Results 1  10
of
5,254
Algorithms for Nonnegative Matrix Factorization
 In NIPS
, 2001
"... Nonnegative matrix factorization (NMF) has previously been shown to be a useful decomposition for multivariate data. Two different multiplicative algorithms for NMF are analyzed. They differ only slightly in the multiplicative factor used in the update rules. One algorithm can be shown to minim ..."
Abstract

Cited by 1246 (5 self)
 Add to MetaCart
to minimize the conventional least squares error while the other minimizes the generalized KullbackLeibler divergence. The monotonic convergence of both algorithms can be proven using an auxiliary function analogous to that used for proving convergence of the ExpectationMaximization algorithm
Object Recognition from Local ScaleInvariant Features
"... An object recognition system has been developed that uses a new class of local image features. The features are invariant to image scaling, translation, and rotation, and partially invariant to illumination changes and affine or 3D projection. These features share similar properties with neurons in ..."
Abstract

Cited by 2739 (13 self)
 Add to MetaCart
in inferior temporal cortex that are used for object recognition in primate vision. Features are efficiently detected through a staged filtering approach that identifies stable points in scale space. Image keys are created that allow for local geometric deformations by representing blurred image gradients
Gradient Domain High Dynamic Range Compression
 PROCEEDINGS OF ACM SIGGRAPH 2002
, 2002
"... We present a new method for rendering high dynamic range images on conventional displays. Our method is conceptually simple, computationally efficient, robust, and easy to use. We manipulate the gradient field of the luminance image by attenuating the magnitudes of large gradients. A new, low dynami ..."
Abstract

Cited by 380 (10 self)
 Add to MetaCart
We present a new method for rendering high dynamic range images on conventional displays. Our method is conceptually simple, computationally efficient, robust, and easy to use. We manipulate the gradient field of the luminance image by attenuating the magnitudes of large gradients. A new, low
An Improved In Situ and Satellite SST Analysis for Climate
 J Clim 15:1609–1625. doi
, 2002
"... A weekly 18 spatial resolution optimum interpolation (OI) sea surface temperature (SST) analysis has been produced at the National Oceanic and Atmospheric Administration (NOAA) using both in situ and satellite data from November 1981 to the present. The weekly product has been available since 1993 a ..."
Abstract

Cited by 391 (12 self)
 Add to MetaCart
global residual biases of roughly 20.038C. The major improvement in the analysis occurs at high latitudes due to the new sea ice algorithm where local differences between the old and new analysis can exceed 18C. Comparisons with other SST products are needed to determine the consistency of the OI
Residual Algorithms: Reinforcement Learning with Function Approximation
 In Proceedings of the Twelfth International Conference on Machine Learning
, 1995
"... A number of reinforcement learning algorithms have been developed that are guaranteed to converge to the optimal solution when used with lookup tables. It is shown, however, that these algorithms can easily become unstable when implemented directly with a general functionapproximation system, such ..."
Abstract

Cited by 307 (6 self)
 Add to MetaCart
, such as a sigmoidal multilayer perceptron, a radialbasisfunction system, a memorybased learning system, or even a linear functionapproximation system. A new class of algorithms, residual gradient algorithms, is proposed, which perform gradient descent on the mean squared Bellman residual, guaranteeing
Stochastic Gradient Boosting
 Computational Statistics and Data Analysis
, 1999
"... Gradient boosting constructs additive regression models by sequentially fitting a simple parameterized function (base learner) to current "pseudo"residuals by leastsquares at each iteration. The pseudoresiduals are the gradient of the loss functional being minimized, with respect to ..."
Abstract

Cited by 285 (1 self)
 Add to MetaCart
Gradient boosting constructs additive regression models by sequentially fitting a simple parameterized function (base learner) to current "pseudo"residuals by leastsquares at each iteration. The pseudoresiduals are the gradient of the loss functional being minimized, with respect
Reinforcement Learning In Continuous Time and Space
 Neural Computation
, 2000
"... This paper presents a reinforcement learning framework for continuoustime dynamical systems without a priori discretization of time, state, and action. Based on the HamiltonJacobiBellman (HJB) equation for infinitehorizon, discounted reward problems, we derive algorithms for estimating value f ..."
Abstract

Cited by 176 (7 self)
 Add to MetaCart
are derived and their correspondences with the conventional residual gradient, TD(0), and TD() algorithms are shown. For policy improvement, two methods, namely, a continuous actorcritic method and a valuegradient based greedy policy, are formulated. As a special case of the latter, a nonlinear feedback
Modelbased Geostatistics
 Applied Statistics
, 1998
"... Conventional geostatistical methodology solves the problem of predicting the realised value of a linear functional of a Gaussian spatial stochastic process, S(x), based on observations Y i = S(x i ) + Z i at sampling locations x i , where the Z i are mutually independent, zeromean Gaussian random v ..."
Abstract

Cited by 228 (9 self)
 Add to MetaCart
Conventional geostatistical methodology solves the problem of predicting the realised value of a linear functional of a Gaussian spatial stochastic process, S(x), based on observations Y i = S(x i ) + Z i at sampling locations x i , where the Z i are mutually independent, zeromean Gaussian random
Adaptive Segmentation of MRI data
, 1995
"... Intensitybased classification of MR images has proven problematic, even when advanced techniques are used. Intrascan and interscan intensity inhomogeneities are a common source of difficulty. While reported methods have had some success in correcting intrascan inhomogeneities, such methods requi ..."
Abstract

Cited by 224 (15 self)
 Add to MetaCart
segmentation of tissue types as well as better visualization of MRI data, that has proven to be effective in a study that includes more than 1000 brain scans. Implementation and results are described for segmenting the brain in the following types of images: axial (dualecho spinecho), coronal (3DFT gradient
Global and regional climate changes due to black carbon,
 Nat. Geosci.,
, 2008
"... Figure 1: Global distribution of BC sources and radiative forcing. a, BC emission strength in tons per year from a study by Bond et al. Full size image (42 KB) Review Nature Geoscience 1, 221 227 (2008 Black carbon in soot is the dominant absorber of visible solar radiation in the atmosphere. Ant ..."
Abstract

Cited by 228 (5 self)
 Add to MetaCart
to as soot 1 . BC in indoor environments is largely due to cooking with biofuels such as wood, dung and crop residue. Outdoors, it is due to fossil fuel combustion (diesel and coal), open biomass burning (associated with deforestation and crop residue burning), and cooking with biofuels 1 . Soot aerosols
Results 1  10
of
5,254