arXiv subjects

Yue Shi

Publications and source records attributed to Yue Shi.

90 records · Page 5Linked to original sources

arXiv2018

Data Security and Privacy Protection Data Security and Privacy Protection in Public Cloud

This paper discusses about the challenges, advantages and shortcomings of existing solutions in data security and privacy in public cloud computing. As in cloud computing, oceans of data will be stored. Data stored in public cloud would face both outside attacks and inside attacks since public cloud provider themselves are untrusted. Conventional encryption could be used for storage, however most data in cloud needs further computation. Decryption before computation will cause large overheads for data operation and lots of inconvenience. Thus, efficient methods to protect data security as well as privacy for large amount of data in cloud are necessary. In the paper, different mechanisms to protect data security and privacy in public cloud are discussed. A data security and privacy enabled multi-cloud architecture is proposed.

cs.CR↗

arXiv2018

Precision of finite-difference representation in 3D coordinate-space Hartree-Fock-Bogoliubov calculations

The current work plans to study the accuracy due to FD approximation to the 3D nuclear HFB problem. By (1) taking the wave functions solved in harmonic oscillator (HO) basis, (2) representing the HFB problem in coordinate space using FD method, the current work carefully evaluates the error due to box discretization by examining the deviation of the resulted HFB matrix, the total energies in the coordinate space, from those calculated with HO method, the latter of which is free from numerical error within its model configuration. To estimate how the error (given by the box discretization schemes suggested above) accumulates with self-consistent iterations, self-consistent HF and HFB calculations (with two-basis method) has been carried out for doubly magic nuclei, $^{40}$Ca, $^{132}$Sn, and $^{110}$Mo. The resulted total energies are compared with those of HO basis, and 3D coordinate space calculations in literatures. The analysis shows that, for grid spacing $\le$0.6\,fm, the off-diagonal elements of the resulted HFB matrix elements (M.E.) are extremely small ($<$1\,keV). The resulted quasi-particle (q.p.) spectra differ from those of HO calculations by a few keV. Self-consistent HF and HFB calculations within the current FD method with the above box discretizatioin schemes give results similar to those calculations of existing HO basis, and coordinate space method.

nucl-th↗

arXiv2017

Joint Text Embedding for Personalized Content-based Recommendation

Learning a good representation of text is key to many recommendation applications. Examples include news recommendation where texts to be recommended are constantly published everyday. However, most existing recommendation techniques, such as matrix factorization based methods, mainly rely on interaction histories to learn representations of items. While latent factors of items can be learned effectively from user interaction data, in many cases, such data is not available, especially for newly emerged items. In this work, we aim to address the problem of personalized recommendation for completely new items with text information available. We cast the problem as a personalized text ranking problem and propose a general framework that combines text embedding with personalized recommendation. Users and textual content are embedded into latent feature space. The text embedding function can be learned end-to-end by predicting user interactions with items. To alleviate sparsity in interaction data, and leverage large amount of text data with little or no user interactions, we further propose a joint text embedding model that incorporates unsupervised text embedding with a combination module. Experimental results show that our model can significantly improve the effectiveness of recommendation systems on real-world datasets.

cs.IR↗

arXiv2017

On Sampling Strategies for Neural Network-based Collaborative Filtering

Recent advances in neural networks have inspired people to design hybrid recommendation algorithms that can incorporate both (1) user-item interaction information and (2) content information including image, audio, and text. Despite their promising results, neural network-based recommendation algorithms pose extensive computational costs, making it challenging to scale and improve upon. In this paper, we propose a general neural network-based recommendation framework, which subsumes several existing state-of-the-art recommendation algorithms, and address the efficiency issue by investigating sampling strategies in the stochastic gradient descent training for the framework. We tackle this issue by first establishing a connection between the loss functions and the user-item interaction bipartite graph, where the loss function terms are defined on links while major computation burdens are located at nodes. We call this type of loss functions "graph-based" loss functions, for which varied mini-batch sampling strategies can have different computational costs. Based on the insight, three novel sampling strategies are proposed, which can significantly improve the training efficiency of the proposed framework (up to $\times 30$ times speedup in our experiments), as well as improving the recommendation performance. Theoretical analysis is also provided for both the computational cost and the convergence. We believe the study of sampling strategies have further implications on general graph-based loss functions, and would also enable more research under the neural network-based recommendation framework.

cs.LG↗

arXiv2016

The Apps You Use Bring The Blogs to Follow

We tackle the blog recommendation problem in Tumblr for mobile users in this paper. Blog recommendation is challenging since most mobile users would suffer from the cold start when there are only a limited number of blogs followed by the user. Specifically to address this problem in the mobile domain, we take into account mobile apps, which typically provide rich information from the users. Based on the assumption that the user interests can be reflected from their app usage patterns, we propose to exploit the app usage data for improving blog recommendation. Building on the state-of-the-art recommendation framework, Factorization Machines (FM), we implement app-based FM that integrates app usage data with the user-blog follow relations. In this approach the blog recommendation is generated not only based on the blogs that the user followed before, but also the apps that the user has often used. We demonstrate in a series of experiments that app-based FM can outperform other alternative approaches to a significant extent. Our experimental results also show that exploiting app usage information is particularly effective for improving blog recommendation quality for cold start users.

cs.IR↗

arXiv2016

Learning Optimal Card Ranking from Query Reformulation

Mobile search has recently been shown to be the major contributor to the growing search market. The key difference between mobile search and desktop search is that information presentation is limited to the screen space of the mobile device. Thus, major search engines have adopted a new type of search result presentation, known as \textit{information cards}, in which each card presents summarized results from one domain/vertical, for a given query, to augment the standard blue-links search results. While it has been widely acknowledged that information cards are particularly suited to mobile user experience, it is also challenging to optimize such result sets. Typically, user engagement metrics like query reformulation are based on whole ranked list of cards for each query and most traditional learning to rank algorithms require per-item relevance labels. In this paper, we investigate the possibility of interpreting query reformulation into effective relevance labels for query-card pairs. We inherit the concept of conventional learning-to-rank, and propose pointwise, pairwise and listwise interpretations for query reformulation. In addition, we propose a learning-to-label strategy that learns the contribution of each card, with respect to a query, where such contributions can be used as labels for training card ranking models. We utilize a state-of-the-art ranking model and demonstrate the effectiveness of proposed mechanisms on a large-scale mobile data from a major search engine, showing that models trained from labels derived from user engagement can significantly outperform ones trained from human judgment labels.

cs.IR↗

arXiv2016

Learning Subclass Representations for Visually-varied Image Classification

In this paper, we present a subclass-representation approach that predicts the probability of a social image belonging to one particular class. We explore the co-occurrence of user-contributed tags to find subclasses with a strong connection to the top level class. We then project each image on to the resulting subclass space to generate a subclass representation for the image. The novelty of the approach is that subclass representations make use of not only the content of the photos themselves, but also information on the co-occurrence of their tags, which determines membership in both subclasses and top-level classes. The novelty is also that the images are classified into smaller classes, which have a chance of being more visually stable and easier to model. These subclasses are used as a latent space and images are represented in this space by their probability of relatedness to all of the subclasses. In contrast to approaches directly modeling each top-level class based on the image content, the proposed method can exploit more information for visually diverse classes. The approach is evaluated on a set of $2$ million photos with 10 classes, released by the Multimedia 2013 Yahoo! Large-scale Flickr-tag Image Classification Grand Challenge. Experiments show that the proposed system delivers sound performance for visually diverse classes compared with methods that directly model top classes.

cs.MM↗

arXiv2015

Properties of nuclei in the nobelium region studied within the covariant, Skyrme, and Gogny energy density functionals

We calculate properties of the ground and excited states of nuclei in the nobelium region for proton and neutron numbers of 92 <= Z <= 104 and 144 <= N <= 156, respectively. We use three different energy-density-functional (EDF) approaches, based on covariant, Skyrme, and Gogny functionals, each within two different parameter sets. A comparative analysis of the results obtained for odd-even mass staggerings, quasiparticle spectra, and moments of inertia allows us to identify single-particle and shell effects that are characteristic to these different models and to illustrate possible systematic uncertainties related to using the EDF modelling

nucl-th↗

arXiv2015

Theoretical study of triaxial shapes of neutron-rich Mo and Ru nuclei

Background: Recently, transition quadrupole moments in rotational bands of even-mass neutron-rich isotopes of molybdenum and ruthenium nuclei have been measured. The new data have provided a challenge for theoretical descriptions invoking stable triaxial deformations. Purpose: To understand experimental data on rotational bands in the neutron-rich Mo-Ru region, we carried out theoretical analysis of moments of inertia, shapes, and transition quadrupole moments of neutron-rich even-even nuclei around $^{110}$Ru using self-consistent mean-field and shell model techniques. Methods: To describe yrast structures in Mo and Ru isotopes, we use nuclear Density Functional Theory (DFT) with the optimized energy density functional UNEDF0. We also apply Triaxial Projected Shell Model (TPSM) to describe yrast and positive-parity, near-yrast band structures. Results: Our self-consistent DFT calculations predict triaxial ground-state deformations in $^{106,108}$Mo and $^{108.110,112}$Ru and reproduce the observed low-frequency behavior of moments of inertia. As the rotational frequency increases, a negative-$γ$ structure, associated with the aligned $ν(h_{11/2})^2$ pair, becomes energetically favored. The computed transition quadrupole moments vary with angular momentum, which reflects deformation changes with rotation; those variations are consistent with experiment. The TPSM calculations explain the observed band structures assuming stable triaxial shapes. Conclusions: The structure of neutron-rich even-even nuclei around $^{110}$Ru is consistent with triaxial shape deformations. Our DFT and TPSM frameworks provide a consistent and complementary description of experimental data.

nucl-th↗

arXiv2014

Adaptive Multi-resolution 3D Hartree-Fock-Bogoliubov Solver for Nuclear Structure

Complex many-body systems, such as triaxial and reflection-asymmetric nuclei, weakly-bound halo states, cluster configurations, nuclear fragments produced in heavy-ion fusion reactions, cold Fermi gases, and pasta phases in neutron star crust, they are all characterized by large sizes and complex topologies, in which many geometrical symmetries characteristic of ground-state configurations are broken. A tool of choice to study such complex forms of matter is an adaptive multi-resolution wavelet analysis. This method has generated much excitement since it provides a common framework linking many diversified methodologies across different fields, including signal processing, data compression, harmonic analysis and operator theory, fractals, and quantum field theory. To describe complex superfluid many-fermion systems, we introduce an adaptive pseudo-spectral method for solving self-consistent equations of nuclear density functional theory in three dimensions, without symmetry restrictions. The new adaptive multi-resolution Hartree-Fock-Bogoliubov (HFB) solver {\madnesshfb} is benchmarked against a two-dimensional coordinate-space solver {\hfbax} based on B-spline technique and three-dimensional solver {\hfodd} based on the harmonic oscillator basis expansion. Several examples are considered, including self-consistent HFB problem for spin-polarized trapped cold fermions and Skyrme-Hartree-Fock (+BCS) problem for triaxial deformed nuclei. The new {\madnesshfb} framework has many attractive features when applied to nuclear and atomic problems involving many-particle superfluid systems. Of particular interest are weakly-bound nuclear configurations close to particle drip lines, strongly elongated and dinuclear configurations such as those present in fission and heavy ion fusion, and exotic pasta phases that appear in the neutron star crust.

nucl-th↗

arXiv2014

Structure of Superheavy Nuclei Along Element 115 Decay Chains

A recent high-resolution $α$, $X$-ray, and $γ$-ray coincidence-spectroscopy experiment offered first glimpse of excitation schemes of isotopes along $α$-decay chains of $Z=115$. To understand these observations and to make predictions about shell structure of superheavy nuclei below $^{288}115$, we employ two complementary mean-field models: self-consistent Skyrme Energy Density Functional approach and the macroscopic-microscopic Nilsson model. We discuss the spectroscopic information carried by the new data. In particular, candidates for the experimentally observed $E1$ transitions in $^{276}$Mt are proposed. We find that the presence and nature of low-energy $E1$ transitions in well-deformed nuclei around $Z=110, N=168$ strongly depends on the strength of the spin-orbit coupling; hence, it provides an excellent constraint on theoretical models of superheavy nuclei. To clarify competing theoretical scenarios, an experimental search for $E1$ transitions in odd-$A$ systems $^{275,277}$Mt, $^{275}$Hs, and $^{277}$Ds is strongly recommended.

nucl-th↗

arXiv2014

Rotational properties of nuclei around $^{254}$No investigated using a spectroscopic-quality Skyrme energy density functional

Nuclei in the $Z\approx100$ mass region represent the heaviest systems where detailed spectroscopic information is experimentally available. Although microscopic-macroscopic and self-consistent models have achieved great success in describing the data in this mass region, a fully satisfying precise theoretical description is still missing. By using fine-tuned parametrizations of the energy density functionals, the present work aims at an improved description of the single-particle properties and rotational bands in the nobelium region. Such locally optimized parameterizations may have better properties when extrapolating towards the superheavy region. Skyrme-Hartree-Fock-Bogolyubov and Lipkin-Nogami methods were used to calculate the quasiparticle energies and rotational bands of nuclei in the nobelium region. Starting from the most recent Skyrme parametrization, UNEDF1, the spin-orbit coupling constants and pairing strengths have been tuned, so as to achieve a better agreement with the excitation spectra and odd-even mass differences in $^{251}$Cf and $^{249}$Bk. The quasiparticle properties of $^{251}$Cf and $^{249}$Bk were very well reproduced. At the same time, crucial deformed neutron and proton shell gaps open up at $N=152$ and $Z=100$, respectively. Rotational bands in Fm, No, and Rf isotopes, where experimental data are available, were also fairly well described. To help future improvements towards a more precise description, small deficiencies of the approach were carefully identified. In the $Z\approx100$ mass region, larger spin-orbit strengths than those from global adjustments lead to improved agreement with data. Puzzling effects of particle-number restoration on the calculated moment of inertia, at odds with the experimental behaviour, require further scrutiny.

nucl-th↗

arXiv2013

Exploiting Social Tags for Cross-Domain Collaborative Filtering

One of the most challenging problems in recommender systems based on the collaborative filtering (CF) concept is data sparseness, i.e., limited user preference data is available for making recommendations. Cross-domain collaborative filtering (CDCF) has been studied as an effective mechanism to alleviate data sparseness of one domain using the knowledge about user preferences from other domains. A key question to be answered in the context of CDCF is what common characteristics can be deployed to link different domains for effective knowledge transfer. In this paper, we assess the usefulness of user-contributed (social) tags in this respect. We do so by means of the Generalized Tag-induced Cross-domain Collaborative Filtering (GTagCDCF) approach that we propose in this paper and that we developed based on the general collective matrix factorization framework. Assessment is done by a series of experiments, using publicly available CF datasets that represent three cross-domain cases, i.e., two two-domain cases and one three-domain case. A comparative analysis on two-domain cases involving GTagCDCF and several state-of-the-art CDCF approaches indicates the increased benefit of using social tags as representatives of explicit links between domains for CDCF as compared to the implicit links deployed by the existing CDCF methods. In addition, we show that users from different domains can already benefit from GTagCDCF if they only share a few common tags. Finally, we use the three-domain case to validate the robustness of GTagCDCF with respect to the scale of datasets and the varying number of domains.

cs.IR↗

arXiv2013

Description of 158Er at ultrahigh spin in nuclear density functional theory

Rotational bands in $^{158}$Er at ultra-high spin have been studied in the framework of relativistic and non-relativistic nuclear density functional theories. Consistent results are obtained across the theoretical models used but some puzzles remain when confronted with experiment. Namely, the many-body configurations which provide good description of experimental transition quadrupole moments and dynamic moments of inertia require substantial increase of the spins of observed bands as compared with experimental estimates, which are still subject to large uncertainties. If, however, the theoretical spins assignments turned out to be correct, the experimental band 1 in $^{158}$Er would be the highest-spin structure ever observed.

nucl-th↗

arXiv2013

Kerman-Onishi conditions in self-consistent tilted-axis-cranking mean-field calculations

\item[Background] For cranked mean-field calculations with arbitrarily oriented rotational frequency vector $\boldsymbolω$ in the intrinsic frame, one has to employ constraints on average values of the quadrupole-moment tensor, so as to keep the nucleus in the principal-axis reference frame. Kerman and Onishi [Nucl. Phys. A {\bf 361}, 179 (1981)] have shown that the Lagrangian multipliers that correspond to the required constraints are proportional to $\boldsymbolω \times \boldsymbol{J}$, where $\boldsymbol{J}$ is the average angular momentum vector. \item[Purpose] We study the validity and consequences of the Kerman-Onishi conditions in the context of self-consistent tilted-axis-cranking (TAC) mean-field calculations. \item[Methods] We perform a two-dimensional self-consistent calculations (with and without pairing) utilizing the symmetry-unrestricted solver {\sc hfodd}. At each tilting angle, we compare the calculated values of quadrupole-moment-tensor Lagrangian multipliers and $\boldsymbolω \times \boldsymbol{J}$. \item[Results] We show that in self-consistent calculations, the Kerman-Onishi conditions are obeyed with high precision. Small deviations seen in the calculations with pairing can be attributed to the truncation of the quasiparticle spectrum. We also provide results of systematic TAC calculations for triaxial strongly deformed bands in $^{160}$Yb. \item[Conclusions] For non-stationary TAC solutions, Kerman-Onishi conditions link the non-zero values of the angle between rotational-frequency and angular-momentum vectors to the constraints on off-diagonal components of the quadrupole-moment tensor. To stabilize the convergence of self-consistent iterations, such constraints have to be taken into account. Only then one can determine the Routhian surfaces as functions of the tilting angles.

nucl-th↗

arXiv2013

GAPfm: Optimal Top-N Recommendations for Graded Relevance Domains

Recommender systems are frequently used in domains in which users express their preferences in the form of graded judgments, such as ratings. If accurate top-N recommendation lists are to be produced for such graded relevance domains, it is critical to generate a ranked list of recommended items directly rather than predicting ratings. Current techniques choose one of two sub-optimal approaches: either they optimize for a binary metric such as Average Precision, which discards information on relevance grades, or they optimize for Normalized Discounted Cumulative Gain (NDCG), which ignores the dependence of an item's contribution on the relevance of more highly ranked items. In this paper, we address the shortcomings of existing approaches by proposing the Graded Average Precision factor model (GAPfm), a latent factor model that is particularly suited to the problem of top-N recommendation in domains with graded relevance data. The model optimizes for Graded Average Precision, a metric that has been proposed recently for assessing the quality of ranked results list for graded relevance. GAPfm learns a latent factor model by directly optimizing a smoothed approximation of GAP. GAPfm's advantages are twofold: it maintains full information about graded relevance and also addresses the limitations of models that optimize NDCG. Experimental results show that GAPfm achieves substantial improvements on the top-N recommendation task, compared to several state-of-the-art approaches. In order to ensure that GAPfm is able to scale to very large data sets, we propose a fast learning algorithm that uses an adaptive item selection strategy. A final experiment shows that GAPfm is useful not only for generating recommendation lists, but also for ranking a given list of rated items.

cs.IR↗

arXiv2013

Deformations and quasiparticle spectra of nuclei in the nobelium region

We have performed self-consistent Skyrme Hartree-Fock-Bogolyubov calculations for nuclei close to $^{254}$No. Self-consistent deformations, including $β_{2,4,6,8}$ as functions of the rotational frequency, were determined for even-even nuclei $^{246,248,250}$Fm, $^{252,254}$No, and $^{256}$Rf. The quasiparticle spectra for N=151 isotones and Z=99 isotopes were calculated and compared with experimental data and the results of Woods-Saxon calculations. We found that our calculations give high-order deformations similar to those obtained for the Woods-Saxon potential, and that the experimental quasiparticle energies are reasonably well reproduced.

nucl-th↗

arXiv2011

Self-consistent tilted-axis-cranking study of triaxial strongly deformed bands in $^{158}$Er at ultrahigh spin

Stimulated by recent experimental discoveries, triaxial strongly deformed (TSD) states in $^{158}$Er at ultrahigh spins have been studied by means of the Skyrme-Hartree-Fock model and the tilted-axis-cranking method. Restricting the rotational axis to one of the principal axes -- as done in previous cranking calculations -- two well-defined TSD minima in the total Routhian surface are found for a given configuration: one with positive and another with negative triaxial deformation $γ$. By allowing the rotational axis to change direction, the higher-energy minimum is shown to be a saddle point. This resolves the long-standing question of the physical interpretation of the two triaxial minima at a very similar quadrupole shape obtained in the principal axis cranking approach. Several TSD configurations have been predicted, including a highly deformed band expected to cross lesser elongated TSD bands at the highest spins. Its transitional quadrupole moment $Q_t \approx 10.5$\,eb is close to the measured value of $\sim$11\,eb; hence, it is a candidate for the structure observed in experiment.

nucl-th↗