arXiv subjects

Wenjie Zhang

Publications and source records attributed to Wenjie Zhang.

At least 199 records · Page 11Linked to original sources

arXiv2022

Efficient k-clique Listing with Set Intersection Speedup [Technical Report]

Listing all k-cliques is a fundamental problem in graph mining, with applications in finance, biology, and social network analysis. However, owing to the exponential growth of the search space as k increases, listing all k-cliques is algorithmically challenging. DDegree and DDegCol are the state-of-the-art algorithms that exploit ordering heuristics based on degree ordering and color ordering, respectively. Both DDegree and DDegCol induce high time and space overhead for set intersections cause they construct and maintain all induced subgraphs. Meanwhile, it is non-trivial to implement the data level parallelism to further accelerate on DDegree and DDegCol. In this paper, we propose two efficient algorithms SDegree and BitCol for k-clique listing. We mainly focus on accelerating the set intersections for k-clique listing. Both SDegree and BitCol exploit the data level parallelism for further acceleration with single instruction multiple data (SIMD) or vector instruction sets. Furthermore, we propose two preprocessing techniques Pre-Core and Pre-List, which run in linear time. The preprocessing techniques significantly reduce the size of the original graph and prevent exploring a large number of invalid nodes. In the theoretical analysis, our algorithms have a comparable time complexity and a slightly lower space complexity than the state-of-the-art algorithms. The comprehensive experiments reveal that our algorithms outperform the state-of-the-art algorithms by 3.75x for degree ordering and 5.67x for color ordering on average.

cs.DB↗

arXiv2022

A Syntax-Guided Edit Decoder for Neural Program Repair

Automated Program Repair (APR) helps improve the efficiency of software development and maintenance. Recent APR techniques use deep learning, particularly the encoder-decoder architecture, to generate patches. Though existing DL-based APR approaches have proposed different encoder architectures, the decoder remains to be the standard one, which generates a sequence of tokens one by one to replace the faulty statement. This decoder has multiple limitations: 1) allowing to generate syntactically incorrect programs, 2) inefficiently representing small edits, and 3) not being able to generate project-specific identifiers. In this paper, we propose Recoder, a syntax-guided edit decoder with placeholder generation. Recoder is novel in multiple aspects: 1) Recoder generates edits rather than modified code, allowing efficient representation of small edits; 2) Recoder is syntax-guided, with the novel provider/decider architecture to ensure the syntactic correctness of the patched program and accurate generation; 3) Recoder generates placeholders that could be instantiated as project-specific identifiers later. We conduct experiments to evaluate Recoder on 395 bugs from Defects4J v1.2 and 420 additional bugs from Defects4J v2.0. Our results show that Recoder repairs 53 bugs on Defects4J v1.2, which achieves 21.4% improvement over the previous state-of-the-art approach for single-hunk bugs (TBar). Importantly, to our knowledge, Recoder is the first DL-based APR approach that has outperformed the traditional APR approaches on this dataset. Furthermore, Recoder also repairs 19 bugs on the additional bugs from Defects4J v2.0, which is 137.5% more than TBar (8 bugs) and 850% more than SimFix (2 bugs). This result suggests that Recoder has better generalizability than existing APR approaches.

cs.SE↗

arXiv2022

Observation of Ultrafast Interfacial Exciton Formation and Recombination in Graphene/MoS2 Heterostructure

In this study,we combined time-resolved terahertz spectroscopy along with transient absorption spectroscopy to revisit the interlayer non-equilibrium carrier dynamics in largely lateral size Gr/MoS2 heterostructure fabricated with chemical vapor deposition method. Our experimental results reveal that, with photon-energy below the A-exciton of MoS2 monolayer, hot electrons transfer from graphene to MoS2 takes place in time scale of less than 0.5 ps, resulting in ultrafast formation of interfacial exciton in the heterostructure, subsequently, recombination relaxation of the interfacial exciton occurs in time scale of ~18 ps. A new model considering carrier heating and photogating effect in graphene is proposed to estimate the amount of carrier transfer in the heterostructure, which shows a good agreement with experimental result. Moreover, when the photon-energy is on-resonance with the A-exciton of MoS2, photogenerated holes in MoS2 are transferred to graphene layer within 0.5 ps, leading to the formation of interfacial exciton, the subsequent photoconductivity (PC) relaxation of graphene and bleaching recovery of A-exciton in MoS2 take place around ~10 ps time scale, ascribing to the interfacial exciton recombination. The faster recombination time of interfacial exciton with on-resonance excitation could come from the reduced interface barrier caused by bandgap renormalization effect. Our study provides deep insight into the understanding of interfacial charge transfer as well as the relaxation dynamics in graphene-based heterostructures, which are promising for the applications of graphene-based optoelectronic devices.

physics.optics↗

arXiv2022

Generalized Equivariance and Preferential Labeling for GNN Node Classification

Existing graph neural networks (GNNs) largely rely on node embeddings, which represent a node as a vector by its identity, type, or content. However, graphs with unattributed nodes widely exist in real-world applications (e.g., anonymized social networks). Previous GNNs either assign random labels to nodes (which introduces artefacts to the GNN) or assign one embedding to all nodes (which fails to explicitly distinguish one node from another). Further, when these GNNs are applied to unattributed node classification problems, they have an undesired equivariance property, which are fundamentally unable to address the data with multiple possible outputs. In this paper, we analyze the limitation of existing approaches to node classification problems. Inspired by our analysis, we propose a generalized equivariance property and a Preferential Labeling technique that satisfies the desired property asymptotically. Experimental results show that we achieve high performance in several unattributed node classification tasks.

cs.LG↗

arXiv2022

Ultrafast Dynamics of Defect-Assisted Auger process in PdSe2 films: Synergistic Interaction Between Defect Trapping and Auger Effect

Strong Coulomb interactions in two-dimensional systems, together with quantum confinement, make many-body processes particularly effective for carrier dynamics, which plays a crucial role in determining carrier lifetime, photoconductivity, and emission yield of the materials. Hereby, by using optical pump and terahertz probe spectroscopy, we have investigated the photocarrier dynamics in the PdSe2 films with different thickness. The experimental results reveal that the photocarrier relaxation consists of two components: a fast component of 2.5 ps that shows the layer-thickness independence, and a slow component has typical lifetime of 7.3 ps decreasing with the layer thickness. Surprisingly, the relaxation times for both fast and slow components are exhibited both pump fluence and temperature independence, which suggests that synergistic interactions between defect trapping and Auger effect dominate the photocarrier dynamics in PdSe2 films. A model involving defect-assisted Auger process is proposed, which can reproduce the experimental results well. The fitting results reveal that the layer dependent lifetime is determined by the defect density rather than carrier occupancy rate after photoexcitation. Our results underscore the interplay between Auger process and defects in two-dimensional semiconductors.

cond-mat.mes-hall↗

arXiv2022

GridTuner: Reinvestigate Grid Size Selection for Spatiotemporal Prediction Models [Technical Report]

With the development of traffic prediction technology, spatiotemporal prediction models have attracted more and more attention from academia communities and industry. However, most existing researches focus on reducing model's prediction error but ignore the error caused by the uneven distribution of spatial events within a region. In this paper, we study a region partitioning problem, namely optimal grid size selection problem (OGSS), which aims to minimize the real error of spatiotemporal prediction models by selecting the optimal grid size. In order to solve OGSS, we analyze the upper bound of real error of spatiotemporal prediction models and minimize the real error by minimizing its upper bound. Through in-depth analysis, we find that the upper bound of real error will decrease then increase when the number of model grids increase from 1 to the maximum allowed value. Then, we propose two algorithms, namely Ternary Search and Iterative Method, to automatically find the optimal grid size. Finally, the experiments verify that the error of prediction has the same trend as its upper bound, and the change trend of the upper bound of real error with respect to the increase of the number of model grids will decrease then increase. Meanwhile, in a case study, by selecting the optimal grid size, the order dispatching results of a state-of-the-art prediction-based algorithm can be improved up to 13.6%, which shows the effectiveness of our methods on tuning the region partition for spatiotemporal prediction models.

cs.DB↗

arXiv2022

Temperature-driven Emergence of Negative Photoconductivity in Semimetal MoTe2 Film Probed with Terahertz Spectroscopy

Layered two-dimensional (2D) materials MoTe2 have been paid special attention due to the rich optoelectronic properties with various phases. The nonequilibrium carrier dynamics as well as its temperature dependence in MoTe2 are of prime importance, as it can shed light on understanding the anomalous optical response and potential applications in far infrared (IR) photodetection. Hereby, we employ time-resolved terahertz (THz) spectroscopy to study the temperature dependent nonequilibrium carrier dynamics in MoTe2 films. After photoexcitation of 1.59 eV, the 1T'-phase MoTe2 at high temperature behaves only THz positive photoconductivity (PPC) with relaxation time of less than 1 ps. In contrast, the Td-phase MoTe2 at low temperature shows ultrafast THz PPC initially followed by emerging THz negative photoconductivity (NPC), and the THz NPC signal relaxes to the equilibrium state in hundreds of ps time scale. Small polaron formation induced by hot carrier has been proposed to be ascribed to the THz NPC in the polar semimetal MoTe2 at low temperature. The polaron formation time after photoexcitation increases slightly with temperature, which is determined to be ~0.4 ps at 5 K and 0.5 ps at 100 K. Our experimental result demonstrates for the first time the dynamical formation of small poalron in MoTe2 Weyl semimetal, this is fundamental importance on the understanding the temperature dependent electron-phonon coupling and quantum phase transition, as well as the designing the MoTe2-based far IR photodetector.

cond-mat.str-el↗

arXiv2021

Towards User Engagement Dynamics in Social Networks

The engagement of each user in a social network is an essential indicator for maintaining a sustainable service. Existing studies use the $coreness$ of a user to well estimate its static engagement in a network. However, when the engagement of a user is weakened or strengthened, the influence on other users' engagement is unclear. Besides, the dynamic of user engagement has not been well captured for evolving social networks. In this paper, we systematically study the network dynamic against the engagement change of each user for the first time. The influence of a user is monitored via two novel concepts: the $collapsed~power$ to measure the effect of user weakening, and the $anchored~power$ to measure the effect of user strengthening. We show that the two concepts can be naturally integrated such that a unified offline algorithm is proposed to compute both the collapsed and anchored followers for each user. When the network structure evolves, online techniques are designed to maintain the users' followers, which is faster than redoing the offline algorithm by around 3 orders of magnitude. Extensive experiments on real-life data demonstrate the effectiveness of our model and the efficiency of our algorithms.

cs.SI↗

arXiv2021

Detecting Communities from Heterogeneous Graphs: A Context Path-based Graph Neural Network Model

Community detection, aiming to group the graph nodes into clusters with dense inner-connection, is a fundamental graph mining task. Recently, it has been studied on the heterogeneous graph, which contains multiple types of nodes and edges, posing great challenges for modeling the high-order relationship between nodes. With the surge of graph embedding mechanism, it has also been adopted to community detection. A remarkable group of works use the meta-path to capture the high-order relationship between nodes and embed them into nodes' embedding to facilitate community detection. However, defining meaningful meta-paths requires much domain knowledge, which largely limits their applications, especially on schema-rich heterogeneous graphs like knowledge graphs. To alleviate this issue, in this paper, we propose to exploit the context path to capture the high-order relationship between nodes, and build a Context Path-based Graph Neural Network (CP-GNN) model. It recursively embeds the high-order relationship between nodes into the node embedding with attention mechanisms to discriminate the importance of different relationships. By maximizing the expectation of the co-occurrence of nodes connected by context paths, the model can learn the nodes' embeddings that both well preserve the high-order relationship between nodes and are helpful for community detection. Extensive experimental results on four real-world datasets show that CP-GNN outperforms the state-of-the-art community detection methods.

cs.SI↗

arXiv2021

Disorder-assisted Robustness of Ultrafast Cooling in High Doped CVD-Graphene

Dirac Fermion, which is the low energy collective excitation near the Dirac cone in monolayer graphene, have gained great attention by low energy Terahertz probe. In the case of undoped graphene, it has been generally understood that the ultrafast terahertz thermal relaxation is mostly driven by the electron-phonon coupling (EOP), which can be prolonged to tens and hundreds of picoseconds. However, for the high doped graphene, which manifests the negative photoinduced terahertz conductivity, there is still no consensus on the dominant aspects of the cooling process on a time scale of a few picoseconds. Here, the competition between the disorders assisted defect scattering and the electron-phonon coupling process in the cooling process of the graphene terahertz dynamics is systematically studied and disentangled. We verify experimentally that the ultrafast disorder assisted lattice-phonon interaction, rather than the electron-phonon coupling process, would play the key role in the ultrafast thermal relaxation of the terahertz dynamics. Furthermore, the cooling process features robustness which is independent on the pump wavelength and external temperature. Our finding is expected to propose a considerable possible cooling channel in CVD-graphene and to increase the hot electron extracting efficiency for the design of graphene-based photoconversion devices.

cond-mat.mes-hall↗

arXiv2021

Generative Adversarial Reward Learning for Generalized Behavior Tendency Inference

Recent advances in reinforcement learning have inspired increasing interest in learning user modeling adaptively through dynamic interactions, e.g., in reinforcement learning based recommender systems. Reward function is crucial for most of reinforcement learning applications as it can provide the guideline about the optimization. However, current reinforcement-learning-based methods rely on manually-defined reward functions, which cannot adapt to dynamic and noisy environments. Besides, they generally use task-specific reward functions that sacrifice generalization ability. We propose a generative inverse reinforcement learning for user behavioral preference modelling, to address the above issues. Instead of using predefined reward functions, our model can automatically learn the rewards from user's actions based on discriminative actor-critic network and Wasserstein GAN. Our model provides a general way of characterizing and explaining underlying behavioral tendencies, and our experiments show our method outperforms state-of-the-art methods in a variety of scenarios, namely traffic signal control, online recommender systems, and scanpath prediction.

cs.LG↗

arXiv2021

HUGE: An Efficient and Scalable Subgraph Enumeration System

Subgraph enumeration is a fundamental problem in graph analytics, which aims to find all instances of a given query graph on a large data graph. In this paper, we propose a system called HUGE to efficiently process subgraph enumeration at scale in the distributed context. HUGE features 1) an optimiser to compute an advanced execution plan without the constraints of existing works; 2) a hybrid communication layer that supports both pushing and pulling communication; 3) a novel two-stage execution mode with a lock-free and zero-copy cache design, 4) a BFS/DFS-adaptive scheduler to bound memory consumption, and 5) two-layer intra- and inter-machine load balancing. HUGE is generic such that all existing distributed subgraph enumeration algorithms can be plugged in to enjoy automatic speed up and bounded-memory execution.

cs.DB↗

arXiv2021

PEFP: Efficient k-hop Constrained s-t Simple Path Enumeration on FPGA

Graph plays a vital role in representing entities and their relationships in a variety of fields, such as e-commerce networks, social networks and biological networks. Given two vertices s and t, one of the fundamental problems in graph databases is to investigate the relationships between s and t. A well-studied problem in such area is k-hop constrained s-t simple path enumeration. Nevertheless, all existing algorithms targeting this problem follow the DFS-based paradigm, which cannot scale up well. Moreover, using hardware devices like FPGA to accelerate graph computation has become popular. Motivated by this, in this paper, we propose the first FPGA-based algorithm PEFP to solve the problem of k-hop constrained s-t simple path enumeration efficiently. On the host side, we propose a preprocessing algorithm Pre-BFS to reduce the graph size and search space. On the FPGA side in PEFP, we propose a novel DFS-based batching technique to save on-chip memory efficiently. In addition, we also propose caching techniques to cache necessary data in BRAM, which overcome the latency bottleneck brought by the read/write operations from/to FPGA DRAM. Finally, we propose a data separation technique to enable dataflow optimization for the path verification module; hence the sub-stages in that module can be executed in parallel. Comprehensive experiments show that PEFP outperforms the state-of-the-art algorithm JOIN by more than 1 order of magnitude by average, and up to 2 orders of magnitude in terms of preprocessing time, query processing time and total time, respectively.

cs.DB↗

arXiv2021

Observation of Negative THz Photoconductivity in Large Area Type-II Dirac Semimetal PtTe2

As a newly emergent type-II Dirac semimetal, Platinum Telluride (PtTe2) stands out from other 2D noble-transition-metal dichalcogenides for the unique structure and novel physical properties, such as high carrier mobility, strong electron-phonon coupling and tunable bandgap, which make the PtTe2 a good candidate for applications in optoelectronics, valleytronics and far infrared detectors. Although the transport properties of PtTe2 have been studied extensively, the dynamics of the nonequilibrium carriers remain nearly uninvestigated. Herein we employ optical pump-terahertz (THz) probe spectroscopy (OPTP) to systematically study the photocarrier dynamics of PtTe2 thin films with varying pump fluence, temperature, and film thickness. Upon photoexcitation the THz photoconductivity (PC) of 5 nm PtTe2 film shows abrupt increase initially, while the THz PC changes into negative value in a subpicosecond time scale, followed by a prolonged recovery process that lasted hundreds of picoseconds (ps). This unusual THz PC response observed in the 5 nm PtTe2 film was found to be absent in a 2 nm PtTe2 film. We assign the unexpected negative THz PC as the small polaron formation due to the strong electron-Eg-mode phonon coupling, which is further substantiated by pump fluence- and temperature-dependent measurements as well as the Raman spectroscopy. Moreover, our investigations give a subpicosecond time scale of sequential carrier cooling and polaron formation. The present study provides deep insights into the underlying dynamics evolution mechanisms of photocarrier in type-II Dirac semimetal upon photoexcitation, which is fundamental importance for designing PtTe2-based optoelectronic devices.

cond-mat.mtrl-sci↗

arXiv2020

X-ray flares from the stellar tidal disruption by a candidate supermassive black hole binary

Optical transient surveys have led to the discovery of dozens of stellar tidal disruption events (TDEs) by massive black hole in the centers of galaxies. Despite extensive searches, X-ray follow-up observations have produced no or only weak X-ray detections in most of them. Here we report the discovery of delayed X-ray brightening around 140 days after the optical outburst in the TDE OGLE16aaa, followed by several flux dips during the decay phase. These properties are unusual for standard TDEs and could be explained by the presence of supermassive black hole binary or patchy obscuration. In either scenario, the X-rays can be produced promptly after the disruption but are blocked in the early phase, possibly by a radiation-dominated ejecta which leads to the bulk of optical and ultraviolet emission. Our findings imply that the reprocessing is important in the TDE early evolution, and X-ray observations are promising in revealing supermassive black hole binaries.

astro-ph.HE↗

arXiv2020

NLocalSAT: Boosting Local Search with Solution Prediction

The Boolean satisfiability problem (SAT) is a famous NP-complete problem in computer science. An effective way for solving a satisfiable SAT problem is the stochastic local search (SLS). However, in this method, the initialization is assigned in a random manner, which impacts the effectiveness of SLS solvers. To address this problem, we propose NLocalSAT. NLocalSAT combines SLS with a solution prediction model, which boosts SLS by changing initialization assignments with a neural network. We evaluated NLocalSAT on five SLS solvers (CCAnr, Sparrow, CPSparrow, YalSAT, and probSAT) with instances in the random track of SAT Competition 2018. The experimental results show that solvers with NLocalSAT achieve 27% ~ 62% improvement over the original SLS solvers.

cs.AI↗

arXiv2020

Efficient and Effective Community Search on Large-scale Bipartite Graphs

Bipartite graphs are widely used to model relationships between two types of entities. Community search retrieves densely connected subgraphs containing a query vertex, which has been extensively studied on unipartite graphs. However, community search on bipartite graphs remains largely unexplored. Moreover, all existing cohesive subgraph models on bipartite graphs can only be applied to measure the structure cohesiveness between two sets of vertices while overlooking the edge weight in forming the community. In this paper, we study the significant (alpha, beta)-community search problem on weighted bipartite graphs. Given a query vertex q, we aim to find the significant (alpha, beta)-community R of q which adopts (alpha, beta)-core to characterize the engagement level of vertices, and maximizes the minimum edge weight (significance) within R. To support fast retrieval of R, we first retrieve the maximal connected subgraph of (alpha, beta)-core containing the query vertex (the (alpha, beta)-community), and the search space is limited to this subgraph with a much smaller size than the original graph. A novel index structure is presented which can be built in O(delta * m) time and takes O(delta * m) space where m is the number of edges in G, delta is bounded by the square root of m and is much smaller in practice. Utilizing the index, the (alpha, beta)-community can be retrieved in optimal time. To further obtain R, we develop peeling and expansion algorithms to conduct searches by shrinking from the (alpha, beta)-community and expanding from the query vertex, respectively. The experimental results on real graphs not only demonstrate the effectiveness of the significant (alpha, beta)-community model but also validate the efficiency of our query processing and indexing techniques.

cs.DB↗

arXiv2020

TextSnake: A Flexible Representation for Detecting Text of Arbitrary Shapes

Driven by deep neural networks and large scale datasets, scene text detection methods have progressed substantially over the past years, continuously refreshing the performance records on various standard benchmarks. However, limited by the representations (axis-aligned rectangles, rotated rectangles or quadrangles) adopted to describe text, existing methods may fall short when dealing with much more free-form text instances, such as curved text, which are actually very common in real-world scenarios. To tackle this problem, we propose a more flexible representation for scene text, termed as TextSnake, which is able to effectively represent text instances in horizontal, oriented and curved forms. In TextSnake, a text instance is described as a sequence of ordered, overlapping disks centered at symmetric axes, each of which is associated with potentially variable radius and orientation. Such geometry attributes are estimated via a Fully Convolutional Network (FCN) model. In experiments, the text detector based on TextSnake achieves state-of-the-art or comparable performance on Total-Text and SCUT-CTW1500, the two newly published benchmarks with special emphasis on curved text in natural images, as well as the widely-used datasets ICDAR 2015 and MSRA-TD500. Specifically, TextSnake outperforms the baseline on Total-Text by more than 40% in F-measure.

cs.CV↗