arXiv subjects

Jianxin Li

Publications and source records attributed to Jianxin Li.

195 records · Page 11Linked to original sources

arXiv2019

Dynamic Network Embedding via Incremental Skip-gram with Negative Sampling

Network representation learning, as an approach to learn low dimensional representations of vertices, has attracted considerable research attention recently. It has been proven extremely useful in many machine learning tasks over large graph. Most existing methods focus on learning the structural representations of vertices in a static network, but cannot guarantee an accurate and efficient embedding in a dynamic network scenario. To address this issue, we present an efficient incremental skip-gram algorithm with negative sampling for dynamic network embedding, and provide a set of theoretical analyses to characterize the performance guarantee. Specifically, we first partition a dynamic network into the updated, including addition/deletion of links and vertices, and the retained networks over time. Then we factorize the objective function of network embedding into the added, vanished and retained parts of the network. Next we provide a new stochastic gradient-based method, guided by the partitions of the network, to update the nodes and the parameter vectors. The proposed algorithm is proven to yield an objective function value with a bounded difference to that of the original objective function. Experimental results show that our proposal can significantly reduce the training time while preserving the comparable performance. We also demonstrate the correctness of the theoretical analysis and the practical usefulness of the dynamic network embedding. We perform extensive experiments on multiple real-world large network datasets over multi-label classification and link prediction tasks to evaluate the effectiveness and efficiency of the proposed framework, and up to 22 times speedup has been achieved.

cs.LG↗

arXiv2019

Fine-grained Event Categorization with Heterogeneous Graph Convolutional Networks

Events are happening in real-world and real-time, which can be planned and organized occasions involving multiple people and objects. Social media platforms publish a lot of text messages containing public events with comprehensive topics. However, mining social events is challenging due to the heterogeneous event elements in texts and explicit and implicit social network structures. In this paper, we design an event meta-schema to characterize the semantic relatedness of social events and build an event-based heterogeneous information network (HIN) integrating information from external knowledge base, and propose a novel Pair-wise Popularity Graph Convolutional Network (PP-GCN) based fine-grained social event categorization model. We propose a Knowledgeable meta-paths Instances based social Event Similarity (KIES) between events and build a weighted adjacent matrix as input to the PP-GCN model. Comprehensive experiments on real data collections are conducted to compare various social event detection and clustering tasks. Experimental results demonstrate that our proposed framework outperforms other alternative social event categorization techniques.

cs.SI↗

arXiv2019

Hierarchical Taxonomy-Aware and Attentional Graph Capsule RCNNs for Large-Scale Multi-Label Text Classification

CNNs, RNNs, GCNs, and CapsNets have shown significant insights in representation learning and are widely used in various text mining tasks such as large-scale multi-label text classification. However, most existing deep models for multi-label text classification consider either the non-consecutive and long-distance semantics or the sequential semantics, but how to consider them both coherently is less studied. In addition, most existing methods treat output labels as independent methods, but ignore the hierarchical relations among them, leading to useful semantic information loss. In this paper, we propose a novel hierarchical taxonomy-aware and attentional graph capsule recurrent CNNs framework for large-scale multi-label text classification. Specifically, we first propose to model each document as a word order preserved graph-of-words and normalize it as a corresponding words-matrix representation which preserves both the non-consecutive, long-distance and local sequential semantics. Then the words-matrix is input to the proposed attentional graph capsule recurrent CNNs for more effectively learning the semantic features. To leverage the hierarchical relations among the class labels, we propose a hierarchical taxonomy embedding method to learn their representations, and define a novel weighted margin loss by incorporating the label representation similarity. Extensive evaluations on three datasets show that our model significantly improves the performance of large-scale multi-label text classification by comparing with state-of-the-art approaches.

cs.IR↗

arXiv2019

Graph Convolutional Neural Networks via Motif-based Attention

Many real-world problems can be represented as graph-based learning problems. In this paper, we propose a novel framework for learning spatial and attentional convolution neural networks on arbitrary graphs. Different from previous convolutional neural networks on graphs, we first design a motif-matching guided subgraph normalization method to capture neighborhood information. Then we implement subgraph-level self-attentional layers to learn different importances from different subgraphs to solve graph classification problems. Analogous to image-based attentional convolution networks that operate on locally connected and weighted regions of the input, we also extend graph normalization from one-dimensional node sequence to two-dimensional node grid by leveraging motif-matching, and design self-attentional layers without requiring any kinds of cost depending on prior knowledge of the graph structure. Our results on both bioinformatics and social network datasets show that we can significantly improve graph classification benchmarks over traditional graph kernel and existing deep models.

cs.LG↗

arXiv2018

Response and phase transition of a Kitaev spin liquid in a local magnetic field

We study the response of the Kitaev spin liquid (KSL) to a local magnetic field perpendicular to the Kitaev honeycomb lattice. The local magnetic field induces a dynamical excitation of a flux pair in the spin liquid and the system can be described by a generally particle-hole asymmetric interacting resonant level model. The dynamical excitation of the flux pair closes the flux gap in the spectrum of the spin correlation function locally for the gapless KSL even from the perturbative response to a weak magnetic field. Beyond the perturbative regime, the p-h asymmetry competes with the magnetic field and results in a rich phase diagram. Moreover, the magnetic field breaks the gauge equivalence of the ferromagnetic and anti-ferromagnetic Kitaev couplings of the ground state and leads to very different behaviors for the two cases. The anti-ferromagnetic case experiences a first order phase transition to the polarized state during magnetization whereas the ferromagnetic case does not. This study can be generalized to the Kitaev model in a uniform magnetic field and may help understand issues in recent experiments on KSL candidates.

cond-mat.str-el↗

arXiv2018

Intermediate gapless phase and topological phase transition of Kitaev model in a uniform magnetic field

We study the Kitaev spin liquid (KSL) in a [001] magnetic field employing the mean field theory (MFT/MF) in the Majorana fermion representation. The MF Hamiltonian of the system has the Bogoliubov de-Gennes (BdG) form of a 2D Weyl superconductor. We discover a robust gapless regime in intermediate magnetic field for both gapless and gapped AFM KSL with Jx = Jy before the system is polarized in high magnetic field. A topological phase transition (PT) connecting two gapless phases with nodal lines takes place at a critical magnetic field hc1 in this regime. While the nodal lines at hc1 is protected by the mirror symmetry with Jx = Jy and disappear at Jx 6= Jy with a gap opening near the critical field, the nodal points at E = 0 can exist at intermediate fields even without mirror symmetry. We reveal that the phase evolution of the KSL in the magnetic field is driven by the competition between the magnetic field and the particle-hole (p-h) asymmetry of the normal state of the BdG Hamiltonian, which results in the robust intermediate gapless phase for AFM KSL. For FM KSL, there is no intermediate PT before polarization. The above phase diagrams are confirmed by dynamical MFT (DMFT) results.

cond-mat.str-el↗

arXiv2018

Inferring Multiplex Diffusion Network via Multivariate Marked Hawkes Process

Understanding the diffusion in social network is an important task. However, this task is challenging since (1) the network structure is usually hidden with only observations of events like "post" or "repost" associated with each node, and (2) the interactions between nodes encompass multiple distinct patterns which in turn affect the diffusion patterns. For instance, social interactions seldom develop on a single channel, and multiple relationships can bind pairs of people due to their various common interests. Most previous work considers only one of these two challenges which is apparently unrealistic. In this paper, we study the problem of \emph{inferring multiplex network} in social networks. We propose the Multiplex Diffusion Model (MDM) which incorporates the multivariate marked Hawkes process and topic model to infer the multiplex structure of social network. A MCMC based algorithm is developed to infer the latent multiplex structure and to estimate the node-related parameters. We evaluate our model based on both synthetic and real-world datasets. The results show that our model is more effective in terms of uncovering the multiplex network structure.

cs.SI↗

arXiv2017

Stacked Kernel Network

Kernel methods are powerful tools to capture nonlinear patterns behind data. They implicitly learn high (even infinite) dimensional nonlinear features in the Reproducing Kernel Hilbert Space (RKHS) while making the computation tractable by leveraging the kernel trick. Classic kernel methods learn a single layer of nonlinear features, whose representational power may be limited. Motivated by recent success of deep neural networks (DNNs) that learn multi-layer hierarchical representations, we propose a Stacked Kernel Network (SKN) that learns a hierarchy of RKHS-based nonlinear features. SKN interleaves several layers of nonlinear transformations (from a linear space to a RKHS) and linear transformations (from a RKHS to a linear space). Similar to DNNs, a SKN is composed of multiple layers of hidden units, but each parameterized by a RKHS function rather than a finite-dimensional vector. We propose three ways to represent the RKHS functions in SKN: (1)nonparametric representation, (2)parametric representation and (3)random Fourier feature representation. Furthermore, we expand SKN into CNN architecture called Stacked Kernel Convolutional Network (SKCN). SKCN learning a hierarchy of RKHS-based nonlinear features by convolutional operation with each filter also parameterized by a RKHS function rather than a finite-dimensional matrix in CNN, which is suitable for image inputs. Experiments on various datasets demonstrate the effectiveness of SKN and SKCN, which outperform the competitive methods.

stat.ML↗

arXiv2017

Intrinsic energy conversion mechanism via telescopic extension and retraction of concentric carbon nanotubes

The conversion of other forms of energy into mechanical work through the geometrical extension and retraction of nanomaterials has a wide variety of potential applications, including for mimicking biomotors. Here, using molecular dynamic simulations, we demonstrate that there exists an intrinsic energy conversion mechanism between thermal energy and mechanical work in the telescopic motions of double-walled carbon nanotubes (DWCNTs). A DWCNT can inherently convert heat into mechanical work in its telescopic extension process, while convert mechanical energy into heat in its telescopic retraction process. These two processes are thermodynamically reversible. The underlying mechanism for this reversibility is that the entropy changes with the telescopic overlapping length of concentric individual tubes. We find also that the entropy effect enlarges with the decreasing intertube space of DWCNTs. As a result, the spontaneously telescopic motion of a condensed DWCNT can be switched to extrusion by rising the system temperature above a critical value. These findings are important for fundamentally understanding the mechanical behavior of concentric nanotubes, and may have general implications in the application of DWCNTs as linear motors in nanodevices.

cond-mat.mes-hall↗

arXiv2016

On the representation and embedding of knowledge bases beyond binary relations

The models developed to date for knowledge base embedding are all based on the assumption that the relations contained in knowledge bases are binary. For the training and testing of these embedding models, multi-fold (or n-ary) relational data are converted to triples (e.g., in FB15K dataset) and interpreted as instances of binary relations. This paper presents a canonical representation of knowledge bases containing multi-fold relations. We show that the existing embedding models on the popular FB15K datasets correspond to a sub-optimal modelling framework, resulting in a loss of structural information. We advocate a novel modelling framework, which models multi-fold relations directly using this canonical representation. Using this framework, the existing TransH model is generalized to a new model, m-TransH. We demonstrate experimentally that m-TransH outperforms TransH by a large margin, thereby establishing a new state of the art.

cs.LG↗

arXiv2014

Distance Landmarks Revisited for Road Graphs

Computing shortest distances is one of the fundamental problems on graphs, and remains a {\em challenging} task today. {\em Distance} {\em landmarks} have been recently studied for shortest distance queries with an auxiliary data structure, referred to as {\em landmark} {\em covers}. This paper studies how to apply distance landmarks for fast {\em exact} shortest distance query answering on large road graphs. However, the {\em direct} application of distance landmarks is {\em impractical} due to the high space and time cost. To rectify this problem, we investigate novel techniques that can be seamlessly combined with distance landmarks. We first propose a notion of {\em hybrid landmark covers}, a revision of landmark covers. Second, we propose a notion of {\em agents}, each of which represents a small subgraph and holds good properties for fast distance query answering. We also show that agents can be computed in {\em linear time}. Third, we introduce graph partitions to deal with the remaining subgraph that cannot be captured by agents. Fourth, we develop a unified framework that seamlessly integrates our proposed techniques and existing optimization techniques, for fast shortest distance query answering. Finally, we experimentally verify that our techniques significantly improve the efficiency of shortest distance queries, using real-life road graphs.

cs.DB↗

arXiv2013

Quasi-SLCA based Keyword Query Processing over Probabilistic XML Data

The probabilistic threshold query is one of the most common queries in uncertain databases, where a result satisfying the query must be also with probability meeting the threshold requirement. In this paper, we investigate probabilistic threshold keyword queries (PrTKQ) over XML data, which is not studied before. We first introduce the notion of quasi-SLCA and use it to represent results for a PrTKQ with the consideration of possible world semantics. Then we design a probabilistic inverted (PI) index that can be used to quickly return the qualified answers and filter out the unqualified ones based on our proposed lower/upper bounds. After that, we propose two efficient and comparable algorithms: Baseline Algorithm and PI index-based Algorithm. To accelerate the performance of algorithms, we also utilize probability density function. An empirical study using real and synthetic data sets has verified the effectiveness and the efficiency of our approaches.

cs.DB↗

arXiv2013

Context-based Diversification for Keyword Queries over XML Data

While keyword query empowers ordinary users to search vast amount of data, the ambiguity of keyword query makes it difficult to effectively answer keyword queries, especially for short and vague keyword queries. To address this challenging problem, in this paper we propose an approach that automatically diversifies XML keyword search based on its different contexts in the XML data. Given a short and vague keyword query and XML data to be searched, we firstly derive keyword search candidates of the query by a classifical feature selection model. And then, we design an effective XML keyword search diversification model to measure the quality of each candidate. After that, three efficient algorithms are proposed to evaluate the possible generated query candidates representing the diversified search intentions, from which we can find and return top-$k$ qualified query candidates that are most relevant to the given keyword query while they can cover maximal number of distinct results.At last, a comprehensive evaluation on real and synthetic datasets demonstrates the effectiveness of our proposed diversification model and the efficiency of our algorithms.

cs.DB↗

arXiv2013

Query-driven Frequent Co-occurring Term Extraction over Relational Data using MapReduce

In this paper we study how to efficiently compute \textit{frequent co-occurring terms} (FCT) in the results of a keyword query in parallel using the popular MapReduce framework. Taking as input a keyword query q and an integer k, an FCT query reports the k terms that are not in q, but appear most frequently in the results of the keyword query q over multiple joined relations. The returned terms of FCT search can be used to do query expansion and query refinement for traditional keyword search. Different from the method of FCT search in a single platform, our proposed approach can efficiently answer a FCT query using the MapReduce Paradigm without pre-computing the results of the original keyword query, which is run in parallel platform. In this work, we can output the final FCT search results by two MapReduce jobs: the first is to extract the statistical information of the data; and the second is to calculate the total frequency of each term based on the output of the first job. At the two MapReduce jobs, we would guarantee the load balance of mappers and the computational balance of reducers as much as possible. Analytical and experimental evaluations demonstrate the efficiency and scalability of our proposed approach using TPC-H benchmark datasets with different sizes.

cs.DB↗

arXiv2011

ELCA Evaluation for Keyword Search on Probabilistic XML Data

As probabilistic data management is becoming one of the main research focuses and keyword search is turning into a more popular query means, it is natural to think how to support keyword queries on probabilistic XML data. With regards to keyword query on deterministic XML documents, ELCA (Exclusive Lowest Common Ancestor) semantics allows more relevant fragments rooted at the ELCAs to appear as results and is more popular compared with other keyword query result semantics (such as SLCAs). In this paper, we investigate how to evaluate ELCA results for keyword queries on probabilistic XML documents. After defining probabilistic ELCA semantics in terms of possible world semantics, we propose an approach to compute ELCA probabilities without generating possible worlds. Then we develop an efficient stack-based algorithm that can find all probabilistic ELCA results and their ELCA probabilities for a given keyword query on a probabilistic XML document. Finally, we experimentally evaluate the proposed ELCA algorithm and compare it with its SLCA counterpart in aspects of result effectiveness, time and space efficiency, and scalability.

cs.DB↗