Filters








50 Hits in 8.2 sec

Oops I Took A Gradient: Scalable Sampling for Discrete Distributions [article]

Will Grathwohl, Kevin Swersky, Milad Hashemi, David Duvenaud, Chris J. Maddison
2021 arXiv   pre-print
We propose a general and scalable approximate sampling strategy for probabilistic models with discrete variables.  ...  Our approach uses gradients of the likelihood function with respect to its discrete inputs to propose updates in a Metropolis-Hastings sampler.  ...  A number of prior works have used gradient information for discrete sampling.  ... 
arXiv:2102.04509v2 fatcat:bnalwykj2nef5kgkc37j4jeuoa

Directly Training Joint Energy-Based Models for Conditional Synthesis and Calibrated Prediction of Multi-Attribute Data [article]

Jacob Kelly, Richard Zemel, Will Grathwohl
2021 arXiv   pre-print
We propose a simple extension which expands the capabilities of EBMs to generating accurate conditional samples.  ...  We build upon recent work and show that architectures for multi-attribute prediction can be reinterpreted as energy-based models (EBMs).  ...  Oops i took a gradient: Scalable sampling for discrete distributions, 2021. Nature, 585(7825):357-362, September 2020. doi: 10. 1038/s41586-020-2649-2.  ... 
arXiv:2108.04227v1 fatcat:dxezrndmkzhzhoeoxktks2xqfe

Dislocation Modeling Of Orientation Gradients In Phase-Transformed Tantalum Thin Films

Ari Kestenbaum
2016
I would also like to thank him for his suggestion to use a genetic algorithm for creating dislocation structures.  ...  I also want to thank him for giving me the opportunity to do research as an undergraduate and for giving me a real problem to work on.  ...  For example, crossover is being performed on two lists A and B, where A has i elements (A 1 , . . . , A i ), B has j elements (B 1 , . . . , B j ), and i < j.  ... 
doi:10.7298/x4h41pdb fatcat:6rten76lfvc6nipht2x7niampe

Inducing Gaussian Process Networks [article]

Alessandro Tibo, Thomas Dyhre Nielsen
2022 arXiv   pre-print
The inducing points, in particular, are learned directly in the feature space, enabling a seamless representation of complex structured domains while also facilitating scalable gradient-based learning  ...  In this paper, we propose inducing Gaussian process networks (IGN), a simple framework for simultaneously learning the feature space as well as the inducing points.  ...  end of spoilers big oops 3. positive -i rented the film i don't think it got a theatrical release here out expecting the worse the previews made the film look awful i ... be proclaimed 'the worst film  ... 
arXiv:2204.09889v1 fatcat:lprxh6xp6zagnpsubqtlq7xpre

Deep Reinforcement Learning, a textbook [article]

Aske Plaat
2022 arXiv   pre-print
The aim of this book is to provide a comprehensive overview of the field of deep reinforcement learning.  ...  The book is written for graduate students of artificial intelligence, and for researchers and practitioners who wish to better understand deep reinforcement learning methods and their challenges.  ...  For a single gradient update step gradient descent produces updated parameters 𝜃 𝑖 = 𝜃 − 𝛼∇ 𝜃 L T 𝑖 (𝜋 𝜃 ) speci c to task 𝑖.  ... 
arXiv:2201.02135v2 fatcat:3icsopexerfzxa3eblpu5oal64

Using machine learning to correct model error in data assimilation and forecast applications [article]

Alban Farchi and Patrick Laloyaux and Massimo Bonavita and Marc Bocquet
2021 arXiv   pre-print
Furthermore, using the hybrid surrogate models for DA yields a significantly better analysis than using the original model.  ...  This yields a class of iterative methods in which, at each iteration a DA step assimilates the observations, and alternates with a ML step to learn the underlying dynamics of the DA analysis.  ...  The computation has been performed using a Bi-Xeon Gold 6154 (18 cores) without a GPU card. The entire training phase took 31 s (wall-clock time) for a total of 2000 epochs.  ... 
arXiv:2010.12605v2 fatcat:ueh7us7lgbeppcum47suiewdim

D6.6: Report on petascale software libraries and programming models

Giovanni Erbacci, Carlo Cavazzoni, Filippo Spiga, Iris Christadler
2009 Zenodo  
provided in terms of performance and efficiency and benefit for parallelism.  ...  The work starts from an analysis of the applications mainly identified in D6.1 and D6.2.2, covering a broad range of scientific areas, and representative of the European HPC usage, to assess the state  ...  I took their CSR-kernel and wrote a C-function to fit the needs of mod2as.  ... 
doi:10.5281/zenodo.6546116 fatcat:ojachtibqfeorbsef6z2lpvbie

Book Reviews

1955 Journal of the American Statistical Association  
The gradient and conjugate gradient methods are presented, and a third method is given for the problem of matrix inversion.  ...  , and Significance Tests for Large Sample Sizes”; “The Chi-square Distribution”; “Correlation”; “Descriptive Sta- tistics”.  ... 
doi:10.1080/01621459.1955.10501305 fatcat:sx2fcjkoz5d7jcushxdt3wgnny

Abstracts Of The 16 Th International Conference On Electrical Bio-Impedance And 17 Th International Conference On Electrical Impedance Tomography [article]

Stig Ollmar, Farhad Abtahi, Björn-Erik Erlandsson, Birgitta Janerot-Sjöberg, Kaj Lindecrantz, Mannan Mridha, Ingrid Nicander, Fernand Seoane, Heikki Teriö
2016 Zenodo  
MethoDs Regional intratidal gas distribution (ITGD) index was calculated for the tidal volume distribution during inspiration [1] ITGD k =SUM ROI (I j,k -I j,0 ) / SUM Lung (I i,k -I i,0 ) * 100% where  ...  A fraction volume approach is demonstrated with explicit Euclidean gradient and minimization over the oblique manifold. we reconstruct the proportion distribution map for each supposed material.  ...  IMAGe MethoDs The CardioExcyte 96 system provides a non-invasive, label-free, high temporal resolution approach for safety screening on iPSCMs (1).  ... 
doi:10.5281/zenodo.55753 fatcat:fh3blon7kfhonbepjfllnwqjbq

International Research Conference on Smart Computing and Systems Engineering SCSE 2020 Proceedings [Full Conference Proceedings]

2020 2020 International Research Conference on Smart Computing and Systems Engineering (SCSE)  
ACKNOWLEDGMENT The authors would like to thank the Department of Census and Department of Irrigation, Sri Lanka for providing the paddy yield and climate data for this study.  ...  would like to extend their heartfelt gratitude and acknowledgment to all the language experts, Sinhala teachers from different schools and especially, the Sinhala Department in the University of Kelaniya for  ...  Smart contracts also bring a lot of scalability for the ticketing Ecosystem.  ... 
doi:10.1109/scse49731.2020.9313027 fatcat:gjk5az2mprgvrpallwh6uhvlfi

Solution of Large-scale Structured Optimization Problems with Schur-complement and Augmented Lagrangian Decomposition Methods

Jose S Rodriguez
2019
We implement an approach that uses a Krylov solver (GMRES) preconditioned with ADMM to solve block-structured linear systems that ar [...]  ...  These problems arise in a great number of engineering applications, including parameter estimation, optimal control, network optimization, and stochastic programming.  ...  i BIn addition to parallel linear solvers, scalable parallel interior-point algorithms require the parallelization of all NLP functions and gradient evaluations.  ... 
doi:10.25394/pgs.8210243.v1 fatcat:i5qjpy3upjc7jalnrw7m4eye5a

A peer-led intervention to promote sexual health in secondary schools: the STASH feasibility study

Kirstin R Mitchell, Carrie Purcell, Ross Forsyth, Sarah Barry, Rachael Hunter, Sharon A Simpson, Lisa McDaid, Lawrie Elliot, Mark McCann, Kirsty Wetherall, Chiara Broccatelli, Julia V Bailey (+1 others)
2020 Public Health Research  
The study also piloted indicative primary outcomes for a full-scale evaluation.  ...  Results A total of 104 students were trained as peer supporters (just over half of those nominated for the role by their peers). Role retention was very high (97%).  ...  Acknowledgement Peer supporters who fulfil the requirements of the role would be provided with University of Glasgow certificates and, if they complete the online questionnaire, a £10 gift voucher.  ... 
doi:10.3310/phr08150 fatcat:ijw4israkrfj5jebcw225gmbdq

An Integrated Scalable Lighting Simulation Tool

Yi Chun Huang
2018
The effort and time reduction goal is structured as the following objectives in a new l [...]  ...  Case studies of actual design scenarios are used to establish quantitatively the effort and time costs baselines for comparison.  ...  in between the samples interpolated if the irradiance gradient between adjacent samples is below a specified threshold.  ... 
doi:10.1184/r1/6714716.v1 fatcat:vc4dvzfr4ze75o4ag6ufwrvkum

World No Tobacco Day reminds us of the desolating effects of smoking on health

Stephen Leeder
2015 Medical Journal of Australia  
ACKNOWLEDGEMENTS We are delighted to offer this preface to the second edition of A race against  ...  D4 Out-of-pocket (OOP) expenditures for CVD as a proportion of overall OOP health expenditures, 2007(29). PloS One 2011; 6: e20821. 12. Kiiskinen U, Vartiainen E, Pekurinen M, Puska P.  ...  A comprehensive primary care service depends on a distribution system for the supply of diagnostics, pharmaceuticals and educational materials.  ... 
doi:10.5694/mja15.c0518 pmid:25971555 fatcat:u6iv3jmstfcovonqdc6sqd3c4y

Astrophysics in 1999

Virginia Trimble, Markus J. Aschwanden
2000 Publications of the Astronomical Society of the Pacific  
The year 1999 saw the arrival of a star with three planets, a universe with three parameters, and a solar corona that could be heated at least three ways.  ...  The year 1999 saw the arrival of a star with three planets, a universe with three parameters, and a solar corona that could be heated at least three ways.  ...  Jensen (who normally used the middle name Hans) later won a Nobel Prize for the work.  ... 
doi:10.1086/316546 fatcat:nsx4ut4s4fhbbk2m7efjmwzqcy
« Previous Showing results 1 — 15 out of 50 results