Searcharxiv⌕ Search

arXiv subjects

Xin Gao

Publications and source records attributed to Xin Gao.

At least 271 records · Page 15Linked to original sources

Predicting Discharge Medications at Admission Time Based on Deep Learning

Predicting discharge medications right after a patient being admitted is an important clinical decision, which provides physicians with guidance on what type of medication regimen to plan for and what possible changes on initial medication may occur during an inpatient stay. It also facilitates medication reconciliation process with easy detection of medication discrepancy at discharge time to improve patient safety. However, since the information available upon admission is limited and patients' condition may evolve during an inpatient stay, these predictions could be a difficult decision for physicians to make. In this work, we investigate how to leverage deep learning technologies to assist physicians in predicting discharge medications based on information documented in the admission note. We build a convolutional neural network which takes an admission note as input and predicts the medications placed on the patient at discharge time. Our method is able to distill semantic patterns from unstructured and noisy texts, and is capable of capturing the pharmacological correlations among medications. We evaluate our method on 25K patient visits and compare with 4 strong baselines. Our methods demonstrate a 20% increase in macro-averaged F1 score than the best baseline.

cs.CL↗

Observation of bulk nodal lines in topological semimetal ZrSiS

ZrSiS is the most intensively studied topological nodal-line semimetal candidate, which is proposed to host multiple nodal lines in its bulk electronic structure. However, previous angle-resolved photoemission spectroscopy (ARPES) experiments with vacuum ultraviolet lights mainly probed the surface states. Here using bulk-sensitive soft X-ray ARPES, we acquire the bulk electronic states of ZrSiS without any interference from surface states. Our results clearly show two groups of three-dimensional bulk nodal lines located on high-symmetry planes and along high-symmetry lines in the bulk Brillouin zone, respectively. The nodal lines on high-symmetry planes are enforced to pin at the Fermi level by carrier compensation and constitute the whole Fermi surfaces. This means that the carriers in ZrSiS are entirely contributed by nodal-line fermions, suggesting that ZrSiS is a remarkable platform for studying physical properties related to nodal lines.

cond-mat.mtrl-sci↗

Fibrations in CICY Threefolds

In this work we systematically enumerate genus one fibrations in the class of 7,890 Calabi-Yau manifolds defined as complete intersections in products of projective spaces, the so-called CICY threefolds. This survey is independent of the description of the manifolds and improves upon past approaches that probed only a particular algebraic form of the threefolds (i.e. searches for "obvious" genus one fibrations as in [1,2]). We also study K3-fibrations and nested fibration structures. That is, K3 fibrations with potentially many distinct elliptic fibrations. To accomplish this survey a number of new geometric tools are developed including a determination of the full topology of all CICY threefolds, including triple intersection numbers. In 2,946 cases this involves finding a new "favorable" description of the manifold in which all divisors descend from a simple ambient space. Our results consist of a survey of obvious fibrations for all CICY threefolds and a complete classification of all genus one fibrations for 4,957 "Kahler favorable" CICYs whose Kahler cones descend from a simple ambient space. Within the CICY dataset, we find 139,597 obvious genus one fibrations, 30,974 obvious K3 fibrations and 208,987 nested combinations. For the Kahler favorable geometries we find a complete classification of 377,559 genus one fibrations. For one manifold with Hodge numbers (19,19) we find an explicit description of an infinite number of distinct genus-one fibrations extending previous results for this particular geometry that have appeared in the literature. The data associated to this scan is available at http://www1.phys.vt.edu/cicydata .

hep-th↗

Robust Cost-Sensitive Learning for Recommendation with Implicit Feedback

Recommendation is the task of improving customer experience through personalized recommendation based on users' past feedback. In this paper, we investigate the most common scenario: the user-item (U-I) matrix of implicit feedback. Even though many recommendation approaches are designed based on implicit feedback, they attempt to project the U-I matrix into a low-rank latent space, which is a strict restriction that rarely holds in practice. In addition, although misclassification costs from imbalanced classes are significantly different, few methods take the cost of classification error into account. To address aforementioned issues, we propose a robust framework by decomposing the U-I matrix into two components: (1) a low-rank matrix that captures the common preference, and (2) a sparse matrix that detects the user-specific preference of individuals. A cost-sensitive learning model is embedded into the framework. Specifically, this model exploits different costs in the loss function for the observed and unobserved instances. We show that the resulting non-smooth convex objective can be optimized efficiently by an accelerated projected gradient method with closed-form solutions. Morever, the proposed algorithm can be scaled up to large-sized datasets after a relaxation. The theoretical result shows that even with a small fraction of 1's in the U-I matrix $M\in\mathbb{R}^{n\times m}$, the cost-sensitive error of the proposed model is upper bounded by $O(\fracα{\sqrt{mn}})$, where $α$ is a bias over imbalanced classes. Finally, empirical experiments are extensively carried out to evaluate the effectiveness of our proposed algorithm. Encouraging experimental results show that our algorithm outperforms several state-of-the-art algorithms on benchmark recommendation datasets.

cs.LG↗

Robust Online Multi-Task Learning with Correlative and Personalized Structures

Multi-Task Learning (MTL) can enhance a classifier's generalization performance by learning multiple related tasks simultaneously. Conventional MTL works under the offline or batch setting, and suffers from expensive training cost and poor scalability. To address such inefficiency issues, online learning techniques have been applied to solve MTL problems. However, most existing algorithms of online MTL constrain task relatedness into a presumed structure via a single weight matrix, which is a strict restriction that does not always hold in practice. In this paper, we propose a robust online MTL framework that overcomes this restriction by decomposing the weight matrix into two components: the first one captures the low-rank common structure among tasks via a nuclear norm and the second one identifies the personalized patterns of outlier tasks via a group lasso. Theoretical analysis shows the proposed algorithm can achieve a sub-linear regret with respect to the best linear model in hindsight. Even though the above framework achieves good performance, the nuclear norm that simply adds all nonzero singular values together may not be a good low-rank approximation. To improve the results, we use a log-determinant function as a non-convex rank approximation. The gradient scheme is applied to optimize log-determinant function and can obtain a closed-form solution for this refined problem. Experimental results on a number of real-world applications verify the efficacy of our method.

cs.LG↗

When coding meets ranking: A joint framework based on local learning

Sparse coding, which represents a data point as a sparse reconstruction code with regard to a dictionary, has been a popular data representation method. Meanwhile, in database retrieval problems, learning the ranking scores from data points plays an important role. Up to now, these two problems have always been considered separately, assuming that data coding and ranking are two independent and irrelevant problems. However, is there any internal relationship between sparse coding and ranking score learning? If yes, how to explore and make use of this internal relationship? In this paper, we try to answer these questions by developing the first joint sparse coding and ranking score learning algorithm. To explore the local distribution in the sparse code space, and also to bridge coding and ranking problems, we assume that in the neighborhood of each data point, the ranking scores can be approximated from the corresponding sparse codes by a local linear function. By considering the local approximation error of ranking scores, the reconstruction error and sparsity of sparse coding, and the query information provided by the user, we construct a unified objective function for learning of sparse codes, the dictionary and ranking scores. We further develop an iterative algorithm to solve this optimization problem.

cs.CV↗

Data Integration with High Dimensionality

We consider a problem of data integration. Consider determining which genes affect a disease. The genes, which we call predictor objects, can be measured in different experiments on the same individual. We address the question of finding which genes are predictors of disease by any of the experiments. Our formulation is more general. In a given data set, there are a fixed number of responses for each individual, which may include a mix of discrete, binary and continuous variables. There is also a class of predictor objects, which may differ within a subject depending on how the predictor object is measured, i.e., depend on the experiment. The goal is to select which predictor objects affect any of the responses, where the number of such informative predictor objects or features tends to infinity as sample size increases. There are marginal likelihoods for each way the predictor object is measured, i.e., for each experiment. We specify a pseudolikelihood combining the marginal likelihoods, and propose a pseudolikelihood information criterion. Under regularity conditions, we establish selection consistency for the pseudolikelihood information criterion with unbounded true model size, which includes a Bayesian information criterion with appropriate penalty term as a special case. Simulations indicate that data integration improves upon, sometimes dramatically, using only one of the data sources.

stat.ML↗

Multiple Fibrations in Calabi-Yau Geometry and String Dualities

In this work we explore the physics associated to Calabi-Yau (CY) n-folds that can be described as a fibration in more than one way. Beginning with F-theory vacua in various dimensions, we consider limits/dualities with M-theory, type IIA, and heterotic string theories. Our results include many M-/F-theory correspondences in which distinct F-theory vacua - associated to different elliptic fibrations of the same CY n-fold - give rise to the same M-theory limit in one dimension lower. Examples include 5-dimensional correspondences between 6-dimensional theories with Abelian, non-Abelian and superconformal structure, as well as examples of higher rank Mordell-Weil geometries. In addition, in the context of heterotic/F-theory duality, we investigate the role played by multiple K3- and elliptic fibrations in known and novel string dualities in 8-, 6- and 4-dimensional theories. Here we systematically summarize nested fibration structures and comment on the roles they play in T-duality, mirror symmetry, and 4-dimensional compactifications of F-theory with G-flux. This investigation of duality structures is made possible by geometric tools developed in a companion paper [1].

hep-th↗

Tools for CICYs in F-theory

We provide a set of tools for analyzing the geometry of elliptically fibered Calabi-Yau manifolds, starting with a description of the total space rather than with a Weierstrass model or a specified type of fiber/base. Such an approach to the subject of F-theory compactification makes certain geometric properties, which are usually hidden, manifest. Specifically, we review how to isolate genus-one fibrations in such geometries and then describe how to find their sections explicitly. This includes a full parameterization of the Mordell-Weil group where non-trivial. We then describe how to analyze the associated Weierstrass models, Jacobians and resolved geometries. We illustrate our discussion with concrete examples which are complete intersections in products of projective spaces (CICYs). The examples presented include cases exhibiting non-abelian symmetries and higher rank Mordell-Weil group. We also make some comments on non-flat fibrations in this context. In a companion paper [1] to this one, these results will be used to analyze the consequences for string dualities of the ubiquity of multiple fibrations in known constructions of Calabi-Yau manifolds.

hep-th↗

Approximate Bayesian estimation in large coloured graphical Gaussian models

Distributed estimation methods have recently been used to compute the maximum likelihood estimate of the precision matrix for large graphical Gaussian models. Our aim, in this paper, is to give a Bayesian estimate of the precision matrix for large graphical Gaussian models with, additionally, symmetry constraints imposed by an underlying graph which is coloured. We take the sample posterior mean of the precision matrix as our estimate. We study its asymptotic behaviour under the regular asymptotic regime when the number of variables p is fixed and under the double asymptotic regime when both p and n grow to infinity. We show in particular, that when the number of parameters of the local models is uniformly bounded, the standard convergence rate we obtain for the asymptotic consistency, in the Frobenius norm, of our estimate of the precision matrix compares well with the rates in the current literature for the maximum likelihood estimate.

stat.ME↗

On Instanton Superpotentials, Calabi-Yau Geometry, and Fibrations

In this paper we explore contributions to non-perturbative superpotentials arising from instantons wrapping effective divisors in smooth Calabi-Yau four-folds. We concentrate on the case of manifolds constructed as complete intersections in products of projective spaces (CICYs) or generalizations thereof (gCICYs). We systematically investigate the structure of the cone of effective (algebraic) divisors in the four-fold geometries and employ the same tools recently developed in arXiv:1507.03235 to construct more general instanton geometries than have previously been considered in the literature. We provide examples of instanton configurations on Calabi-Yau manifolds that are elliptically and $K3$-fibered and explore their consequences in the context of string dualities. The examples discussed include manifolds containing infinite families of divisors with arithmetic genus, $χ(D, \mathcal O_D)=1$ and superpotentials exhibiting modular symmetry.

hep-th↗

Phase transitions in a holographic s+p model with backreaction

In a previous paper (arXiv:1309.2204, JHEP 1311 (2013) 087), we present a holographic s+p superconductor model with a scalar triplet charged under an SU(2) gauge field in the bulk. We also study the competition and coexistence of the s-wave and p-wave orders in the probe limit. In this work we continue to study the model by considering the full back-reaction The model shows a rich phase structure and various condensate behaviors such as the "n-type" and "u-type" ones, which are also known as reentrant phase transitions in condensed matter physics. The phase transitions to the p-wave phase or s+p coexisting phase become first order in strong back-reaction cases. In these first order phase transitions, the free energy curve always forms a swallow tail shape, in which the unstable s+p solution can also play an important role. The phase diagrams of this model are given in terms of the dimension of the scalar order and the temperature in the cases of eight different values of the back reaction parameter, which show that the region for the s+p coexisting phase is enlarged with a small or medium back reaction parameter, but is reduced in the strong back-reaction cases.

hep-th↗

On Beurling's uncertainty principle

We generalise a result of Hedenmalm to show that if a function $f$ on $\mathbb{R}$ is such that $\int_{\mathbb{R}^2} \bigl|f(x) \, \hat f(y)\bigr| \,e^{λ\left|xy\right|} \,dx\,dy = O( (1-λ)^{-N} )$ as $λ\to 1-$, then $f$ is the product of a polynomial and a gaussian.

math.CA↗

A New Construction of Calabi-Yau Manifolds: Generalized CICYs

We present a generalization of the complete intersection in products of projective space (CICY) construction of Calabi-Yau manifolds. CICY three-folds and four-folds have been studied extensively in the physics literature. Their utility stems from the fact that they can be simply described in terms of a `configuration matrix', a matrix of integers from which many of the details of the geometries can be easily extracted. The generalization we present is to allow negative integers in the configuration matrices which were previously taken to have positive semi-definite entries. This broadening of the complete intersection construction leads to a larger class of Calabi-Yau manifolds than that considered in previous work, which nevertheless enjoys much of the same degree of calculational control. These new Calabi-Yau manifolds are complete intersections in (not necessarily Fano) ambient spaces with an effective anticanonical class. We find examples with topology distinct from any that has appeared in the literature to date. The new manifolds thus obtained have many interesting features. For example, they can have smaller Hodge numbers than ordinary CICYs and lead to many examples with elliptic and K3-fibration structures relevant to F-theory and string dualities.

hep-th↗

Bayesian precision matrix estimation for graphical Gaussian models with edge and vertex symmetries

Graphical Gaussian models with edge and vertex symmetries were introduced by \citet{HojLaur:2008} who also gave an algorithm to compute the maximum likelihood estimate of the precision matrix for such models. In this paper, we take a Bayesian approach to the estimation of the precision matrix. We consider only those models where the symmetry constraints are imposed on the precision matrix and which thus form a natural exponential family with the precision matrix as the canonical parameter. We first identify the Diaconis-Ylvisaker conjugate prior for these models and develop a scheme to sample from the prior and posterior distributions. We thus obtain estimates of the posterior mean of the precision matrix. Second, in order to verify the precision of our estimate, we derive the explicit analytic expression of the expected value of the precision matrix when the graph underlying our model is a tree, a complete graph on three vertices and a decomposable graph on four vertices with various symmetries. In those cases, we compare our estimates with the exact value of the mean of the prior distribution. We also verify the accuracy of our estimates of the posterior mean on simulated data for graphs with up to thirty vertices and various symmetries.

stat.ME↗

Multiple Comparisons using Composite Likelihood in Clustered Data

We study the problem of multiple hypothesis testing for multidimensional data when inter-correlations are present. The problem of multiple comparisons is common in many applications. When the data is multivariate and correlated, existing multiple comparisons procedures based on maximum likelihood estimation could be prohibitively computationally intensive. We propose to construct multiple comparisons procedures based on composite likelihood statistics. We focus on data arising in three ubiquitous cases: multivariate Gaussian, probit, and quadratic exponential models. To help practitioners assess the quality of our proposed methods, we assess their empirical performance via Monte Carlo simulations. It is shown that composite likelihood based procedures maintain good control of the familywise type I error rate in the presence of intra-cluster correlation, whereas ignoring the correlation leads to erratic performance. Using data arising from a diabetic nephropathy study, we show how our composite likelihood approach makes an otherwise intractable analysis possible.

math.ST↗

Dimensional oxidation and modular completion of non-geometric type IIB action

Utilizing a setup of type IIB superstring theory compactified on an orientifold of T^6/(Z2xZ2), we propose a modular invariant dimensional oxidation of the four-dimensional scalar potential. In the oxidized ten-dimensional supergravity action, the standard NS-NS and RR three form fluxes (H-, F-) as well as the non-geometric fluxes (Q-, P-) are found to nicely rearrange themselves to form generalized flux-combinations. As an application towards moduli stabilization, using the same S-duality invariant scalar potential, we examine the recently proposed No-Go theorem (in arXiv:1409.7075) about creating a mass-hierarchy between universal-axion and the dilaton relevant for axionic-inflation. Considering a two-field dynamics of universal axion and dilaton while assuming the other moduli/axions being stabilized, we find a part of the No-Go arguments to be quite robust even with the inclusion of non-geometric (Q-, P-) fluxes.

hep-th↗

Regularized maximum correntropy machine

In this paper we investigate the usage of regularized correntropy framework for learning of classifiers from noisy labels. The class label predictors learned by minimizing transitional loss functions are sensitive to the noisy and outlying labels of training samples, because the transitional loss functions are equally applied to all the samples. To solve this problem, we propose to learn the class label predictors by maximizing the correntropy between the predicted labels and the true labels of the training samples, under the regularized Maximum Correntropy Criteria (MCC) framework. Moreover, we regularize the predictor parameter to control the complexity of the predictor. The learning problem is formulated by an objective function considering the parameter regularization and MCC simultaneously. By optimizing the objective function alternately, we develop a novel predictor learning algorithm. The experiments on two chal- lenging pattern classification tasks show that it significantly outperforms the machines with transitional loss functions.

cs.LG↗