arXiv subjects

Yang Xiang

Publications and source records attributed to Yang Xiang.

At least 271 records · Page 15Linked to original sources

arXiv2017

Task-specific Word Identification from Short Texts Using a Convolutional Neural Network

Task-specific word identification aims to choose the task-related words that best describe a short text. Existing approaches require well-defined seed words or lexical dictionaries (e.g., WordNet), which are often unavailable for many applications such as social discrimination detection and fake review detection. However, we often have a set of labeled short texts where each short text has a task-related class label, e.g., discriminatory or non-discriminatory, specified by users or learned by classification algorithms. In this paper, we focus on identifying task-specific words and phrases from short texts by exploiting their class labels rather than using seed words or lexical dictionaries. We consider the task-specific word and phrase identification as feature learning. We train a convolutional neural network over a set of labeled texts and use score vectors to localize the task-specific words and phrases. Experimental results on sentiment word identification show that our approach significantly outperforms existing methods. We further conduct two case studies to show the effectiveness of our approach. One case study on a crawled tweets dataset demonstrates that our approach can successfully capture the discrimination-related words/phrases. The other case study on fake review detection shows that our approach can identify the fake-review words/phrases.

cs.CL↗

arXiv2017

Wikipedia Vandal Early Detection: from User Behavior to User Embedding

Wikipedia is the largest online encyclopedia that allows anyone to edit articles. In this paper, we propose the use of deep learning to detect vandals based on their edit history. In particular, we develop a multi-source long-short term memory network (M-LSTM) to model user behaviors by using a variety of user edit aspects as inputs, including the history of edit reversion information, edit page titles and categories. With M-LSTM, we can encode each user into a low dimensional real vector, called user embedding. Meanwhile, as a sequential model, M-LSTM updates the user embedding each time after the user commits a new edit. Thus, we can predict whether a user is benign or vandal dynamically based on the up-to-date user embedding. Furthermore, those user embeddings are crucial to discover collaborative vandals.

cs.CR↗

arXiv2017

SNE: Signed Network Embedding

Several network embedding models have been developed for unsigned networks. However, these models based on skip-gram cannot be applied to signed networks because they can only deal with one type of link. In this paper, we present our signed network embedding model called SNE. Our SNE adopts the log-bilinear model, uses node representations of all nodes along a given path, and further incorporates two signed-type vectors to capture the positive or negative relationship of each edge along the path. We conduct two experiments, node classification and link prediction, on both directed and undirected signed networks and compare with four baselines including a matrix factorization method and three state-of-the-art unsigned network embedding models. The experimental results demonstrate the effectiveness of our signed network embedding.

cs.SI↗

arXiv2017

Representation Learning Models for Entity Search

We focus on the problem of learning distributed representations for entity search queries, named entities, and their short descriptions. With our representation learning models, the entity search query, named entity and description can be represented as low-dimensional vectors. Our goal is to develop a simple but effective model that can make the distributed representations of query related entities similar to the query in the vector space. Hence, we propose three kinds of learning strategies, and the difference between them mainly lies in how to deal with the relationship between an entity and its description. We analyze the strengths and weaknesses of each learning strategy and validate our methods on public datasets which contain four kinds of named entities, i.e., movies, TV shows, restaurants and celebrities. The experimental results indicate that our proposed methods can adapt to different types of entity search queries, and outperform the current state-of-the-art methods based on keyword matching and vanilla word2vec models. Besides, the proposed methods can be trained fast and be easily extended to other similar tasks.

cs.CL↗

arXiv2016

Energy of low angle grain boundaries based on continuum dislocation structure

In this paper, we present a continuum model to compute the energy of low angle grain boundaries for any given degrees of freedom (arbitrary rotation axis, rotation angle and boundary plane orientation) based on a continuum dislocation structure. In our continuum model, we minimize the grain boundary energy associated with the dislocation structure subject to the constraint of Frank's formula for dislocations with all possible Burgers vectors. This constrained minimization problem is solved by the penalty method by which it is turned into an unconstrained minimization problem. The grain boundary dislocation structure is approximated by a network of straight dislocations that predicts the energy and dislocation densities of the grain boundaries. The grain boundary energy based on the calculated dislocation structure is able to incorporate its anisotropic nature. We use our continuum model to systematically study the energy of $<111>$ low angle grain boundaries in fcc Al with any boundary plane orientation and all six possible Burgers vectors. Comparisons with results of the atomistic simulations show that our continuum model is able to give excellent predictions of the energy and dislocation densities of low angle grain boundaries. We also study the energy of low angle grain boundaries in fcc Al with varying rotation axis while the rest degrees of freedom are fixed. With minor modifications, our model can also apply to dislocation structures and energy of heterogeneous interfaces.

cond-mat.mtrl-sci↗

arXiv2016

Dislocation climb models from atomistic scheme to dislocation dynamics

We develop a mesoscopic dislocation dynamics model for vacancy-assisted dislocation climb by upscalings from a stochastic model on the atomistic scale. Our models incorporate microscopic mechanisms of (i) bulk diffusion of vacancies, (ii) vacancy exchange dynamics between bulk and dislocation core, (iii) vacancy pipe diffusion along the dislocation core, and (iv) vacancy attachment-detachment kinetics at jogs leading to the motion of jogs. Our mesoscopic model consists of the vacancy bulk diffusion equation and a dislocation climb velocity formula. The effects of pipe diffusion and the jog structure on dislocations are incorporated by a Robin boundary condition near the dislocations for the bulk diffusion equation and a new contribution in the dislocation climb velocity due to vacancy pipe diffusion driven by the stress variation along the dislocation. Our climb formulation is able to quantitatively describe the translation of prismatic loops at low temperatures when the bulk diffusion is negligible. Using this new formulation, we derive analytical formulas for the climb velocity of a straight edge dislocation and a prismatic circular loop. Our dislocation climb formulation can be implemented in dislocation dynamics simulations to incorporate all the above four microscopic mechanisms of dislocation climb.

cond-mat.mtrl-sci↗

arXiv2016

Continuum dynamics of the formation, migration and dissociation of self-locked dislocation structures on parallel slip planes

In continuum models of dislocations, proper formulations of short-range elastic interactions of dislocations are crucial for capturing various types of dislocation patterns formed in crystalline materials. In this article, the continuum dynamics of straight dislocations distributed on two parallel slip planes is modelled through upscaling the underlying discrete dislocation dynamics. Two continuum velocity field quantities are introduced to facilitate the discrete-to-continuum transition. The first one is the local migration velocity of dislocation ensembles which is found fully independent of the short-range dislocation correlations. The second one is the decoupling velocity of dislocation pairs controlled by a threshold stress value, which is proposed to be the effective flow stress for single slip systems. Compared to the almost ubiquitously adopted Taylor relationship, the derived flow stress formula exhibits two features that are more consistent with the underlying discrete dislocation dynamics: i) the flow stress increases with the in-plane component of the dislocation density only up to a certain value, hence the derived formula admits a minimum inter-dislocation distance within slip planes; ii) the flow stress smoothly transits to zero when all dislocations become geometrically necessary dislocations. A regime under which inhomogeneities in dislocation density grow is identified, and is further validated through comparison with discrete dislocation dynamical simulation results. Based on the findings in this article and in our previous works, a general strategy for incorporating short-range dislocation correlations in continuum models of dislocations is proposed.

cond-mat.mtrl-sci↗

arXiv2016

The Dependent Random Measures with Independent Increments in Mixture Models

When observations are organized into groups where commonalties exist amongst them, the dependent random measures can be an ideal choice for modeling. One of the propositions of the dependent random measures is that the atoms of the posterior distribution are shared amongst groups, and hence groups can borrow information from each other. When normalized dependent random measures prior with independent increments are applied, we can derive appropriate exchangeable probability partition function (EPPF), and subsequently also deduce its inference algorithm given any mixture model likelihood. We provide all necessary derivation and solution to this framework. For demonstration, we used mixture of Gaussians likelihood in combination with a dependent structure constructed by linear combinations of CRMs. Our experiments show superior performance when using this framework, where the inferred values including the mixing weights and the number of clusters both respond appropriately to the number of completely random measure used.

stat.ML↗

arXiv2016

Smoothed Hierarchical Dirichlet Process: A Non-Parametric Approach to Constraint Measures

Time-varying mixture densities occur in many scenarios, for example, the distributions of keywords that appear in publications may evolve from year to year, video frame features associated with multiple targets may evolve in a sequence. Any models that realistically cater to this phenomenon must exhibit two important properties: the underlying mixture densities must have an unknown number of mixtures, and there must be some "smoothness" constraints in place for the adjacent mixture densities. The traditional Hierarchical Dirichlet Process (HDP) may be suited to the first property, but certainly not the second. This is due to how each random measure in the lower hierarchies is sampled independent of each other and hence does not facilitate any temporal correlations. To overcome such shortcomings, we proposed a new Smoothed Hierarchical Dirichlet Process (sHDP). The key novelty of this model is that we place a temporal constraint amongst the nearby discrete measures $\{G_j\}$ in the form of symmetric Kullback-Leibler (KL) Divergence with a fixed bound $B$. Although the constraint we place only involves a single scalar value, it nonetheless allows for flexibility in the corresponding successive measures. Remarkably, it also led us to infer the model within the stick-breaking process where the traditional Beta distribution used in stick-breaking is now replaced by a new constraint calculated from $B$. We present the inference algorithm and elaborate on its solutions. Our experiment using NIPS keywords has shown the desirable effect of the model.

stat.ML↗

arXiv2016

Fog Computing: Focusing on Mobile Users at the Edge

With smart devices, particular smartphones, becoming our everyday companions, the ubiquitous mobile Internet and computing applications pervade people daily lives. With the surge demand on high-quality mobile services at anywhere, how to address the ubiquitous user demand and accommodate the explosive growth of mobile traffics is the key issue of the next generation mobile networks. The Fog computing is a promising solution towards this goal. Fog computing extends cloud computing by providing virtualized resources and engaged location-based services to the edge of the mobile networks so as to better serve mobile traffics. Therefore, Fog computing is a lubricant of the combination of cloud computing and mobile applications. In this article, we outline the main features of Fog computing and describe its concept, architecture and design goals. Lastly, we discuss some of the future research issues from the networking perspective.

cs.NI↗

arXiv2016

A View of Fog Computing from Networking Perspective

With smart devices, particular smartphones, becoming our everyday companions, the ubiquitous mobile Internet and computing applications pervade people's daily lives. With the surge demand on high-quality mobile services at anywhere, how to address the ubiquitous user demand and accommodate the explosive growth of mobile traffics is the key issue of the next generation mobile networks. The Fog computing is a promising solution towards this goal. Fog computing extends cloud computing by providing virtualized resources and engaged location-based services to the edge of the mobile networks so as to better serve mobile traffics. Therefore, Fog computing is a lubricant of the combination of cloud computing and mobile applications. In this article, we outline the main features of Fog computing and describe its concept, architecture and design goals. Lastly, we discuss some of the future research issues from the networking perspective.

cs.NI↗

arXiv2016

Homogenisation of a Row of Dislocation Dipoles from Discrete Dislocation Dynamics

Conventional discrete-to-continuum approaches have seen their limitation in describing the collective behaviour of the multi-polar configurations of dislocations, which are widely observed in crystalline materials. The reason is that dislocation dipoles, which play an important role in determining the mechanical properties of crystals, often get smeared out when traditional homogenisation methods are applied. To address such difficulties, the collective behaviour of a row of dislocation dipoles is studied by using matched asymptotic techniques. The discrete-to-continuum transition is facilitated by introducing two field variables respectively describing the dislocation pair density potential and the dislocation pair width. It is found that the dislocation pair width evolves much faster than the pair density. Such hierarchy in evolution time scales enables us to describe the dislocation dynamics at the coarse-grained level by an evolution equation for the slowly varying variable (the pair density) coupled with an equilibrium equation for the fast varying variable (the pair width). The time-scale separation method adopted here paves a way for properly incorporating dipole-like (zero net Burgers vector but non-vanishing) dislocation structures, known as the statistically stored dislocations (SSDs) into macroscopic models of crystal plasticity in three dimensions. Moreover, the natural transition between different equilibrium patterns found here may also shed light on understanding the emergence of the persistent slip bands (PSBs) in fatigue metals induced by cyclic loads.

cond-mat.mtrl-sci↗

arXiv2015

A continuum model for dislocation dynamics in three dimensions using the dislocation density potential functions and its application in understanding the micro-pillar size effect

In this paper, we present a dislocation-density-based three-dimensional continuum model, where the dislocation substructures are represented by pairs of dislocation density potential functions (DDPFs), denoted by $ϕ$ and $ψ$. The slip plane distribution is characterized by the contour surfaces of $ψ$, while the distribution of dislocation curves on each slip plane is identified by the contour curves of $ϕ$ which represents the plastic slip on the slip plane. By using DDPFs, we can explicitly write down an evolution equation system, which is shown consistent with the underlying discrete dislocation dynamics. The system includes i) A constitutive stress rule, which describes how the total stress field is determined in the presence of dislocation networks and applied loads; ii) A plastic flow rule, which describes how dislocation ensembles evolve. The proposed continuum model is validated through comparisons with discrete dislocation dynamics simulation results and experimental data. As an application of the proposed model, the "smaller-being-stronger" size effect observed in single-crystal micro-pillars is studied. A scaling law for the pillar flow stress $σ_{\text{flow}}$ against its (non-dimensionalized) size $D$ is derived to be $σ_{\text{flow}}\sim\log(D)/D$.

cond-mat.mtrl-sci↗

arXiv2015

Synergy, suppression and immorality: forward differences of the entropy function

Conditional mutual information is important in the selection and interpretation of graphical models. Its empirical version is well known as a generalised likelihood ratio test and that it may be represented as a difference in entropy. We consider the forward difference expansion of the entropy function defined on all subsets of the variables under study. The elements of this expansion are invariant to permutation of their suffices and relate higher order mutual informations to lower order ones. The third order difference is expressible as an, apparently assymmetric, difference between a marginal and a conditional mutual information. Its role in the decomposition for explained information provides a technical definition for synergy between three random variables. Positive values occur when two variables provide alternative explanations for a third; negative values, termed synergies, occur when the sum of explained information is greater than the sum of its parts. Synergies tend to be infrequent; they connect the seemingly unrelated concepts of suppressor variables in regression, on the one hand, and unshielded colliders in Bayes networks (immoralities), on the other. We give novel characterizations of these phenomena that generalise to categorical variables and to higher dimensions. We propose an algorithm for systematically computing low order differences from a given graph. Examples from small scale real-life studies indicate the potential of these techniques for empirical statistical analysis.

stat.ME↗

arXiv2014

Simple linear algorithms for mining graph cores

Batagelj and Zaversnik proposed a linear algorithm for the well-known $k$-core decomposition problem. However, when $k$-cores are desired for a given $k$, we find that a simple linear algorithm requiring no sorting works for mining $k$-cores. In addition, this algorithm can be extended to mine $(k_1, k_2,\ldots, k_p)$-cores from $p$-partite graphs in linear time, and this mining approach can be efficiently implemented in a distributed computing environment with a lower message complexity bound in comparison with the best known method of distributed $k$-core decomposition.

cs.DS↗

arXiv2013

Can Uncertainty Management be Realized in a Finite Totally Ordered Probability Algebra?

In this paper, the feasibility of using finite totally ordered probability models under Alelinnas's Theory of Probabilistic Logic [Aleliunas, 1988] is investigated. The general form of the probability algebra of these models is derived and the number of possible algebras with given size is deduced. Based on this analysis, we discuss problems of denominator-indifference and ambiguity-generation that arise in reasoning by cases and abductive reasoning. An example is given that illustrates how these problems arise. The investigation shows that a finite probability model may be of very limited usage.

cs.AI↗

arXiv2013

Exploring Localization in Bayesian Networks for Large Expert Systems

Current Bayesian net representations do not consider structure in the domain and include all variables in a homogeneous network. At any time, a human reasoner in a large domain may direct his attention to only one of a number of natural subdomains, i.e., there is ?localization' of queries and evidence. In such a case, propagating evidence through a homogeneous network is inefficient since the entire network has to be updated each time. This paper presents multiply sectioned Bayesian networks that enable a (localization preserving) representation of natural subdomains by separate Bayesian subnets. The subnets are transformed into a set of permanent junction trees such that evidential reasoning takes place at only one of them at a time. Probabilities obtained are identical to those that would be obtained from the homogeneous network. We discuss attention shift to a different junction tree and propagation of previously acquired evidence. Although the overall system can be large, computational requirements are governed by the size of only one junction tree.

cs.AI↗

arXiv2013

A Method for Implementing a Probabilistic Model as a Relational Database

This paper discusses a method for implementing a probabilistic inference system based on an extended relational data model. This model provides a unified approach for a variety of applications such as dynamic programming, solving sparse linear equations, and constraint propagation. In this framework, the probability model is represented as a generalized relational database. Subsequent probabilistic requests can be processed as standard relational queries. Conventional database management systems can be easily adopted for implementing such an approximate reasoning system.

cs.AI↗