317 Hits in 7.6 sec

Exact and Consistent Interpretation of Piecewise Linear Models Hidden behind APIs: A Closed Form Solution [article]

Zicun Cong, Lingyang Chu, Lanjun Wang, Xia Hu, Jian Pei
2020 arXiv   pre-print
In this paper, we propose an elegant closed form solution named OpenAPI to compute exact and consistent interpretations for the family of Piecewise Linear Models (PLM), which includes many popular classification  ...  Existing works interpret the predictions of a model hidden behind an API by heuristically probing the response of the API with perturbed input instances.  ...  Can we compute exact and consistent interpretations of decisions made by predictive models hidden behind cloud service APIs?  ... 
arXiv:1906.06857v2 fatcat:m7x6o6iirzaevfozaaurm7ym3a

Advancing Artificial Neural Network Parameterisation for Atmospheric Turbulence using a Variational Multiscale Model

M. Janssens, S. J. Hulshoff
2021 Journal of Advances in Modeling Earth Systems  
Intercomparison studies of existing atmospheric General Circulation Models (GCMs) exhibit a large spread in their predictions of atmospheric 2 CO E concentration at which a 2 K temperature rise with respect  ...  Such clouds are currently approximated by phenomenological unresolved scales models, "parameterizations", of considerably lower fidelity than the resolved simulation.  ...  Acknowledgments We gratefully acknowledge Pier Siebesma and Harm Jonker for their considerable contribution to shaping this study and for selecting an appropriate test case; we also thank Harm for running  ... 
doi:10.1029/2021ms002490 fatcat:3sxnwtabgbgthcpdunor55ywta

The hidden models of model checking

Willem Visser, Matthew B. Dwyer, Michael Whalen
2012 Journal of Software and Systems Modeling  
We illustrate the value of hidden models by surveying successful examples of their use in different areas of model checking (hardware and software) and how a lack of suitable models hamper a new area (  ...  In the past, applying formal analysis, such as model checking, to industrial problems required a team of formal methods experts and a great deal of effort.  ...  In RoVerGene, these networks are modeled as a system of piecewise-affine equations which are translated to a form that is amenable to model checking.  ... 
doi:10.1007/s10270-012-0281-9 fatcat:hpfpl3ufwrhpxbhg7koafunnuq

Tractable Density Estimation on Learned Manifolds with Conformal Embedding Flows [article]

Brendan Leigh Ross, Jesse C. Cresswell
2021 arXiv   pre-print
To this end, we present a series of conformal building blocks and apply them in experiments with synthetic and real-world data to demonstrate that flows can model manifold-supported distributions without  ...  Recent attempts to remedy this limitation have introduced geometric complications that defeat a central benefit of normalizing flows: exact density estimation.  ...  Acknowledgments and Disclosure of Funding We thank Gabriel Loaiza-Ganem and Anthony Caterini for their valuable discussions and advice.  ... 
arXiv:2106.05275v2 fatcat:j4lrysqoobbetbwkd5pa4dxssu

How can big data and machine learning benefit environment and water management: A survey of methods, applications, and future directions

Alexander Y. Sun, Bridget R Scanlon
2019 Environmental Research Letters  
Acknowledgments The authors were partially supported by funding from Jackson School of Geosciences, the University of Texas at Austin.  ...  The authors are grateful to Dr Michael Fienen and an anonymous reviewer for their constructive comments on the original manuscript.  ...  (typically close to zero), r j denotes the average activation of the jth hidden unit that is averaged over the training set, and s is the total number of hidden units in a layer.  ... 
doi:10.1088/1748-9326/ab1b7d fatcat:vx4thuy45vhlnmhu7bk2hwh2g4

icepack: a new glacier flow modeling package in Python, version 1.0

Daniel R. Shapero, Jessica A. Badgeley, Andrew O. Hoffman, Ian R. Joughin
2021 Geoscientific Model Development  
We introduce a new software package called "icepack" for modeling the flow of glaciers and ice sheets.  ...  The icepack package is built on the finite element modeling library Firedrake, which uses the Unified Form Language (UFL), a domain-specific language embedded into Python for describing weak forms of partial  ...  This research has been supported by the National Aeronautics and Space Administration (grant nos. 80NSSC20K0954 and 80NSSC20K1627) and the National Science Foundation (grant nos. 1835321 and 1643285).  ... 
doi:10.5194/gmd-14-4593-2021 fatcat:sdw2qyvwdrcl5pyg7v4zdnkcqi

Preventing self-intersection under free-form deformation

J.E. Gain, N.A. Dodgson
2001 IEEE Transactions on Visualization and Computer Graphics  
AbstractÐFree-Form Deformation (FFD) is a versatile and efficient modeling technique which transforms an object by warping the surrounding space.  ...  In this paper, we present the following novel results: a set of theoretical conditions for preventing self-intersection by ensuring the injectivity (one-to-one mapping) of FFD, an exact (necessary and  ...  ae INTRODUCTION F REE-FORM Deformation (FFD) [7] , [29] and its extensions form a class of highly intuitive, versatile, and efficient free-form modeling and animation tools.  ... 
doi:10.1109/2945.965344 fatcat:aqthhitqfrab3kbnfi7tr5q3li

A self-organization based optical flow estimator with GPU implementation

Manish P. Shiralkar, Robert J. Schalkoff
2011 Machine Vision and Applications  
This gradient-based estimator handles the ill-posedness in motion estimation via a novel combination of regression and a self organization strategy.  ...  The self organization step uses a uniquely designed pair of training set (Q = H c ) and the initial weights set (W = H c ∪ H p ).  ...  Moreover, we will see in the next chapter that using the interpretation of SOM as mixture model plus additional regularization, how a special case of SOM leads to non-linear interpolation of optical flow  ... 
doi:10.1007/s00138-011-0352-x fatcat:7hn6rwfhpbcqdndjg4fmn6ajbi

Photo-Realistic Rendering - Recent Trends and Developments [article]

Philipp Slusallek
1997 Eurographics State of the Art Reports  
This report gives an overview of advanced algorithms for photo-realistic rendering and discusses recent trends and developments.  ...  The goal to create photo-realistic images has lead to the development of completely new software techniques for dealing with the inherent geometric and optical complexity of real world scenes.  ...  Figure 6 : 6 Figure 6: The model of a station consisting of approximately 32000 patches and 24 light sources.  ... 
doi:10.2312/egst.19971005 fatcat:ls3wan52lnfwhlwo6mxmzzfv4e

Point-to-point shortest paths on dynamic time-dependent road networks

Giacomo Nannicini
2010 4OR  
are linear or piecewise linear functions, whereas it is a Mixed-Integer Nonlinear Program (MINLP) if the arc weights are nonlinear functions.  ...  However, applying Dijkstra's algorithm on a graph with several millions of vertices and arcs, such as a continental road network, may require several seconds of CPU time.  ...  Acknowledgements First of all, I want to thank Leo Liberti for giving me the possibility of working on this thesis: he always encouraged me to try harder, and much of this work would not have been done  ... 
doi:10.1007/s10288-010-0121-0 fatcat:wwp2d2dyeretfm2qabxwy73wjy

ExaHyPE: An engine for parallel dynamically adaptive simulations of wave problems

Anne Reinarz, Dominic E. Charrier, Michael Bader, Luke Bovard, Michael Dumbser, Kenneth Duru, Francesco Fambri, Alice-Agnes Gabriel, Jean-Matthieu Gallard, Sven Köppel, Lukas Krenz, Leonhard Rannabauer (+4 others)
2020 Computer Physics Communications  
: The ExaHyPE PDE engine offers robust algorithms to solve linear and non-linear hyperbolic systems of PDEs written in first order form.  ...  Due to the robustness and shock capturing abilities of ExaHyPE's numerical methods, users of the engine can simulate linear and non-linear hyperbolic PDEs with very high accuracy.  ...  Acknowledgements and funding This project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 671698,  ... 
doi:10.1016/j.cpc.2020.107251 fatcat:mnpkmiqumrhhpgnc6lph53e3ya

SpECTRE: A task-based discontinuous Galerkin code for relativistic astrophysics

Lawrence E. Kidder, Scott E. Field, Francois Foucart, Erik Schnetter, Saul A. Teukolsky, Andy Bohn, Nils Deppe, Peter Diener, François Hébert, Jonas Lippuner, Jonah Miller, Christian D. Ott (+2 others)
2017 Journal of Computational Physics  
A task-based parallelism model allows efficient use of the largest supercomputers for problems with a heterogeneous workload over disparate spatial and temporal scales.  ...  We introduce a new relativistic astrophysics code, SpECTRE, that combines a discontinuous Galerkin method with a task-based parallelism model.  ...  Brügmann, Marcus Bugner, Mani Chandra, Roland Haas, Daniel Hemberger, Jan Hesthaven, Cameron Hummels, Fraser Hutchison, Jonathan Lifflander, Geoffrey Lovelace, Phil Miller, Harald Pfeiffer, David Radice, and  ... 
doi:10.1016/ fatcat:m6f2qmbs6vhxhdtnqkekimpp7i

State of the Art Report on Real-time Rendering with Hardware Tessellation [article]

H. Schäfer, M. Nießner, B. Keinert, M. Stamminger, C. Loop
2014 Eurographics State of the Art Reports  
For a long time, GPUs have primarily been optimized to render more and more triangles with increasingly flexible shading.  ...  In this state of the art report, we provide an overview of recent work and challenges in this topic by summarizing, discussing and comparing methods for the rendering of smooth and highly detailed surfaces  ...  These are piecewise linear upper and lower bounding polylines that can be computed using a precomputed table of numbers and a curve's Bézier control points.  ... 
doi:10.2312/egst.20141037 fatcat:5tahtayzszd3fletv4yk3rnh2i

Distance Geometry and Data Science [article]

Leo Liberti
2019 arXiv   pre-print
We discuss applications, solution methods, dimensional reduction techniques and some of their limits.  ...  In this survey we discuss the fundamental problem of mapping graphs to vectors, and its relation with mathematical programming.  ...  Maculan, A. Mucherino. I have first heard of concentration of measure as I passed by D. Malioutov's office at the T.J.  ... 
arXiv:1909.08544v1 fatcat:5m6n4lozsra6fbo27gqymi2fte

14th International Symposium on Mathematical Programming

1990 Mathematical programming  
A v ariation of bundle methods is presented by means of which such a t ype of problems can be solved.In this variation a dynamic polyhedral model of dom is generated by means of normal vectors.In addition  ...  to the well known serious steps und null steps of bundle methods a third type of steps is used to generate the model of dom .  ...  consistent subsystems -inconsistent linear systems -motion analysis -piecewise linear models -time series modeling Although linear models play an important role in the study of a n umber of signal and  ... 
doi:10.1007/bf01580875 fatcat:3jtclwmntzgjxkqs5uecombdaa
« Previous Showing results 1 — 15 out of 317 results