sentence1
stringlengths 16
446
| sentence2
stringlengths 14
436
|
|---|---|
finally , generative adversarial networks have recently made great progress in learning complete data generation models from unlabeled data . | furthermore , generative adversarial networks , have shown promise as well , generating almost photo-realistic images for particular datasets . |
harmonic morphisms and twistorial maps in this section we shall work in the complex analytic category . | harmonic morphisms between weyl spaces in this section we shall work in the smooth and . |
rheological studies suggest that the microgels collapse to a smaller diameter at the interface , leading to a reduced interfacial coverage and thus a reduced emulsion stability . | rheological studies suggest that the microgels collapse to a smaller diameter at a liquid interface , leading to a reduced interfacial coverage and thus a reduced emulsion stability . |
recently , convolutional neural networks have achieved remarkable success in semantic image segmentation . | convolutional neural networks have achieved state-of-the-art performance on visual tasks such as image and video recognition in the last few years . |
the instability is the signature to a phase transition into a state with coherent radiation . | this instability is the origin of the formation of a critical point in the bootstrap matter . |
a monoid in m is the same as an m-enriched category with one object . | thus a monoid is a group without inverses . |
we also test our algorithm on a public , block-design fmri dataset from a study on face and object representation in human ventral temporal cortex . | we used a popular fmri dataset from a study on face and object representation in human ventral temporal cortex . |
but we also obtain the following theorem on more general µl-stable parabolic higgs bundles . | we also see the uniqueness of the adapted pluri-harmonic metric for parabolic higgs bundles . |
we shall derive the consistent equations for the lagrangian perturbations with the backreaction term qd . | using the lagrangian perturbation s , we shall rewrite the backreaction term . |
the third maps are the homology isomorphism . | the second map is the hurewicz homomorphism . |
we obtain gorenstein algebras with hilbert functions , respectively , and . | repeating the procedure , we obtain gorenstein hilbert functions and . |
the calculations were performed using the projector augmented wave method , as implemented in the vienna ab-initio simulation package . | the first-principles calculations were carried in the framework of the projector augmented-wave formalism as implemented in the vienna ab initio simulation package . |
feng et al by using the the distance between the test set and unrelated training sets , in addition to the distance between the test image set and the related training set . | feng et al extends the work of chen by using the the distance between the test set and unrelated training sets , in addition to the distance between the test image set and the related training set . |
the data were calibrated and cleaned using the miriad package . | the sma data were reduced using the miriad software package . |
in practice , one usually has cavities with full-cell termination , and in this case one has to detune the frequences of the end cells to obtain a flat field distribution in the cavity . | in practice one usually has cavities with full-cell termination , and in this case one has to detune the frequency of the end cells to obtain a flat field distribution in the cavity . |
we also study the improvement in the convergence obtained by using the pms criterion for m . | more rapid convergence can again be obtained by using the pms criterion for µ . |
conversely , every quadruple or and additional data mentioned above . | conversely , every quadruple and some additional data with the properties just listed . |
cai and yuan introduced an adaptive block thresholding estimator which is simultaneously rate optimal rate over large collections of bandable covariance matrices . | cai and yuan proposed a block thresholding procedure which is shown to be adaptively rate-optimal over a wide range of collections of bandable covariance matrices . |
the generalized gradient approximation in the parametrization of perdew , burke and ernzerhof was used as approximation for the exchange and correlation functional . | a perdew-burke-ernzerhof form of generalized gradient approximation was employed as the exchange-correlation functional . |
recent methods based on convolutional neural networks have been shown to produce results of high accuracy for a wide range of challenging computer vision tasks like image recognition . | recent development of deep convolutional neural networks has led to great success in a variety of tasks including image classfication and others . |
kolmogorov complexity is , however , not computable , but we can approximate it in a well-founded and computable way through the minimum description length principle . | kolmogorov complexity is not computable , but can be approximated through the minimum description length principle , which we use to instantiate this framework . |
dft calculations were performed by using the vienna ab initio simulation package . | the vienna ab-initio simulation package was used to perform the required computations . |
recently , deterministic deep neural networks have demonstrated state-of-the-art performance on many supervised tasks , eg , speech recognition . | convolutional neural networks have recently been shown to perform well on large scale visual recognition tasks . |
a reasonable approximate can be acquired using variational approximation which is shown to work reasonably well in various applications . | a reasonable approximate can be acquired using variational approximation , which is shown to work reasonably well in various applications . |
this non perturbative information happens to be equivalent to the knowledge of the exact monodromy group in the rigid case . | moreover , this information is totaly equivalent to the knowledge of the second monodromy generator of the rigid theory . |
this stabilizer is a lie group , of dimension 14 , called g2 . | this stabilizer is a subgroup gi subgroup gi are symplectic reflections . |
additionally , if the manifold m is compact and admits a ricci flat metric , then its first chern class must vanish and the manifold is called calabi-yau . | then y is not a calabi-yau , but it is a transverse calabi-yau . |
the 1365-orbit stabilizer in m is a maximal parabolic subgroup p of m . | the stabilizer is the spin symmetry lifting a possible momentum degeneracy . |
the generator consists of 6 basic residual blocks between two down-sample layers and two upsample layers , assembled by instance normalization . | the content encoder is made of several 2d convolutional layers followed by several residual blocks . |
deep network based language learning has received great success recently and has been applied in different applications , for example , machine translation . | recently , neural networks have achieved very impressive success on a wide range of fields like computer vision . |
recently , deep learning methods have shown great success in various computer vision tasks , such as object recognition , object detection , and image classification . | in deep learning for computer vision , object classification and other tasks have improved performance for image understanding . |
deep convolutional neural networks have made significant breakthroughs in many visual understanding tasks including image classification . | deep neural networks have had great success in learning to predict various quantities from images , eg , object classes . |
lastly , user pairing and the performance of noma has been also studied from an information theory perspective , as discussed in , to mimo systems . | user pairing and the performance of noma have been also studied from an information theory perspective , as discussed in for the mimo bc . |
the computational costs for uniformization are similar in structure to those of direct sampling . | rejection sampling requires less cpu time than direct sampling and uniformization . |
due to the requirement of a robust solution , compared to iterative methods are often favored for vlsi circuit simulation , and thus adopted by state-of-the-art power grid solvers in tau pg simulation contest . | due to the requirement of a robust solution , compared to iterative matrix solvers are often favored for vlsi circuit simulation , and thus adopted by state-of-the-art power grid solvers in tau pg simulation contest . |
at tc the quantum correlations are entirely vanishing meanwhile the classical ones remain finite for all temperatures . | as expected , at tc the quantum correlations are zero meanwhile the classical ones remain finite . |
the diameter is the maximum distance between two vertices and the girth is the length of a shortest circuit . | here the diameter is the average distance between a connected pair of vertices along the shortest pathways . |
based on the original analytic definition in , an algebraic definition of k -- stability was introduced in . | the notion of k-stability was later generalized to a more algebraic formulation in . |
for all methods , we use an imagenet-pretrained resnet-18 architecture as a feature extractor . | we achieve the best results with a resnet-101 as the underlying feature extractor . |
characterize those group rings znsm which are s-right multiplication ideal ring . | characterize those s-semigroup rings which are s-artinian . |
ogawa et al proposed rules to identify and remove the least influential data in order to reduce the computation cost when training support vector machines . | ogawa et al proposed rules to identify and remove the least influential data when training support vector machines to reduce the computation cost . |
it turns out to be related to morita-millermumford classes in the cohomology of moduli spaces of algebraic curves , cf . | however , the odd case is also quite interesting and apparently related to the homology of moduli spaces of algebraic curves , cf . |
the sparse representation problem has wide applicability , for example , in communications , and compressive sampling . | the sparse signal estimation problem has many applications , for example , in linear regression , and recently in compressive sampling . |
therefore the power spectrum is the superposition of lorenzians weighted by the same gaussian function introduced to describe the polidispersity in the correlation analysis . | the power spectrum is the legendre transform of the two-point correlation function , and is more commonly encountered for individually for the each wmap frequency band , theoretical predictions . |
while neural networks are known to be robust to random noise , they have been shown to be vulnerable to adversarially-crafted perturbations . | although convolutional neural networks have been largely successful in various applications , they have been shown to be quite vulnerable to additive adversarial perturbations . |
this phenomena is the brane version of the field identification fixed points . | irreversible phenomena that contribute to an entropy production are called dissipative phenomena . |
in recent years , deep learning techniques have achieved profound breakthroughs in many computer vision applications , including the classification of natural and medical images . | convolutional neural networks have achieved state-of-the-art performance on visual tasks such as image and video recognition in the last few years . |
the network was trained with the adam optimiser for a maximum of 40 epochs . | the model was trained for 3 epochs using the adam optimizer . |
the atca data was calibrated using the miriad software package . | the data reduction and image analysis were done with the miriad data reduction tools . |
the second approach applies post-processing in case of bp decoding failure . | in the second approach , a post-processing is used in case bp decoding fails . |
it remains an open question whether instance-dependent noise may be included into our approach . | it remains an open question whether more realistic instance-dependent noise may modelled within our approach . |
bertsimas and shioda provide a mixed integer quadratic programming formulation for linear regression with a cardinality constraint . | bertsimas and shioda provide a mixed integer quadratic programming formulation with a cardinality constraint for linear regression . |
this indicates that the spontaneous recoil does not completely wash out the interference . | indeed , the averaging over the spontaneous recoil does not destroy the interference pattern . |
we state by thatf 1 , f 2 l has only result equality with f 1 and by that it has both result and effect equality with f 2 . | we state by that f 1 , f 2 l has only result equality with f 1 and by that it has both result and effect equality with f 2 . |
in his seminal work , shannon showed that separation of source and channel coding is optimal for discrete memoryless point-to-point channels . | in his 1948 seminal work , shannon showed that the separation between source and channel coding is optimal for point-to-point communication systems . |
in 2006 , ryu and takayanagi proposed that the quantum entanglement entropy can be directly obtained from minimal surface . | in particular , ryu and takayanagi proposed a direct connection between the entanglement entropy of a cft to a dual bulk geometry . |
to solve this problem , he et al developed a new structure called deep residual learning , which adds a shortcut from the formal layer to the trained layer . | to overcome these problems , he et al proposed a residual learning technique to ease the training of networks that enables them to be substantially deeper . |
once the final topic set was established , we used a focused web crawler to download all web pages returned by google for each topic . | once the final topic set is established , we use a focused web crawler , to download all web pages returned by google for each topic . |
treleani , double parton scatterings in high-energy hadronic collisions , nucl . | treleani , double parton scatterings in b-quark pairs production at the lhc , phys . |
huang et al proposed a dense network with several densely connected blocks , which make the receptive filed of prediction more dense . | huang et al proposed the densely connected networks , in which each layer connects with all its previous layers . |
deep neural networks have achieved impressive performance on tasks across a variety of domains , including vision . | deep convolutional neural networks have revolutionized computer vision , achieving unprecedented performance in high-level vision tasks such as classification . |
here a multiset is a collection in which the order of the entries does not matter , but multiplicities do . | the multiset m is a set and has strength 3 4 proof . |
the geometry we consider is a one-dimensional ring coupled to a bubble . | the geometry associated with the scalars in this model is known as special geometry . |
the neural network architectures include densenet121 and resnet34 . | the network architectures include vgg11 and two other shallow networks , simple and simpler . |
thus , si-si bond length and si-si-si bond angle distributions are almost independent of preparation procedure and the initial structure from which the hydrogenated sample is generated . | the si-si bond length and si-si-si bond angle distributions are nearly independent of sample preparation procedure , but si-h bond length distributions are sample dependent . |
convolutional neural networks have recently exhibited great performance in various fields such as computer vision . | the effectiveness of the deep convolutional neural networks has been demonstrated for various computer vision tasks such as image classification and so on . |
hence , gulrajani et al proposed penalizing the gradient norm to enforce lipschitz constraint instead of clipping . | in order to improve the training stability of gan , arjovsky et al further introduce gradient penalty for enforcing the lipschitz constraint in discriminator . |
in the same spirit , the fal method based on a fractional steepest descent approach was proposed in . | in the same spirit , a fractional adaptive learning method based on a fractional steepest descent approach was proposed in . |
we use the adam optimizer and apply dropout at all lstm layers . | we use the adam optimizer with the default parameters and minibatches of size 16 . |
the data were reduced using the common astronomy software application package with a standard pipeline . | the mwa data were calibrated with the common astronomy software applications , using observations of a calibrator source . |
machine learning approaches , especially deep neural networks , are transforming a wide range of application domains , such as computer vision . | deep learning , especially convolutional neural networks , has revolutionized various machine learning tasks with grid-like input data , such as image classification . |
a considerable amount of research has been devoted to exploiting multiple antennas for providing communication secrecy . | recently , considerable research has investigated secrecy in wiretap channels with multiple antennas . |
we use the sample-weighted implementation of logistic regression in scikit-learn as the base classifier , to compare the effect of the reduction approach . | for non-private logistic regression , we have used the prepackaged scikit-learn logistic regression classifier . |
the visibility of the stone strctures has been adjusted with image processing software . | the visibility of the stone structures has been adjusted with image processing software . |
deep neural networks have achieved outstanding performance on prediction tasks like visual object and speech recognition . | deep learning has witnessed great success in image recognition using convolutional neural networks and has been widely explored in neuroimaging field . |
to enrich training samples , some extra images can be generated by gans-based methods , which can reduce the model over-fitting . | besides , to avoid the model over-fitting , some extra images are generated by gans-based methods to enhance the diversity of training data . |
in the present study , we address a model , where the structural obstacles of the environment are spatially correlated on a mesoscopic scale . | here , however , we address a case where the structural obstacles of the environment are spatially correlated on a mesoscopic scale . |
in the present paper we have established a direct connection between the more detailed polymerization model of sandars and the simpler model equation approach with only two ordinary differential equations . | in the present paper we have modified the polymerization model of sandars such that polymerization is only possible on one of the two ends of the polymer . |
wyner proposed the wiretap channel model as a basic framework for pls . | these ideas were later developed by wyner , in which he introduced the wiretap channel . |
we used the idl mpfit routine , which performs weighted least-squares curve fitting of the data taking into account the noise for each spectral band . | we performed maximum likelihood fits to the data using the levenbergmarquardt algorithm mpfit , a local minimization routine . |
if the section is also flat in the induced metric , then the action is called hyperpolar . | an action admitting a section that is flat in the induced metric is called hyperpolar . |
the network is trained with the adam optimizer , a mini-batch size of 64 and early stopping . | the network was trained with the adam optimiser for a maximum of 40 epochs . |
now we proceed on to define smarandache semivector spaces . | in section two we introduce bivector spaces and give its smarandache analogues . |
for the discriminator , we use a network architecture similar to patchgan , and with wn and trelu . | for the discriminator network , we employ the fully convolutional architecture used in , which can be applied to arbitrarily-sized images . |
to make good use of global image-level priors , zhao et al proposed to use the pyramid pooling module to collet levels of information from multiple scales . | zhao et al exploit global context information by different region-based context aggregation through pyramid pooling . |
if a bi-algebra has in addition unit , counit and antipode operators , it is called a hopf algebra . | the hopf algebra a denotes the tensor algebra of a graded q-vector space v . |
the parameters are initialized using the method proposed by he et al and optimized with adam . | the weight parameters were initialized by using the method proposed by he et al for all experiments . |
these objects are expected to have formed during phase transitions in the early universe through spontaneous symmetry breakings . | topological defects are believed to have formed in the numerous phase transitions in the early universe due to the kibble mechanism . |
since the free energy is a state variable , in the thermodynamic limit it should not exhibit hysteresis , and our results raise qualms about the conventional assumption of local thermodynamic equilibrium in most continuum modeling . | thus , even if the solvation free energy is a significant fraction of the overall surface energy , it is unlikely to contribute significantly to the relative stability of these surfaces . |
d eep neural networks have advanced to show the state-of-the-art performance in many of computer vision applications , such as image classification . | recently deep neural networks have attained impressive performance in many fields such as image classification . |
recently , comparison-based settings have become increasingly popular , schultz and joachims , 2003 , agarwal et al , 2007 , van der maaten and weinberger , 2012 , amid and ukkonen , 2015 , balcan et al , 2016 . | the comparison-based setting has recently become popular in machine learning literature , schultz and joachims , 2003 , agarwal et al , 2007 , van der maaten and weinberger , 2012 , amid and ukkonen , 2015 , balcan et al , 2016 . |
the latter theory is well known to appear in another duality relating it to type iia string theory on k3 . | this can also be seen by noting that this theory is dual to the type iia string compactified on k3 which has no chern-simons terms . |
in mammalian autophagy , peroxisomes are targeted for degradation by exogenous ubiquitin labeling . | in mammalian autophagy , peroxisomes are targeted for degradation by exogenous ubiquitin labelling . |
deep neural networks have defined the state-of-the-art in a wide range of problems in computer vision , speech analysis , and natural language processing . | neural networks proved to be extremely effective in solving a broad variety of problems in computer vision , speech recognition and natural language processing . |
these different descriptions must provide the same physical information of the zero modes if we restrict to the adhm group invariant quantities . | it is therefore natural to describe the zero modes in terms of the adhm group invariant quantities . |
events are reconstructed with the particle flow algorithm , which combines information from the subdetectors to optimize reconstruction and identification of produced stable particles , namely charged and neutral hadrons , photons , electrons , and muons . | particle candidates in cms are reconstructed using a particle-flow algorithm , which identifies muons , electrons , photons , and neutral and charged hadrons through a combination of information from the various subdetectors . |
indeed , finding a global minimum of a general non-convex function and training certain types of neural networks are both np-hard . | finding a global minimum of a general nonconvex function is np-hard , and nonconvex optimization to train certain types of neural networks is also known to be np-hard . |
deep neural networks have been found to be quite effective for solving problems in the domain of computer vision . | deep neural networks have been significantly successful in many artificial intelligence tasks such as im- age classification . |
the heterostructure consists of a fe halfspace and a sc of fe and 2 ml of sc are determined selfconsistently . | the heterostructure consists of superconducting , s , and ferromagnetic , f , layers . |
we will illustrate this situation by some simple examples . | we will first give some examples of this definition . |
Subsets and Splits