-
Adaptive County Level COVID-19 Forecast Models: Analysis and Improvement
Authors:
Stewart W Doe,
Tyler Russell Seekins,
David Fitzpatrick,
Dawsin Blanchard,
Salimeh Yasaei Sekeh
Abstract:
Accurately forecasting county level COVID-19 confirmed cases is crucial to optimizing medical resources. Forecasting emerging outbreaks pose a particular challenge because many existing forecasting techniques learn from historical seasons trends. Recurrent neural networks (RNNs) with LSTM-based cells are a logical choice of model due to their ability to learn temporal dynamics. In this paper, we a…
▽ More
Accurately forecasting county level COVID-19 confirmed cases is crucial to optimizing medical resources. Forecasting emerging outbreaks pose a particular challenge because many existing forecasting techniques learn from historical seasons trends. Recurrent neural networks (RNNs) with LSTM-based cells are a logical choice of model due to their ability to learn temporal dynamics. In this paper, we adapt the state and county level influenza model, TDEFSI-LONLY, proposed in Wang et a. [l2020] to national and county level COVID-19 data. We show that this model poorly forecasts the current pandemic. We analyze the two week ahead forecasting capabilities of the TDEFSI-LONLY model with combinations of regularization techniques. Effective training of the TDEFSI-LONLY model requires data augmentation, to overcome this challenge we utilize an SEIR model and present an inter-county mixing extension to this model to simulate sufficient training data. Further, we propose an alternate forecast model, {\it County Level Epidemiological Inference Recurrent Network} (\alg{}) that trains an LSTM backbone on national confirmed cases to learn a low dimensional time pattern and utilizes a time distributed dense layer to learn individual county confirmed case changes each day for a two weeks forecast. We show that the best, worst, and median state forecasts made using CLEIR-Net model are respectively New York, South Carolina, and Montana.
△ Less
Submitted 1 July, 2020; v1 submitted 16 June, 2020;
originally announced June 2020.
-
Slimming Neural Networks using Adaptive Connectivity Scores
Authors:
Madan Ravi Ganesh,
Dawsin Blanchard,
Jason J. Corso,
Salimeh Yasaei Sekeh
Abstract:
In general, deep neural network (DNN) pruning methods fall into two categories: 1) Weight-based deterministic constraints, and 2) Probabilistic frameworks. While each approach has its merits and limitations there are a set of common practical issues such as, trial-and-error to analyze sensitivity and hyper-parameters to prune DNNs, which plague them both. In this work, we propose a new single-shot…
▽ More
In general, deep neural network (DNN) pruning methods fall into two categories: 1) Weight-based deterministic constraints, and 2) Probabilistic frameworks. While each approach has its merits and limitations there are a set of common practical issues such as, trial-and-error to analyze sensitivity and hyper-parameters to prune DNNs, which plague them both. In this work, we propose a new single-shot, fully automated pruning algorithm called Slimming Neural networks using Adaptive Connectivity Scores (SNACS). Our proposed approach combines a probabilistic pruning framework with constraints on the underlying weight matrices, via a novel connectivity measure, at multiple levels to capitalize on the strengths of both approaches while solving their deficiencies. In \alg{}, we propose a fast hash-based estimator of Adaptive Conditional Mutual Information (ACMI), that uses a weight-based scaling criterion, to evaluate the connectivity between filters and prune unimportant ones. To automatically determine the limit up to which a layer can be pruned, we propose a set of operating constraints that jointly define the upper pruning percentage limits across all the layers in a deep network. Finally, we define a novel sensitivity criterion for filters that measures the strength of their contributions to the succeeding layer and highlights critical filters that need to be completely protected from pruning. Through our experimental validation we show that SNACS is faster by over 17x the nearest comparable method and is the state of the art single-shot pruning method across three standard Dataset-DNN pruning benchmarks: CIFAR10-VGG16, CIFAR10-ResNet56 and ILSVRC2012-ResNet50.
△ Less
Submitted 17 December, 2021; v1 submitted 22 June, 2020;
originally announced June 2020.
-
Materials for hydrogen-based energy storage: Past, recent progress and future outlook
Authors:
Volodymyr A. Yartys,
Marcello Baricco,
Jose Bellosta von Colbe,
Didier Blanchard,
Robert C. Bowman Jr.,
Darren P. Broom,
Craig E. Buckley,
Fei Chang,
** Chen,
Young Whan Cho,
Jean-Claude Crivello,
Fermin Cuevas,
William I. F. David,
Petra E. de Jongh,
Roman V. Denys,
Martin Dornheim,
Michael Felderhoff,
Yaroslav Filinchuk,
George E. Froudakis,
David M. Grant,
Bjørn C. Hauback,
Ladislav Havela,
Teng He,
Michael Hirscher,
Terry D. Humphries
, et al. (23 additional authors not shown)
Abstract:
Magnesium hydride owns the largest share of publications on solid materials for hydrogen storage. The Magnesium group of international experts contributing to IEA Task 32 Hydrogen Based Energy Storage recently published two review papers presenting the activities of the group focused on magnesium hydride based materials and on Mg based compounds for hydrogen and energy storage. This review article…
▽ More
Magnesium hydride owns the largest share of publications on solid materials for hydrogen storage. The Magnesium group of international experts contributing to IEA Task 32 Hydrogen Based Energy Storage recently published two review papers presenting the activities of the group focused on magnesium hydride based materials and on Mg based compounds for hydrogen and energy storage. This review article not only overviews the latest activities on both fundamental aspects of Mg-based hydrides and their applications, but also presents a historic overview on the topic and outlines projected future developments. Particular attention is paid to the theoretical and experimental studies of Mg-H system at extreme pressures, kinetics and thermodynamics of the systems based on MgH2,nanostructuring, new Mg-based compounds and novel composites, and catalysis in the Mg based H storage systems. Finally, thermal energy storage and upscaled H storage systems accommodating MgH2 are presented.
△ Less
Submitted 7 May, 2020;
originally announced May 2020.
-
Is getting the right answer just about choosing the right words? The role of syntactically-informed features in short answer scoring
Authors:
Derrick Higgins,
Chris Brew,
Michael Heilman,
Ramon Ziai,
Lei Chen,
Aoife Cahill,
Michael Flor,
Nitin Madnani,
Joel Tetreault,
Daniel Blanchard,
Diane Napolitano,
Chong Min Lee,
John Blackmore
Abstract:
Developments in the educational landscape have spurred greater interest in the problem of automatically scoring short answer questions. A recent shared task on this topic revealed a fundamental divide in the modeling approaches that have been applied to this problem, with the best-performing systems split between those that employ a knowledge engineering approach and those that almost solely lever…
▽ More
Developments in the educational landscape have spurred greater interest in the problem of automatically scoring short answer questions. A recent shared task on this topic revealed a fundamental divide in the modeling approaches that have been applied to this problem, with the best-performing systems split between those that employ a knowledge engineering approach and those that almost solely leverage lexical information (as opposed to higher-level syntactic information) in assigning a score to a given response. This paper aims to introduce the NLP community to the largest corpus currently available for short-answer scoring, provide an overview of methods used in the shared task using this data, and explore the extent to which more syntactically-informed features can contribute to the short answer scoring task in a way that avoids the question-specific manual effort of the knowledge engineering approach.
△ Less
Submitted 5 March, 2014; v1 submitted 4 March, 2014;
originally announced March 2014.
-
Junction between a plate and a rod of comparable thickness in nonlinear elasticity. Part II
Authors:
Dominique Blanchard,
Georges Griso
Abstract:
We analyze the asymptotic behavior of a junction problem between a plate and a perpendicular rod made of a nonlinear elastic material. The two parts of this multi-structure have small thicknesses of the same order $δ$. We use the decomposition techniques obtained for the large deformations and the displacements in order to derive the limit energy as $δ$ tends to 0.
We analyze the asymptotic behavior of a junction problem between a plate and a perpendicular rod made of a nonlinear elastic material. The two parts of this multi-structure have small thicknesses of the same order $δ$. We use the decomposition techniques obtained for the large deformations and the displacements in order to derive the limit energy as $δ$ tends to 0.
△ Less
Submitted 21 October, 2012; v1 submitted 11 March, 2012;
originally announced March 2012.
-
Decomposition of deformations of thin rods. Application to nonlinear elasticity
Authors:
Dominique Blanchard,
Georges Griso
Abstract:
This paper deals with the introduction of a decomposition of the deformations of curved thin beams, with section of order $δ$, which takes into account the specific geometry of such beams. A deformation $v$ is split into an elementary deformation and a war**. The elementary deformation is the analog of a Bernoulli-Navier's displacement for linearized deformations replacing the infinitesimal rota…
▽ More
This paper deals with the introduction of a decomposition of the deformations of curved thin beams, with section of order $δ$, which takes into account the specific geometry of such beams. A deformation $v$ is split into an elementary deformation and a war**. The elementary deformation is the analog of a Bernoulli-Navier's displacement for linearized deformations replacing the infinitesimal rotation by a rotation in SO(3) in each cross section of the rod. Each part of the decomposition is estimated with respect to the $L^2$ norm of the distance from gradient $v$ to SO(3). This result relies on revisiting the rigidity theorem of Friesecke-James-Müller in which we estimate the constant for a bounded open set star-shaped with respect to a ball. Then we use the decomposition of the deformations to derive a few asymptotic geometrical behavior: large deformations of extensional type, inextensional deformations and linearized deformations. To illustrate the use of our decomposition in nonlinear elasticity, we consider a St Venant-Kirchhoff material and upon various scaling on the applied forces we obtain the $Γ$-limit of the rescaled elastic energy. We first analyze the case of bending forces of order $δ^2$ which leads to a nonlinear inextensional model. Smaller pure bending forces give the classical linearized model. A coupled extensional-bending model is obtained for a class of forces of order $δ^2$ in traction and of order $δ^3$ in bending.
△ Less
Submitted 12 September, 2011;
originally announced September 2011.
-
Asymptotic behavior of a structure made by a plate and a straight rod
Authors:
Dominique Blanchard,
Georges Griso
Abstract:
This paper is devoted to describe the asymptotic behavior of a structure made by a thin plate and a thin rod in the framework of nonlinear elasticity. We scale the applied forces in such a way that the level of the total elastic energy leads to the Von-Kármán's equations (or the linear model for smaller forces) in the plate and to a one dimensional rod-model at the limit. The junction conditions i…
▽ More
This paper is devoted to describe the asymptotic behavior of a structure made by a thin plate and a thin rod in the framework of nonlinear elasticity. We scale the applied forces in such a way that the level of the total elastic energy leads to the Von-Kármán's equations (or the linear model for smaller forces) in the plate and to a one dimensional rod-model at the limit. The junction conditions include in particular the continuity of the bending in the plate and the stretching in the rod at the junction.
△ Less
Submitted 26 July, 2011;
originally announced July 2011.
-
Asymptotic behavior of structures made of straight rods
Authors:
Dominique Blanchard,
Georges Griso
Abstract:
This paper is devoted to describe the deformations and the elastic energy for structures made of straight rods of thickness $2δ$ when $δ$ tends to 0. This analysis relies on the decomposition of the large deformation of a single rod introduced in [6] and on the extension of this technique to a multi-structure. We characterize the asymptotic behavior of the infimum of the total elastic energy as th…
▽ More
This paper is devoted to describe the deformations and the elastic energy for structures made of straight rods of thickness $2δ$ when $δ$ tends to 0. This analysis relies on the decomposition of the large deformation of a single rod introduced in [6] and on the extension of this technique to a multi-structure. We characterize the asymptotic behavior of the infimum of the total elastic energy as the minimum of a limit functional for an energy of order $δ^β$ ($2<β\le 4$).
△ Less
Submitted 18 October, 2010;
originally announced October 2010.
-
Compressed Sensing: How sharp is the Restricted Isometry Property
Authors:
Jeffrey D. Blanchard,
Coralia Cartis,
Jared Tanner
Abstract:
Compressed Sensing (CS) seeks to recover an unknown vector with $N$ entries by making far fewer than $N$ measurements; it posits that the number of compressed sensing measurements should be comparable to the information content of the vector, not simply $N$. CS combines the important task of compression directly with the measurement task. Since its introduction in 2004 there have been hundreds o…
▽ More
Compressed Sensing (CS) seeks to recover an unknown vector with $N$ entries by making far fewer than $N$ measurements; it posits that the number of compressed sensing measurements should be comparable to the information content of the vector, not simply $N$. CS combines the important task of compression directly with the measurement task. Since its introduction in 2004 there have been hundreds of manuscripts on CS, a large fraction of which develop algorithms to recover a signal from its compressed measurements. Because of the paradoxical nature of CS -- exact reconstruction from seemingly undersampled measurements -- it is crucial for acceptance of an algorithm that rigorous analyses verify the degree of undersampling the algorithm permits. The Restricted Isometry Property (RIP) has become the dominant tool used for the analysis in such cases. We present here an asymmetric form of RIP which gives tighter bounds than the usual symmetric one. We give the best known bounds on the RIP constants for matrices from the Gaussian ensemble. Our derivations illustrate the way in which the combinatorial nature of CS is controlled. Our quantitative bounds on the RIP allow precise statements as to how aggressively a signal can be undersampled, the essential question for practitioners. We also document the extent to which RIP gives precise information about the true performance limits of CS, by comparing with approaches from high-dimensional geometry.
△ Less
Submitted 28 April, 2010;
originally announced April 2010.
-
Phase Transitions for Greedy Sparse Approximation Algorithms
Authors:
Jeffrey D. Blanchard,
Coralia Cartis,
Jared Tanner,
Andrew Thompson
Abstract:
A major enterprise in compressed sensing and sparse approximation is the design and analysis of computationally tractable algorithms for recovering sparse, exact or approximate, solutions of underdetermined linear systems of equations. Many such algorithms have now been proven to have optimal-order uniform recovery guarantees using the ubiquitous Restricted Isometry Property (RIP). However, it is…
▽ More
A major enterprise in compressed sensing and sparse approximation is the design and analysis of computationally tractable algorithms for recovering sparse, exact or approximate, solutions of underdetermined linear systems of equations. Many such algorithms have now been proven to have optimal-order uniform recovery guarantees using the ubiquitous Restricted Isometry Property (RIP). However, it is unclear when the RIP-based sufficient conditions on the algorithm are satisfied. We present a framework in which this task can be achieved; translating these conditions for Gaussian measurement matrices into requirements on the signal's sparsity level, length, and number of measurements. We illustrate this approach on three of the state-of-the-art greedy algorithms: CoSaMP, Subspace Pursuit (SP), and Iterative Hard Thresholding (IHT). Designed to allow a direct comparison of existing theory, our framework implies that, according to the best known bounds, IHT requires the fewest number of compressed sensing measurements and has the lowest per iteration computational cost of the three algorithms compared here.
△ Less
Submitted 11 April, 2010;
originally announced April 2010.
-
A simplified model for elastic thin shells
Authors:
Dominique Blanchard,
Georges Griso
Abstract:
We introduce a simplified model for the minimization of the elastic energy in thin shells. This model is not obtained by an asymptotic analysis. The thickness of the shell remains a parameter as in Reisner-Mindlin's model for plates and Koiter's model for shells in the linear case. The simplified model admits always minimizers by contrast with the original one. We show the relevance of our approac…
▽ More
We introduce a simplified model for the minimization of the elastic energy in thin shells. This model is not obtained by an asymptotic analysis. The thickness of the shell remains a parameter as in Reisner-Mindlin's model for plates and Koiter's model for shells in the linear case. The simplified model admits always minimizers by contrast with the original one. We show the relevance of our approach by proving that the minimum of the simplified model and the infimum of the full model have the same limit as the thickness tends to 0.
△ Less
Submitted 20 November, 2010; v1 submitted 18 February, 2010;
originally announced February 2010.
