Searcharxiv⌕ Search

arXiv subjects

Xin Gao

Publications and source records attributed to Xin Gao.

At least 289 records · Page 16Linked to original sources

Semi-Supervised Sparse Coding

Sparse coding approximates the data sample as a sparse linear combination of some basic codewords and uses the sparse codes as new presentations. In this paper, we investigate learning discriminative sparse codes by sparse coding in a semi-supervised manner, where only a few training samples are labeled. By using the manifold structure spanned by the data set of both labeled and unlabeled samples and the constraints provided by the labels of the labeled samples, we learn the variable class labels for all the samples. Furthermore, to improve the discriminative ability of the learned sparse codes, we assume that the class labels could be predicted from the sparse codes directly using a linear classifier. By solving the codebook, sparse codes, class labels and classifier parameters simultaneously in a unified objective function, we develop a semi-supervised sparse coding algorithm. Experiments on two real-world pattern recognition problems demonstrate the advantage of the proposed methods over supervised sparse coding methods on partially labeled data sets.

stat.ML↗

A Tag Identification Approach Based On Fragile Watermark

This paper proposes a tag identify approach based on fragile Watermark that based on Least significant bit of the replacement that we first use a special way to initialize the cover to ensure that we can use random positions to embed the information of tag. Using this way enhance the security of other to get the right information of this tag. Finally as long as the covered information can be decoded, the completeness and accuracy of the tag information can be guaranteed. the result of simulation experiment show that this approach has high sensitivity and security .

cs.MM↗

Fractional chaotic inflation in the lights of PLANCK and BICEP2

In the lights of current BICEP2 observations accompanied with the PLANCK satellite results, it has been observed that the simple single field chaotic inflationary models provide a good agreement with their spectral index n_s and large tensor-to-scalar ratio r (0.15 <r <0.26). To explore the other simple models, we consider the fractional-chaotic inflationary potentials of the form V_0 phi^(a/b) where a and b are relatively prime. We show that such kind of inflaton potentials can be realized elegantly in the supergravity framework with generalized shift symmetry and a nature bound a/b < 4 for consistency. Especially, for the number of e-folding from 50 to 60 and some a/b from 2 to 3, our predictions are nicely within at least 1 $σ$ region in the r-n_s plane. We also present a systematic investigation of such chaotic inflationary models with fractional exponents to explore the possibilities for the enhancement in the magnitude of running of spectral index (α_{n_s}) beyond the simplistic models.

hep-ph↗

Learning manifold to regularize nonnegative matrix factorization

Inthischapterwediscusshowtolearnanoptimalmanifoldpresentationto regularize nonegative matrix factorization (NMF) for data representation problems. NMF,whichtriestorepresentanonnegativedatamatrixasaproductoftwolowrank nonnegative matrices, has been a popular method for data representation due to its ability to explore the latent part-based structure of data. Recent study shows that lots of data distributions have manifold structures, and we should respect the manifold structure when the data are represented. Recently, manifold regularized NMF used a nearest neighbor graph to regulate the learning of factorization parameter matrices and has shown its advantage over traditional NMF methods for data representation problems. However, how to construct an optimal graph to present the manifold prop- erly remains a difficultproblem due to the graph modelselection, noisy features, and nonlinear distributed data. In this chapter, we introduce three effective methods to solve these problems of graph construction for manifold regularized NMF. Multiple graph learning is proposed to solve the problem of graph model selection, adaptive graph learning via feature selection is proposed to solve the problem of constructing a graph from noisy features, while multi-kernel learning-based graph construction is used to solve the problem of learning a graph from nonlinearly distributed data.

cs.LG↗

Combining Universal and Odd RR Axions for Aligned Natural Inflation

We successfully embed the Kim-Nilles-Peloso (KNP) alignment mechanism for enhancing the axion decay constant in the context of large volume type IIB orientifolds. The flat direction is generated in the plane of ($C_0$-$C_2$) axions corresponding to the involutively even universal axion $C_0$ and odd axion $C_2$, respectively. The moduli stabilization with large volume scheme has been established as well.

hep-th↗

Cosmological observables in multi-field inflation with a non-flat field space

Using $δN$ formalism, in the context of a generic multi-field inflation driven on a non-flat field space background, we revisit the analytic expressions of the various cosmological observables such as scalar/tensor power spectra, scalar/tensor spectral tilts, non-Gaussianity parameters, tensor-to-scalar ratio, and the various runnings of these observables. In our backward formalism approach, the subsequent expressions of observables automatically include terms beyond the leading order slow-roll expansion correcting many of the expression at subleading order. To connect our analysis properly with the earlier results, we rederive the (well) known (single field) expressions in the limiting cases of our generic formulae. Further, in the light of PLANCK results, we examine for the compatibility of the consistency relations within the slow-roll regime of a two-field roulette poly-instanton inflation realized in the context of large volume scenarios.

hep-th↗

Maximum mutual information regularized classification

In this paper, a novel pattern classification approach is proposed by regularizing the classifier learning to maximize mutual information between the classification response and the true class label. We argue that, with the learned classifier, the uncertainty of the true class label of a data sample should be reduced by knowing its classification response as much as possible. The reduced uncertainty is measured by the mutual information between the classification response and the true class label. To this end, when learning a linear classifier, we propose to maximize the mutual information between classification responses and true class labels of training samples, besides minimizing the classification error and reduc- ing the classifier complexity. An objective function is constructed by modeling mutual information with entropy estimation, and it is optimized by a gradi- ent descend method in an iterative algorithm. Experiments on two real world pattern classification problems show the significant improvements achieved by maximum mutual information regularization.

cs.LG↗

Normal modes and time evolution of a holographic superconductor after a quantum quench

We employ holographic techniques to investigate the dynamics of the order parameter of a strongly coupled superconductor after a perturbation that drives the system out of equilibrium. The gravity dual that we employ is the ${\rm AdS}_5$ Soliton background at zero temperature. We first analyze the normal modes associated to the superconducting order parameter which are purely real since the background has no horizon. We then study the full time evolution of the order parameter after a quench. For sufficiently a weak and slow perturbation we show that the order parameter undergoes simple undamped oscillations in time with a frequency that agrees with the lowest normal model computed previously. This is expected as the soliton background has no horizon and therefore, at least in the probe and large $N$ limits considered, the system will never return to equilibrium. For stronger and more abrupt perturbations higher normal modes are excited and the pattern of oscillations becomes increasingly intricate. We identify a range of parameters for which the time evolution of the order parameter become quasi chaotic. The details of the chaotic evolution depend on the type of perturbation used. Therefore it is plausible to expect that it is possible to engineer a perturbation that leads to the almost complete destruction of the oscillating pattern and consequently to quasi equilibration induced by superposition of modes with different frequencies.

hep-th↗

Large Margin Image Set Representation and Classification

In this paper, we propose a novel image set representation and classification method by maximizing the margin of image sets. The margin of an image set is defined as the difference of the distance to its nearest image set from different classes and the distance to its nearest image set of the same class. By modeling the image sets by using both their image samples and their affine hull models, and maximizing the margins of the images sets, the image set representation parameter learning problem is formulated as an minimization problem, which is further optimized by an expectation -maximization (EM) strategy with accelerated proximal gradient (APG) optimization in an iterative algorithm. To classify a given test image set, we assign it to the class which could provide the largest margin. Experiments on two applications of video-sequence-based face recognition demonstrate that the proposed method significantly outperforms state-of-the-art image set classification methods in terms of both effectiveness and efficiency.

cs.CV↗

On Classifying the Divisor Involutions in Calabi-Yau Threefolds

In order to support the odd moduli in models of (type IIB) string compactification, we classify the Calabi-Yau threefolds with h^{1,1}<=4 which exhibit pairs of identical divisors, with different line-bundle charges, mapping to each other under possible divisor exchange involutions. For this purpose, the divisors of interest are identified as completely rigid surface, Wilson surface, K3 surface and some other deformation surfaces. Subsequently, various possible exchange involutions are examined under the symmetry of Stanley-Reisner Ideal. In addition, we search for the Calabi-Yau theefolds which contain a divisor with several disjoint components. Under certain reflection involution, such spaces also have nontrivial odd components in (1,1)-cohomology class. String compactifications on such Calabi-Yau orientifolds with non-zero h^{1,1}_-(CY_3/σ) could be promising for concrete model building in both particle physics and cosmology. In the spirit of using such Calabi-Yau orientifolds in the context of LARGE volume scenario, we also present some concrete examples of (strong/weak) swiss-cheese type volume form.

hep-th↗

F-term Stabilization of Odd Axions in LARGE Volume Scenario

In the context of the LARGE volume scenario, stabilization of axionic moduli is revisited. This includes both even and odd axions with their scalar potential being generated by F-term contributions via various tree-level and non-perturbative effects like fluxed E3-brane instantons and fluxed poly-instantons. In all the cases, we estimate the decay constants and masses of the axions involved.

hep-th↗

Competition between the s-wave and p-wave superconductivity phases in a holographic model

We build a holographic superconductor model with a scalar triplet charged under an SU(2) gauge field in the bulk. In this model, the s-wave and p-wave condensates can be consistently realized. We find that there are totally four phases in this model, namely, the normal phase without any condensate, s-wave phase, p-wave phase and the s+p coexisting phase. By calculating Gibbs free energy, the s+p coexisting phase turns out to be thermodynamically favored once it can appear. The phase diagram with the dimension of the scalar operator and temperature is drawn. The temperature range for the s+p coexisting phase is very narrow, which shows the competition between the s-wave and p-wave orders in the superconductor model.

hep-th↗

Dimensional Oxidation of Non-geometric Fluxes in Type II Orientifolds

Some aspects of string compactifications with non-geometric fluxes are revisited in the light of recent progress in double field theory. After rederiving the general form of these fluxes, we consider the proposed flux induced four-dimensional effective superpotential and oxidize its induced scalar potential to terms in a ten-dimensional action. This analysis is performed independently for an explicit toroidal type IIA and its T-dual type IIB orientifold. We show in detail that the result of this bottom-up approach is compatible with the gauged supergravity motivated flux formulation of the double field theory action in both the NS-NS and the R-R sector.

hep-th↗

Multiple graph regularized protein domain ranking

Background Protein domain ranking is a fundamental task in structural biology. Most protein domain ranking methods rely on the pairwise comparison of protein domains while neglecting the global manifold structure of the protein domain database. Recently, graph regularized ranking that exploits the global structure of the graph defined by the pairwise similarities has been proposed. However, the existing graph regularized ranking methods are very sensitive to the choice of the graph model and parameters, and this remains a difficult problem for most of the protein domain ranking methods. Results To tackle this problem, we have developed the Multiple Graph regularized Ranking algorithm, MultiG- Rank. Instead of using a single graph to regularize the ranking scores, MultiG-Rank approximates the intrinsic manifold of protein domain distribution by combining multiple initial graphs for the regularization. Graph weights are learned with ranking scores jointly and automatically, by alternately minimizing an ob- jective function in an iterative algorithm. Experimental results on a subset of the ASTRAL SCOP protein domain database demonstrate that MultiG-Rank achieves a better ranking performance than single graph regularized ranking methods and pairwise similarity based ranking methods. Conclusion The problem of graph model and parameter selection in graph regularized protein domain ranking can be solved effectively by combining multiple graphs. This aspect of generalization introduces a new frontier in applying multiple graphs to solving protein domain ranking applications.

cs.LG↗

On Non-Gaussianities in Two-Field Poly-Instanton Inflation

In the context of Type IIB LARGE volume orientifold setup equipped with poly-instanton corrections, the standard single-field poly-instanton inflation driven by a 'Wilson' divisor volume modulus is generalized by the inclusion of respective axion modulus. This two-field dynamics results in a "Roulette" type inflation with the presence of several inflationary trajectories which could produce 50 (or more) e-foldings. The evolution of various trajectories along with physical observables are studied. The possibility of generating primordial non-Gaussianities in the slow-roll as well as in the beyond slow-roll region is investigated. We find that although the non- linearity parameters are quite small during the slow-roll regime, the same are significantly enhanced in the beyond slow-roll regime investigated up to the end of inflation.

hep-th↗

A Note on Poly-Instanton Effects in Type IIB Orientifolds on Calabi-Yau Threefolds

The zero mode structure for the generation of poly-instanton corrections for Euclidian D3-branes wrapping complex surfaces in Type IIB orientifolds with O7- and O3-planes is analyzed. Working examples of such surfaces and explicit embeddings into compact Calabi-Yau threefolds are presented, with special emphasis on geometries capable of realizing the LARGE volume scenario.

hep-th↗

Moduli Stabilization and Inflationary Cosmology with Poly-Instantons in Type IIB Orientifolds

Equipped with concrete examples of Type IIB orientifolds featuring poly-instanton corrections to the superpotential, the effects on moduli stabilization and inflationary cosmology are analyzed. Working in the framework of the LARGE volume scenario, the Kaehler modulus related to the size of the four-cycle supporting the poly-instanton contributes sub-dominantly to the scalar potential. It is shown that this Kaehler modulus gets stabilized and, by displacing it from its minimum, can play the role of an inflaton. Subsequent cosmological implications are discussed and compared to experimental data.

hep-th↗

Nonparametric Clustering of Mixed Data Using Modified Chi-square Tests

We propose a non-parametric method to cluster mixed data containing both continuous and discrete random variables. The product space of continuous and categorical sample spaces is approximated locally by analyzing neighborhoods with cluster patterns. Detection of cluster patterns on the product space is determined by using a modified Chi-square test. The proposed method does not impose a global distance function which could be difficult to specify in practice. Results from simulation studies have shown that our proposed methods out-performed the benchmark method, AutoClass, for various settings.

stat.ME↗