rice University HoustonTX United States School of Mathematical Sciences University of Electronic Science and Technology of China Chengdu Sichuan China College of Computer Science Chongqing University Chongqing China Institute for Data and Decision Analytics Chinese University of Hong Kong Shenzhen China

The K-means algorithm is arguably the most popular data clustering method, commonly applied to processed datasets in some "feature spaces", as is in spectral clustering. Highly sensitive to initializations, however, K-means encounters a scalability bottleneck with respect to the number of clusters K as this number grows in big data applications. In this work, we promote a closely related model called K-indicators model and construct an efficient, semi-convex-relaxation algorithm that requires no randomized initializations. We present extensive empirical results to show advantages of the new algorithm when K is large. In particular, using the new algorithm to start the K-means algorithm, without any replication, can significantly outperform the standard K-means with a large number of currently state-ofthe-A rt random replications. Copyright © 2019, The Authors. All rights reserved.

关键词： K-means clustering

Multi-omic Microbiome Profiles in the Female Reproductive Tract in Early Pregnancy

学校读者我要写书评

暂无评论

Infectious Microbes & Diseases 2019年第2期1卷 49-60页

作者： Sophonie Jean Bernice Huang Hardik I.Parikh David J.Edwards J.Paul Brooks Naren Gajenthra Kumar Nihar U.Sheth Vishal Koparde Ekaterina Smirnova Snehalata Huzurbazar Philippe H.Girerd Dayanjan S.Wijesinghe Jerome F.StraussIII Myrna G.Serrano Jennifer M.Fettweis Kimberly K.Jefferson Gregory A.Buck Department of Microbiology and Immunology Virginia Commonwealth UniversityRichmondVA Center for Microbiome Engineering and Data Analysis Virginia Commonwealth UniversityRichmondVA Department of Statistical Sciences and Operations Research Virginia Commonwealth UniversityRichmondVA Supply Chain Management and Analytics School of BusinessVirginia Commonwealth UniversityRichmondVA Life Sciences Virginia Commonwealth UniversityRichmondVA DNAnexus Inc. Mountain ViewCA Frederick National Laboratory for Cancer Research FrederickMD Department of Biostatistics Virginia Commonwealth UniversityRichmondVA Department of Statistics University of WyomingLaramieWY Obstetrics and Gynecology Virginia Commonwealth UniversityRichmondVA Pharmacotherapy and Outcomes Science Virginia Commonwealth UniversityRichmondVA Department of Computer Science Virginia Commonwealth UniversityRichmondVA.

The vaginal microbiome likely influences host signaling compounds within the reproductive tract,including pro-inflammatory signals,which may play an important role during *** lactobacilli are associated with positive pregnancy outcome,whereas bacterial vaginosis,a dysbiosis of the vaginal microbiome,is associated with an increased risk of adverse pregnancy outcomes including preterm *** the host response could be predicted based on the taxonomic composition of the vaginal microbiome,particularly early in pregnancy,then those predictions could potentially be used to personalize intervention methods to reduce preterm birth and other adverse *** this proof of principle study,we apply multivariate strategies to analyze 16S rRNA-based taxonomic surveys in conjunction with targeted immuno-proteomic and lipidomic data from vaginal samples from 58 women enrolled in the Multi-Omic Microbiome Study-Pregnancy Initiative during early *** between the vaginal microbiome and the vaginal lipidome have not been previously *** from this study reveal significant multiple pairwise associations between microbial taxa,specific eicosanoids and sphingomyelins,and *** the biologic significance of these associations is not yet known,these results support the utility of such multi-omic approaches as a means to predict the impact of the microbiome on the host.

关键词： vaginal microbiome preterm birth multi-omics metagenomics cytokine profiles lipidomics

JS08.6.A A connectivity signature for glioblastoma

学校读者我要写书评

暂无评论

Neuro-Oncology 2022年第SUPPLEMENT_2期24卷 ii8-ii9页

作者： Hoffmann, D C Hai, L Wagener, R Mandelbaum, H Xie, R Hausmann, D Dominguez Azorín, D Weil, S Sievers, P Cebulla, G Ito, J Warnken, U Venkataramani, V Ernst, K Reibold, D Will, R Suvà, M L Herold-Mende, C Sahm, F Winkler, F Schlesner, M Wick, W Kessler, T Clinical Cooperation Unit Neurooncology German Cancer Consortium (DKTK) German Cancer Research Center (DKFZ) Heidelberg Germany Department of Neurology and Neurooncology Program National Center for Tumor Diseases Heidelberg University Hospital Heidelberg Germany Faculty of Biosciences Heidelberg University Heidelberg Germany Bioinformatics and Omics Data Analytics DKFZ Heidelberg Germany Department of Neuropathology Institute of Pathology University Hospital Heidelberg Heidelberg Germany Clinical Cooperation Unit Neuropathology DKTK DKFZ Heidelberg Germany Department of Neuroanatomy Institute for Anatomy and Cell Biology Heidelberg University Heidelberg Germany Hopp Children's Cancer Center at the NCT Heidelberg (KiTZ) Heidelberg Germany Division of Pediatric Neurooncology DKTK DKFZ Heidelberg Germany Genomics and Proteomics Core Facility DKTK DKFZ Heidelberg Germany Broad Institute of Harvard and MIT Cambridge MA United States Department of Pathology and Center for Cancer Research Massachusetts General Hospital and Harvard Medical School Boston MA United States Department of Neurosurgery Heidelberg University Hospital Heidelberg Germany Biomedical Informatics Data Mining and Data Analytics Faculty of Applied Computer Science and Medical Faculty University of Augsburg Augsburg Germany