The Experts below are selected from a list of 56553 Experts worldwide ranked by ideXlab platform
Robert E Schapire - One of the best experts on this subject based on the ideXlab platform.
-
maximum Entropy Density estimation with generalized regularization and an application to species distribution modeling
Journal of Machine Learning Research, 2007Co-Authors: Miroslav Dudik, Steven J Phillips, Robert E SchapireAbstract:We present a unified and complete account of maximum Entropy Density estimation subject to constraints represented by convex potential functions or, alternatively, by convex regularization. We provide fully general performance guarantees and an algorithm with a complete convergence proof. As special cases, we easily derive performance guarantees for many known regularization types, including l1, l2, l22, and l2 + l22 style regularization. We propose an algorithm solving a large and general subclass of generalized maximum Entropy problems, including all discussed in the paper, and prove its convergence. Our approach generalizes and unifies techniques based on information geometry and Bregman divergences as well as those based more directly on compactness. Our work is motivated by a novel application of maximum Entropy to species distribution modeling, an important problem in conservation biology and ecology. In a set of experiments on real-world data, we demonstrate the utility of maximum Entropy in this setting. We explore effects of different feature types, sample sizes, and regularization levels on the performance of maxent, and discuss interpretability of the resulting models.
-
hierarchical maximum Entropy Density estimation
International Conference on Machine Learning, 2007Co-Authors: Miroslav Dudik, David M Blei, Robert E SchapireAbstract:We study the problem of simultaneously estimating several densities where the datasets are organized into overlapping groups, such as a hierarchy. For this problem, we propose a maximum Entropy formulation, which systematically incorporates the groups and allows us to share the strength of prediction across similar datasets. We derive general performance guarantees, and show how some previous approaches, such as hierarchical shrinkage and hierarchical priors, can be derived as special cases. We demonstrate the proposed technique on synthetic data and in a real-world application to modeling the geographic distributions of species hierarchically grouped in a taxonomy. Specifically, we model the geographic distributions of species in the Australian wet tropics and Northeast New South Wales. In these regions, small numbers of samples per species significantly hinder effective prediction. Substantial benefits are obtained by combining information across taxonomic groups.
-
correcting sample selection bias in maximum Entropy Density estimation
Neural Information Processing Systems, 2005Co-Authors: Miroslav Dudik, Steven J Phillips, Robert E SchapireAbstract:We study the problem of maximum Entropy Density estimation in the presence of known sample selection bias. We propose three bias correction approaches. The first one takes advantage of unbiased sufficient statistics which can be obtained from biased samples. The second one estimates the biased distribution and then factors the bias out. The third one approximates the second by only using samples from the sampling distribution. We provide guarantees for the first two approaches and evaluate the performance of all three approaches in synthetic experiments and on real data from species habitat modeling, where maxent has been successfully applied and where sample selection bias is a significant problem.
-
NIPS - Correcting sample selection bias in maximum Entropy Density estimation
2005Co-Authors: Miroslav Dudik, Steven J Phillips, Robert E SchapireAbstract:We study the problem of maximum Entropy Density estimation in the presence of known sample selection bias. We propose three bias correction approaches. The first one takes advantage of unbiased sufficient statistics which can be obtained from biased samples. The second one estimates the biased distribution and then factors the bias out. The third one approximates the second by only using samples from the sampling distribution. We provide guarantees for the first two approaches and evaluate the performance of all three approaches in synthetic experiments and on real data from species habitat modeling, where maxent has been successfully applied and where sample selection bias is a significant problem.
Miroslav Dudik - One of the best experts on this subject based on the ideXlab platform.
-
maximum Entropy Density estimation with generalized regularization and an application to species distribution modeling
Journal of Machine Learning Research, 2007Co-Authors: Miroslav Dudik, Steven J Phillips, Robert E SchapireAbstract:We present a unified and complete account of maximum Entropy Density estimation subject to constraints represented by convex potential functions or, alternatively, by convex regularization. We provide fully general performance guarantees and an algorithm with a complete convergence proof. As special cases, we easily derive performance guarantees for many known regularization types, including l1, l2, l22, and l2 + l22 style regularization. We propose an algorithm solving a large and general subclass of generalized maximum Entropy problems, including all discussed in the paper, and prove its convergence. Our approach generalizes and unifies techniques based on information geometry and Bregman divergences as well as those based more directly on compactness. Our work is motivated by a novel application of maximum Entropy to species distribution modeling, an important problem in conservation biology and ecology. In a set of experiments on real-world data, we demonstrate the utility of maximum Entropy in this setting. We explore effects of different feature types, sample sizes, and regularization levels on the performance of maxent, and discuss interpretability of the resulting models.
-
hierarchical maximum Entropy Density estimation
International Conference on Machine Learning, 2007Co-Authors: Miroslav Dudik, David M Blei, Robert E SchapireAbstract:We study the problem of simultaneously estimating several densities where the datasets are organized into overlapping groups, such as a hierarchy. For this problem, we propose a maximum Entropy formulation, which systematically incorporates the groups and allows us to share the strength of prediction across similar datasets. We derive general performance guarantees, and show how some previous approaches, such as hierarchical shrinkage and hierarchical priors, can be derived as special cases. We demonstrate the proposed technique on synthetic data and in a real-world application to modeling the geographic distributions of species hierarchically grouped in a taxonomy. Specifically, we model the geographic distributions of species in the Australian wet tropics and Northeast New South Wales. In these regions, small numbers of samples per species significantly hinder effective prediction. Substantial benefits are obtained by combining information across taxonomic groups.
-
correcting sample selection bias in maximum Entropy Density estimation
Neural Information Processing Systems, 2005Co-Authors: Miroslav Dudik, Steven J Phillips, Robert E SchapireAbstract:We study the problem of maximum Entropy Density estimation in the presence of known sample selection bias. We propose three bias correction approaches. The first one takes advantage of unbiased sufficient statistics which can be obtained from biased samples. The second one estimates the biased distribution and then factors the bias out. The third one approximates the second by only using samples from the sampling distribution. We provide guarantees for the first two approaches and evaluate the performance of all three approaches in synthetic experiments and on real data from species habitat modeling, where maxent has been successfully applied and where sample selection bias is a significant problem.
-
NIPS - Correcting sample selection bias in maximum Entropy Density estimation
2005Co-Authors: Miroslav Dudik, Steven J Phillips, Robert E SchapireAbstract:We study the problem of maximum Entropy Density estimation in the presence of known sample selection bias. We propose three bias correction approaches. The first one takes advantage of unbiased sufficient statistics which can be obtained from biased samples. The second one estimates the biased distribution and then factors the bias out. The third one approximates the second by only using samples from the sampling distribution. We provide guarantees for the first two approaches and evaluate the performance of all three approaches in synthetic experiments and on real data from species habitat modeling, where maxent has been successfully applied and where sample selection bias is a significant problem.
Paul M. Hohler - One of the best experts on this subject based on the ideXlab platform.
-
A sticky business: the status of the conjectured viscosity/Entropy Density bound
Journal of High Energy Physics, 2008Co-Authors: Aleksey Cherman, Thomas D. Cohen, Paul M. HohlerAbstract:There have been a number of forms of a conjecture that there is a universal lower bound on the ratio, eta/s, of the shear viscosity, eta, to Entropy Density, s, with several different domains of validity. We examine the various forms of the conjecture. We argue that a number of variants of the conjecture are not viable due to the existence of theoretically consistent counterexamples. We also note that much of the evidence in favor of a bound does not apply to the variants which have not yet been ruled out.
-
a sticky business the status of the conjectured viscosity Entropy Density bound
Journal of High Energy Physics, 2008Co-Authors: Aleksey Cherman, Thomas D. Cohen, Paul M. HohlerAbstract:There have been a number of forms of a conjecture that there is a universal lower bound on the ratio, η/s, of the shear viscosity, η, to Entropy Density, s, with several different domains of validity. We examine the various forms of the conjecture. We argue that a number of variants of the conjecture are not viable due to the existence of theoretically consistent counterexamples. We also note that much of the evidence in favor of a bound does not apply to the variants which have not yet been ruled out.
-
a sticky business the status of the conjectured viscosity Entropy Density bound
arXiv: High Energy Physics - Theory, 2007Co-Authors: Aleksey Cherman, Thomas D. Cohen, Paul M. HohlerAbstract:There have been a number of forms of a conjecture that there is a universal lower bound on the ratio, eta/s, of the shear viscosity, eta, to Entropy Density, s, with several different domains of validity. We examine the various forms of the conjecture. We argue that a number of variants of the conjecture are not viable due to the existence of theoretically consistent counterexamples. We also note that much of the evidence in favor of a bound does not apply to the variants which have not yet been ruled out.
Antal Jakovác - One of the best experts on this subject based on the ideXlab platform.
-
Shear viscosity over Entropy Density ratio with extended quasiparticles
Physical Review D, 2016Co-Authors: M. Horváth, Antal JakovácAbstract:We consider an effective field theory description of beyond-quasi-particle excitations aiming to associate the transport properties of the system with the spectral Density of states. Tuning various properties of the many-particle correlations, we investigate how the robust microscopic features are translated into the macroscopic observables like shear viscosity and Entropy Density. The liquid-gas crossover is analysed using several examples. A thermal constraint on the fluidity measure, the ratio of shear viscosity to Entropy Density, is discussed.
-
Nonuniversal lower bound for the shear viscosity to Entropy Density ratio
Physical Review D, 2010Co-Authors: Antal JakovácAbstract:The lower bound of the shear viscosity to Entropy Density ratio is examined using an exact representation of the ratio through the Density of states. Under certain assumptions it can be shown that the lower bound of the ratio is not universal; its value is determined by the Entropy Density. Some examples of physical systems are discussed in the paper where one can expect violation of the conformal $1/4\ensuremath{\pi}$ value.
-
Viscosity to Entropy Density ratio: violation of the lower bound at small temperatures
arXiv: High Energy Physics - Theory, 2009Co-Authors: Antal JakovácAbstract:We show that in theories where the lowest energy excitations are not quasiparticles but they form a continuum, the shear viscosity to Entropy Density ratio goes to zero as the temperature goes to zero. In these theories therefore there is no lower bound for the shear viscosity to Entropy Density ratio, in contrast to the predictions coming from the AdS/CFT correspondence.
Steven J Phillips - One of the best experts on this subject based on the ideXlab platform.
-
maximum Entropy Density estimation with generalized regularization and an application to species distribution modeling
Journal of Machine Learning Research, 2007Co-Authors: Miroslav Dudik, Steven J Phillips, Robert E SchapireAbstract:We present a unified and complete account of maximum Entropy Density estimation subject to constraints represented by convex potential functions or, alternatively, by convex regularization. We provide fully general performance guarantees and an algorithm with a complete convergence proof. As special cases, we easily derive performance guarantees for many known regularization types, including l1, l2, l22, and l2 + l22 style regularization. We propose an algorithm solving a large and general subclass of generalized maximum Entropy problems, including all discussed in the paper, and prove its convergence. Our approach generalizes and unifies techniques based on information geometry and Bregman divergences as well as those based more directly on compactness. Our work is motivated by a novel application of maximum Entropy to species distribution modeling, an important problem in conservation biology and ecology. In a set of experiments on real-world data, we demonstrate the utility of maximum Entropy in this setting. We explore effects of different feature types, sample sizes, and regularization levels on the performance of maxent, and discuss interpretability of the resulting models.
-
correcting sample selection bias in maximum Entropy Density estimation
Neural Information Processing Systems, 2005Co-Authors: Miroslav Dudik, Steven J Phillips, Robert E SchapireAbstract:We study the problem of maximum Entropy Density estimation in the presence of known sample selection bias. We propose three bias correction approaches. The first one takes advantage of unbiased sufficient statistics which can be obtained from biased samples. The second one estimates the biased distribution and then factors the bias out. The third one approximates the second by only using samples from the sampling distribution. We provide guarantees for the first two approaches and evaluate the performance of all three approaches in synthetic experiments and on real data from species habitat modeling, where maxent has been successfully applied and where sample selection bias is a significant problem.
-
NIPS - Correcting sample selection bias in maximum Entropy Density estimation
2005Co-Authors: Miroslav Dudik, Steven J Phillips, Robert E SchapireAbstract:We study the problem of maximum Entropy Density estimation in the presence of known sample selection bias. We propose three bias correction approaches. The first one takes advantage of unbiased sufficient statistics which can be obtained from biased samples. The second one estimates the biased distribution and then factors the bias out. The third one approximates the second by only using samples from the sampling distribution. We provide guarantees for the first two approaches and evaluate the performance of all three approaches in synthetic experiments and on real data from species habitat modeling, where maxent has been successfully applied and where sample selection bias is a significant problem.