Filters








4,962 Hits in 8.8 sec

Heavy tails with parameter adaptation in random projection based continuous EDA

Momodou L. Sanyang, Ata Kaban
2015 2015 IEEE Congress on Evolutionary Computation (CEC)  
Since we are going to have large scale multivariate Gaussian EDA of [12] by proposing M subspaces, M independent random projection matrices are to use random matrices with heavy tailed entries.  ...  Heavy tails with Parameter Adaptation in Random Projection based continuous EDA Momodou L.  ... 
doi:10.1109/cec.2015.7257140 dblp:conf/cec/SanyangK15 fatcat:2ohu3e3irbbevadbglnrlmxgi4

Robust probabilistic superposition and comparison of protein structures

Martin Mechelke, Michael Habeck
2010 BMC Bioinformatics  
We develop two iterative procedures, an Expectation Maximization algorithm and a Gibbs sampler, to estimate the local weights, the optimal superposition, and the parameters of the heavy-tailed distribution  ...  Conclusions: Heavy-tailed distributions are well-suited to describe large-scale conformational differences in protein structures.  ...  To estimate these distributions and an optimal superposition we employ a scale mixture representation of the heavy-tailed models.  ... 
doi:10.1186/1471-2105-11-363 pmid:20594332 pmcid:PMC2912885 fatcat:n2z7eoqavjfb7o5u7gqg56h5qe

Towards large scale continuous EDA

Ata Kaban, Jakramate Bootkrajang, Robert John Durrant
2013 Proceeding of the fifteenth annual conference on Genetic and evolutionary computation conference - GECCO '13  
Estimation of distribution algorithms (EDA) are a major branch of evolutionary algorithms (EA) with some unique advantages in principle.  ...  Our concept is to introduce an ensemble of random projections of the set of fittest search points to low dimensions as a basis for developing a new and generic divide-and-conquer methodology.  ...  Some authors employ heavy tail search distributions, for example [29] propose a univariate EDA (UMDAc) with Gaussian and Lévy search distribution for large scale EDA.  ... 
doi:10.1145/2463372.2463423 dblp:conf/gecco/KabanBD13 fatcat:oirpdoylxfh2fjvvj3agildkwq

Two approaches of using heavy tails in high dimensional EDA

Momodou L. Sanyang, Hanno Muehlbrandt, Ata Kaban
2014 2014 IEEE International Conference on Data Mining Workshop  
We consider the problem of high dimensional black-box optimisation via Estimation of Distribution Algorithms (EDA).  ...  To get around of the difficulty of multivariate model building in high dimensions we employ a recently proposed random projections (RP) ensemble based approach which we modify to get samples from a multivariate  ...  A recent approach with state of the art performance [13] introduces an ensemble of random projections (RP) to low dimensions.  ... 
doi:10.1109/icdmw.2014.184 dblp:conf/icdm/SanyangMK14 fatcat:az4n4e2tenei3g5ypr4fql2uby

On Universal Restart Strategies for Backtracking Search [chapter]

Huayue Wu, Peter van Beek
2007 Lecture Notes in Computer Science  
We consider the commonly occurring scenario where one is to solve an ensemble of instances using a backtracking algorithm and wish to learn a good restart strategy for the ensemble.  ...  Previous studies have shown that a technique called randomization and restarts can dramatically improve the performance of a backtracking algorithm on some instances.  ...  This work was made possible by the facilities of the Shared Hierarchical Academic Research Computing Network (SHARCNET).  ... 
doi:10.1007/978-3-540-74970-7_48 fatcat:pffay3izerbkzdi6nbus3hx6o4

Bayesian compressed sensing imaging using a Gaussian Scale Mixture

George Tzagkarakis, Panagiotis Tsakalides
2010 2010 IEEE International Conference on Acoustics, Speech and Signal Processing  
coefficients should be sparse is enforced by fitting directly its prior probability distribution by means of a Gaussian Scale Mixture (GSM).  ...  The inherent property of compressed sensing (CS) theory working simultaneously as a sensing and compression protocol, using a small subset of random incoherent projection coefficients, enables a potentially  ...  In the present work, we did not make any assumption for the probability density function of the scaling factor A of the GSM. As a future work, we intend in posing a heavy-tailed distribution on A.  ... 
doi:10.1109/icassp.2010.5495397 dblp:conf/icassp/TzagkarakisT10 fatcat:rzxtamsyrzcmliise4qaybks7y

Dynamic Metric Learning from Pairwise Comparisons [article]

Kristjan Greenewald and Stephen Kelley and Alfred Hero III
2016 arXiv   pre-print
Specifically, we create a retro-initialized composite objective mirror descent (COMID) ensemble (RICE) consisting of a set of parallel COMID learners with different learning rates, demonstrate RICE-OCELAD  ...  We apply the OCELAD framework to an ensemble of online learners.  ...  New learner at scale j > 0: initialize to the last estimate of learner at scale j − 1.  ... 
arXiv:1610.03090v1 fatcat:mubd56vqvrccvpftkl33hj4tyy

Forest-based methods and ensemble model output statistics for rainfall ensemble forecasting [article]

Maxime Taillardat, Philippe Naveau
2017 arXiv   pre-print
We present statistical post-processing methods based on Quantile Regression Forests (QRF) and Gradient Forests (GF) with a parametric extension for heavy-tailed distributions.  ...  Our goal is to improve ensemble quality for all types of precipitation events, heavy-tailed included, subject to a good overall performance.  ...  Acknowledgments Part of the work of P. Naveau has been supported by the ANR-DADA, LEFE-INSU-Multirisk, AMERISKA, A2C2, CHAVANA and Extremoscope projects.  ... 
arXiv:1711.10937v1 fatcat:bhuq5hz4eveqhkacrye3znqgjy

Forest-based and semi-parametric methods for the postprocessing of rainfall ensemble forecasting

Maxime Taillardat, Anne-Laure Fougères, Philippe Naveau, Olivier Mestre
2019 Weather and forecasting  
We present statistical post-processing methods based on Quantile Regression Forests (QRF) and Gradient Forests (GF) with a parametric extension for heavy-tailed distributions.  ...  Our goal is to improve ensemble quality for all types of precipitation events, heavy-tailed included, subject to a good overall performance.  ...  Acknowledgments Part of the work of P. Naveau has been supported by the ANR-DADA, LEFE-INSU-Multirisk, AMERISKA, A2C2, CHAVANA and Extremoscope projects.  ... 
doi:10.1175/waf-d-18-0149.1 fatcat:5cta7h65rzamtp33m3cphkzfra

Compressive sensing signal reconstruction by weighted median regression estimates

Jose L. Paredes, Gonzalo R. Arce
2010 2010 IEEE International Conference on Acoustics, Speech and Signal Processing  
This stage is achieved by a hard threshold operator with adaptable thresholding parameter that is suitably tuned as the algorithm progresses.  ...  We compare the performance of the proposed approach to those yielded by state-of-the-art CS reconstruction algorithms showing that our approach achieves a better performance for different noise distributions  ...  heavy-tailed noise.  ... 
doi:10.1109/icassp.2010.5495738 dblp:conf/icassp/ParedesA10 fatcat:tfm4dhvxpzhq7lgvgtbxk5sb24

Network design for heavy rainfall analysis

T. Rietsch, P. Naveau, N. Gilardi, A. Guillou
2013 Journal of Geophysical Research - Atmospheres  
1] The analysis of heavy rainfall distributional properties is a complex object of study in hydrology and climatology, and it is essential for impact studies.  ...  We combine this theory with a machine learning algorithm based on neural networks and a Query By Committee approach.  ...  Part of this work has been supported by the EU-FP7 ACQWA project (www.acqwa.ch) under contract 212250, by the 13,085 RIETSCH ET AL.: NETWORK DESIGN FOR HEAVY RAINFALL PEPER-GIS-ADEME project, by the ANR  ... 
doi:10.1002/2013jd020867 fatcat:553cvp4525dvralvu3aeyi2vb4

A two-stage super learner for healthcare expenditures [article]

Ziyue Wu, Seth A. Berkowitz, Patrick J. Heagerty, David Benkeser
2021 medRxiv   pre-print
The method can flexibly incorporate a range of individual estimation approaches for each stage of estimation, including both regression-based approaches and machine learning algorithms such as random forests  ...  The super learner is an ensemble machine learning approach that can combine several algorithms to improve estimation.  ...  of 4 spine-related RVUs are all highly skewed with heavy upper tails despite the difference in scale.  ... 
doi:10.1101/2021.03.26.21254428 fatcat:n4ozytnj7vfkjet5tp5aavlpee

Cauchy Markov Random Field Priors for Bayesian Inversion [article]

Neil K. Chada, Lassi Roininen, Jarkko Suuronen
2022 arXiv   pre-print
The use of Cauchy Markov random field priors in statistical inverse problems can potentially lead to posterior distributions which are non-Gaussian, high-dimensional, multimodal and heavy-tailed.  ...  Thorough MCMC statistics are provided for all test cases, including potential scale reduction factors.  ...  JS and LR acknowledge Academy of Finland project funding (grant numbers 334816 and 336787). NKC is supported by KAUST baseline funding.  ... 
arXiv:2105.12488v2 fatcat:pmetbznplnhs7lgiwaprxxfyr4

Robust compressive sensing of sparse signals: a review

Rafael E. Carrillo, Ana B. Ramirez, Gonzalo R. Arce, Kenneth E. Barner, Brian M. Sadler
2016 EURASIP Journal on Advances in Signal Processing  
Among the scenarios in which robust performance is required, applications where the sampling process is performed in the presence of impulsive noise, i.e., measurements are corrupted by outliers, are of  ...  Compressive sensing generally relies on the 2 norm for data fidelity, whereas in many applications, robust estimators are needed.  ...  In Fig. 5 , the top shows the clean random measurements obtained with the DCT ensemble and the bottom shows the measurements contaminated with heavy-tailed noise.  ... 
doi:10.1186/s13634-016-0404-5 fatcat:5guvueoul5buhl7h2k7jq3zh64

Efficient variational inference in large-scale Bayesian compressed sensing

George Papandreou, Alan L. Yuille
2011 2011 IEEE International Conference on Computer Vision Workshops (ICCV Workshops)  
large-scale problems with the same memory and time complexity requirements as conventional point estimation techniques.  ...  We study linear models under heavy-tailed priors from a probabilistic viewpoint.  ...  Office of Naval Research under the MURI grant N000141010933 and by the Korean Ministry of Education, Science, and Technology, under the National Research Foundation WCU program R31-10008. We thank S.  ... 
doi:10.1109/iccvw.2011.6130406 dblp:conf/iccvw/PapandreouY11 fatcat:e57shqa2kzb23lwmusalxzyyr4
« Previous Showing results 1 — 15 out of 4,962 results