arXiv subjects

Jiawei Zhang

Publications and source records attributed to Jiawei Zhang.

At least 325 records · Page 18Linked to original sources

arXiv2021

Label Contrastive Coding based Graph Neural Network for Graph Classification

Graph classification is a critical research problem in many applications from different domains. In order to learn a graph classification model, the most widely used supervision component is an output layer together with classification loss (e.g.,cross-entropy loss together with softmax or margin loss). In fact, the discriminative information among instances are more fine-grained, which can benefit graph classification tasks. In this paper, we propose the novel Label Contrastive Coding based Graph Neural Network (LCGNN) to utilize label information more effectively and comprehensively. LCGNN still uses the classification loss to ensure the discriminability of classes. Meanwhile, LCGNN leverages the proposed Label Contrastive Loss derived from self-supervised learning to encourage instance-level intra-class compactness and inter-class separability. To power the contrastive learning, LCGNN introduces a dynamic label memory bank and a momentum updated encoder. Our extensive evaluations with eight benchmark graph datasets demonstrate that LCGNN can outperform state-of-the-art graph classification models. Experimental results also verify that LCGNN can achieve competitive performance with less training data because LCGNN exploits label information comprehensively.

cs.LG↗

arXiv2020

Online Resource Allocation with Stochastic Resource Consumption

We consider an online resource allocation problem where multiple resources, each with an individual initial capacity, are available to serve random requests arriving sequentially over multiple discrete time periods. At each time period, one request arrives and its associated reward and size are drawn independently from a known distribution that could be resource-dependent. Upon its arrival and revealing itself, an online decision has to be made on whether or not to accept the request. If accepted, another online decision should also be made to specify on assigning which resource to serve the request, then a certain amount of the resource equal to the size of the request will be consumed and a reward will be collected. The objective of the decision maker is to maximize the total collected reward subject to the capacity constraints. We develop near-optimal policies for our problem under two settings separately. When the reward distribution has a finite support, we assume that given each reward realization, the size distributions for different resources are perfectly positive correlated and we propose an adaptive threshold policy. We show that with certain regularity conditions on the size distribution, our policy enjoys an optimal $O(\log T)$ regret bound, where $T$ denotes the total time periods. When the support of the reward distribution is not necessarily finite, we develop another adaptive threshold policy with a $O(\log T)$ regret bound when both the reward and the size of each request are resource-independent and the size distribution satisfies the same conditions.

math.OC↗

arXiv2020

Sub-graph Contrast for Scalable Self-Supervised Graph Representation Learning

Graph representation learning has attracted lots of attention recently. Existing graph neural networks fed with the complete graph data are not scalable due to limited computation and memory costs. Thus, it remains a great challenge to capture rich information in large-scale graph data. Besides, these methods mainly focus on supervised learning and highly depend on node label information, which is expensive to obtain in the real world. As to unsupervised network embedding approaches, they overemphasize node proximity instead, whose learned representations can hardly be used in downstream application tasks directly. In recent years, emerging self-supervised learning provides a potential solution to address the aforementioned problems. However, existing self-supervised works also operate on the complete graph data and are biased to fit either global or very local (1-hop neighborhood) graph structures in defining the mutual information based loss terms. In this paper, a novel self-supervised representation learning method via Subgraph Contrast, namely \textsc{Subg-Con}, is proposed by utilizing the strong correlation between central nodes and their sampled subgraphs to capture regional structure information. Instead of learning on the complete input graph data, with a novel data augmentation strategy, \textsc{Subg-Con} learns node representations through a contrastive loss defined based on subgraphs sampled from the original graph instead. Compared with existing graph representation learning approaches, \textsc{Subg-Con} has prominent performance advantages in weaker supervision requirements, model learning scalability, and parallelization. Extensive experiments verify both the effectiveness and the efficiency of our work compared with both classic and state-of-the-art graph representation learning approaches on multiple real-world large-scale benchmark datasets from different domains.

cs.LG↗

arXiv2020

Heterogeneous Deep Graph Infomax

Graph representation learning is to learn universal node representations that preserve both node attributes and structural information. The derived node representations can be used to serve various downstream tasks, such as node classification and node clustering. When a graph is heterogeneous, the problem becomes more challenging than the homogeneous graph node learning problem. Inspired by the emerging information theoretic-based learning algorithm, in this paper we propose an unsupervised graph neural network Heterogeneous Deep Graph Infomax (HDGI) for heterogeneous graph representation learning. We use the meta-path structure to analyze the connections involving semantics in heterogeneous graphs and utilize graph convolution module and semantic-level attention mechanism to capture local representations. By maximizing local-global mutual information, HDGI effectively learns high-level node representations that can be utilized in downstream graph-related tasks. Experiment results show that HDGI remarkably outperforms state-of-the-art unsupervised graph representation learning methods on both classification and clustering tasks. By feeding the learned representations into a parametric model, such as logistic regression, we even achieve comparable performance in node classification tasks when comparing with state-of-the-art supervised end-to-end GNN models.

cs.LG↗

arXiv2020

AIM 2020 Challenge on Learned Image Signal Processing Pipeline

This paper reviews the second AIM learned ISP challenge and provides the description of the proposed solutions and results. The participating teams were solving a real-world RAW-to-RGB mapping problem, where to goal was to map the original low-quality RAW images captured by the Huawei P20 device to the same photos obtained with the Canon 5D DSLR camera. The considered task embraced a number of complex computer vision subtasks, such as image demosaicing, denoising, white balancing, color and contrast correction, demoireing, etc. The target metric used in this challenge combined fidelity scores (PSNR and SSIM) with solutions' perceptual results measured in a user study. The proposed solutions significantly improved the baseline results, defining the state-of-the-art for practical image signal processing pipeline modeling.

cs.CV↗

arXiv2020

Distributed Stochastic Consensus Optimization with Momentum for Nonconvex Nonsmooth Problems

While many distributed optimization algorithms have been proposed for solving smooth or convex problems over the networks, few of them can handle non-convex and non-smooth problems. Based on a proximal primal-dual approach, this paper presents a new (stochastic) distributed algorithm with Nesterov momentum for accelerated optimization of non-convex and non-smooth problems. Theoretically, we show that the proposed algorithm can achieve an $ε$-stationary solution under a constant step size with $\mathcal{O}(1/ε^2)$ computation complexity and $\mathcal{O}(1/ε)$ communication complexity. When compared to the existing gradient tracking based methods, the proposed algorithm has the same order of computation complexity but lower order of communication complexity. To the best of our knowledge, the presented result is the first stochastic algorithm with the $\mathcal{O}(1/ε)$ communication complexity for non-convex and non-smooth problems. Numerical experiments for a distributed non-convex regression problem and a deep neural network based classification problem are presented to illustrate the effectiveness of the proposed algorithms.

math.OC↗

arXiv2020

EnsemFDet: An Ensemble Approach to Fraud Detection based on Bipartite Graph

Fraud detection is extremely critical for e-commerce business. It is the intent of the companies to detect and prevent fraud as early as possible. Existing fraud detection methods try to identify unexpected dense subgraphs and treat related nodes as suspicious. Spectral relaxation-based methods solve the problem efficiently but hurt the performance due to the relaxed constraints. Besides, many methods cannot be accelerated with parallel computation or control the number of returned suspicious nodes because they provide a set of subgraphs with diverse node sizes. These drawbacks affect the real-world applications of existing methods. In this paper, we propose an Ensemble-based Fraud Detection (EnsemFDet) method to scale up fraud detection in bipartite graphs by decomposing the original problem into subproblems on small-sized subgraphs. By oversampling the graph and solving the subproblems, the ensemble approach further votes suspicious nodes without sacrificing the prediction accuracy. Extensive experiments have been done on real transaction data from JD.com, which is one of the world's largest e-commerce platforms. Experimental results demonstrate the effectiveness, practicability, and scalability of EnsemFDet. More specifically, EnsemFDet is up to 100x faster than the state-of-the-art methods due to its parallelism with all aspects of data.

cs.LG↗

arXiv2020

Cross-Scale Internal Graph Neural Network for Image Super-Resolution

Non-local self-similarity in natural images has been well studied as an effective prior in image restoration. However, for single image super-resolution (SISR), most existing deep non-local methods (e.g., non-local neural networks) only exploit similar patches within the same scale of the low-resolution (LR) input image. Consequently, the restoration is limited to using the same-scale information while neglecting potential high-resolution (HR) cues from other scales. In this paper, we explore the cross-scale patch recurrence property of a natural image, i.e., similar patches tend to recur many times across different scales. This is achieved using a novel cross-scale internal graph neural network (IGNN). Specifically, we dynamically construct a cross-scale graph by searching k-nearest neighboring patches in the downsampled LR image for each query patch in the LR image. We then obtain the corresponding k HR neighboring patches in the LR image and aggregate them adaptively in accordance to the edge label of the constructed graph. In this way, the HR information can be passed from k HR neighboring patches to the LR query patch to help it recover more detailed textures. Besides, these internal image-specific LR/HR exemplars are also significant complements to the external information learned from the training dataset. Extensive experiments demonstrate the effectiveness of IGNN against the state-of-the-art SISR methods including existing non-local networks on standard benchmarks.

cs.CV↗

arXiv2020

Graph Neural Lasso for Dynamic Network Regression

The regression of multiple inter-connected sequence data is a problem in various disciplines. Formally, we name the regression problem of multiple inter-connected data entities as the "dynamic network regression" in this paper. Within the problem of stock forecasting or traffic speed prediction, we need to consider both the trends of the entities and the relationships among the entities. A majority of existing approaches can't capture that information together. Some of the approaches are proposed to deal with the sequence data, like LSTM. The others use the prior knowledge in a network to get a fixed graph structure and do prediction on some unknown entities, like GCN. To overcome the limitations in those methods, we propose a novel graph neural network, namely Graph Neural Lasso (GNL), to deal with the dynamic network problem. GNL extends the GDU (gated diffusive unit) as the base neuron to capture the information behind the sequence. Rather than using a fixed graph structure, GNL can learn the dynamic graph structure automatically. By adding the attention mechanism in GNL, we can learn the dynamic relations among entities within each network snapshot. Combining these two parts, GNL is able to model the dynamic network problem well. Experimental results provided on two networked sequence datasets, i.e., Nasdaq-100 and METR-LA, show that GNL can address the network regression problem very well and is also very competitive among the existing approaches.

cs.LG↗

arXiv2020

EfficientFCN: Holistically-guided Decoding for Semantic Segmentation

Both performance and efficiency are important to semantic segmentation. State-of-the-art semantic segmentation algorithms are mostly based on dilated Fully Convolutional Networks (dilatedFCN), which adopt dilated convolutions in the backbone networks to extract high-resolution feature maps for achieving high-performance segmentation performance. However, due to many convolution operations are conducted on the high-resolution feature maps, such dilatedFCN-based methods result in large computational complexity and memory consumption. To balance the performance and efficiency, there also exist encoder-decoder structures that gradually recover the spatial information by combining multi-level feature maps from the encoder. However, the performances of existing encoder-decoder methods are far from comparable with the dilatedFCN-based methods. In this paper, we propose the EfficientFCN, whose backbone is a common ImageNet pre-trained network without any dilated convolution. A holistically-guided decoder is introduced to obtain the high-resolution semantic-rich feature maps via the multi-scale features from the encoder. The decoding task is converted to novel codebook generation and codeword assembly task, which takes advantages of the high-level and low-level features from the encoder. Such a framework achieves comparable or even better performance than state-of-the-art methods with only 1/3 of the computational cost. Extensive experiments on PASCAL Context, PASCAL VOC, ADE20K validate the effectiveness of the proposed EfficientFCN.

cs.CV↗

arXiv2020

Probing efficient n-type lanthanide dopants for Mg$_3$Sb$_2$ thermoelectrics

The recent discovery of n-type Mg$_3$Sb$_2$ thermoelectric has ignited intensive research activities on searching for potential n-type dopants for this material. Using first-principles defect calculations, here we conduct a systematic computational screening of potential efficient n-type lanthanide dopants for Mg$_3$Sb$_2$. In addition to La, Ce, Pr, and Tm, we find that high electron concentration ($\geq$ 10$^{20}$ cm$^{-3}$ at the growth temperature of 900 K) can be achieved by doping on the Mg sites with Nd, Gd, Ho, and Lu, which are generally more efficient than other lanthanide dopants and the anion-site dopant Te. Experimentally, we confirm Nd and Tm as effective n-type dopants for Mg$_3$Sb$_2$ since doping with Nd and Tm shows superior thermoelectric figure of merit zT $\geq$ 1.3 with higher electron concentration than doping with Te. Through codoping with Nd (Tm) and Te, simultaneous power factor improvement and thermal conductivity reduction are achieved. As a result, we obtain high zT values of about 1.65 and 1.75 at 775 K in n-type Mg$_{3.5}$Nd$_{0.04}$Sb$_{1.97}$Te$_{0.03}$ and Mg$_{3.5}$Tm$_{0.03}$Sb$_{1.97}$Te$_{0.03}$, respectively, which are among the highest values for n-type Mg$_3$Sb$_2$ without alloying with Mg$_3$Bi$_2$. This work sheds light on exploring promising n-type dopants for the design of Mg$_3$Sb$_2$ thermoelectrics.

cond-mat.mtrl-sci↗

arXiv2020

A Proximal Alternating Direction Method of Multiplier for Linearly Constrained Nonconvex Minimization

Consider the minimization of a nonconvex differentiable function over a polyhedron. A popular primal-dual first-order method for this problem is to perform a gradient projection iteration for the augmented Lagrangian function and then update the dual multiplier vector using the constraint residual. However, numerical examples show that this approach can exhibit "oscillation" and may not converge. In this paper, we propose a proximal alternating direction method of multipliers for the multi-block version of this problem. A distinctive feature of this method is the introduction of a "smoothed" (i.e., exponentially weighted) sequence of primal iterates, and the inclusion, at each iteration, to the augmented Lagrangian function a quadratic proximal term centered at the current smoothed primal iterate. The resulting proximal augmented Lagrangian function is inexactly minimized (via a gradient projection step) at each iteration while the dual multiplier vector is updated using the residual of the linear constraints. When the primal and dual stepsizes are chosen sufficiently small, we show that suitable "smoothing" can stabilize the "oscillation", and the iterates of the new proximal ADMM algorithm converge to a stationary point under some mild regularity conditions. Furthermore, when the objective function is quadratic, we establish the linear convergence of the algorithm. Our proof is based on a new potential function and a novel use of error bounds.

math.OC↗

arXiv2020

Scalable Heterogeneous Social Network Alignment through Synergistic Graph Partition

Social network alignment has been an important research problem for social network analysis in recent years. With the identified shared users across networks, it will provide researchers with the opportunity to achieve a more comprehensive understanding of users' social activities both within and across networks. Social network alignment is a very difficult problem. Besides the challenges introduced by the network heterogeneity, the network alignment problem can be reduced to a combinatorial optimization problem with an extremely large search space. The learning effectiveness and efficiency of existing alignment models will be degraded significantly as the network size increases. In this paper, we will focus on studying the scalable heterogeneous social network alignment problem, and propose to address it with a novel two-stage network alignment model, namely \textbf{S}calable \textbf{H}eterogeneous \textbf{N}etwork \textbf{A}lignment (SHNA). Based on a group of intra- and inter-network meta diagrams, SHNA first partitions the social networks into a group of sub-networks synergistically. Via the partially known anchor links, SHNA will extract the partitioned sub-network correspondence relationships. Instead of aligning the complete input network, SHNA proposes to identify the anchor links between the matched sub-network pairs, while those between the unmatched sub-networks will be pruned to effectively shrink the search space. Extensive experiments have been done to compare SHNA with the state-of-the-art baseline methods on a real-world aligned social networks dataset. The experimental results have demonstrated both the effectiveness and efficiency of the {\our} model in addressing the problem.

cs.SI↗

arXiv2020

Meta Diagram based Active Social Networks Alignment

Network alignment aims at inferring a set of anchor links matching the shared entities between different information networks, which has become a prerequisite step for effective fusion of multiple information networks. In this paper, we will study the network alignment problem to fuse online social networks specifically. Social network alignment is extremely challenging to address due to several reasons, i.e., lack of training data, network heterogeneity and one-to-one constraint. Existing network alignment works usually require a large number of training data, but such a demand can hardly be met in applications, as manual anchor link labeling is extremely expensive. Significantly different from other homogeneous network alignment works, information in online social networks is usually of heterogeneous categories, the incorporation of which in model building is not an easy task. Furthermore, the one-to-one cardinality constraint on anchor links renders their inference process intertwistingly correlated. To resolve these three challenges, a novel network alignment model, namely ActiveIter, is introduced in this paper. ActiveIter defines a set of inter-network meta diagrams for anchor link feature extraction, adopts active learning for effective label query and uses greedy link selection for anchor link cardinality filtering. Extensive experiments are conducted on real-world aligned networks datasets, and the experimental results have demonstrated the effectiveness of ActiveIter compared with other state-of-the-art baseline methods.

cs.SI↗

arXiv2020

A global dual error bound and its application to the analysis of linearly constrained nonconvex optimization

Error bound analysis, which estimates the distance of a point to the solution set of an optimization problem using the optimality residual, is a powerful tool for the analysis of first-order optimization algorithms. In this paper, we use global error bound analysis to study the iteration complexity of a first-order algorithm for a linearly constrained nonconvex minimization problem. we develop a global dual error bound analysis for a regularized version of this nonconvex problem by using a novel ``decomposition'' technique. Equipped with this global dual error bound, we prove that a suitably designed primal-dual first order method can generate an $ε$-stationary solution of the linearly constrained nonconvex minimization problem within $\mathcal{O}(1/ε^2)$ iterations, which is the best known iteration complexity for this class of nonconvex problems.

math.OC↗

arXiv2020

G5: A Universal GRAPH-BERT for Graph-to-Graph Transfer and Apocalypse Learning

The recent GRAPH-BERT model introduces a new approach to learning graph representations merely based on the attention mechanism. GRAPH-BERT provides an opportunity for transferring pre-trained models and learned graph representations across different tasks within the same graph dataset. In this paper, we will further investigate the graph-to-graph transfer of a universal GRAPH-BERT for graph representation learning across different graph datasets, and our proposed model is also referred to as the G5 for simplicity. Many challenges exist in learning G5 to adapt the distinct input and output configurations for each graph data source, as well as the information distributions differences. G5 introduces a pluggable model architecture: (a) each data source will be pre-processed with a unique input representation learning component; (b) each output application task will also have a specific functional component; and (c) all such diverse input and output components will all be conjuncted with a universal GRAPH-BERT core component via an input size unification layer and an output representation fusion layer, respectively. The G5 model removes the last obstacle for cross-graph representation learning and transfer. For the graph sources with very sparse training data, the G5 model pre-trained on other graphs can still be utilized for representation learning with necessary fine-tuning. What's more, the architecture of G5 also allows us to learn a supervised functional classifier for data sources without any training data at all. Such a problem is also named as the Apocalypse Learning task in this paper. Two different label reasoning strategies, i.e., Cross-Source Classification Consistency Maximization (CCCM) and Cross-Source Dynamic Routing (CDR), are introduced in this paper to address the problem.

cs.LG↗

arXiv2020

Learning a Reinforced Agent for Flexible Exposure Bracketing Selection

Automatically selecting exposure bracketing (images exposed differently) is important to obtain a high dynamic range image by using multi-exposure fusion. Unlike previous methods that have many restrictions such as requiring camera response function, sensor noise model, and a stream of preview images with different exposures (not accessible in some scenarios e.g. some mobile applications), we propose a novel deep neural network to automatically select exposure bracketing, named EBSNet, which is sufficiently flexible without having the above restrictions. EBSNet is formulated as a reinforced agent that is trained by maximizing rewards provided by a multi-exposure fusion network (MEFNet). By utilizing the illumination and semantic information extracted from just a single auto-exposure preview image, EBSNet can select an optimal exposure bracketing for multi-exposure fusion. EBSNet and MEFNet can be jointly trained to produce favorable results against recent state-of-the-art approaches. To facilitate future research, we provide a new benchmark dataset for multi-exposure selection and fusion.

cs.CV↗

arXiv2020

CG-BERT: Conditional Text Generation with BERT for Generalized Few-shot Intent Detection

In this paper, we formulate a more realistic and difficult problem setup for the intent detection task in natural language understanding, namely Generalized Few-Shot Intent Detection (GFSID). GFSID aims to discriminate a joint label space consisting of both existing intents which have enough labeled data and novel intents which only have a few examples for each class. To approach this problem, we propose a novel model, Conditional Text Generation with BERT (CG-BERT). CG-BERT effectively leverages a large pre-trained language model to generate text conditioned on the intent label. By modeling the utterance distribution with variational inference, CG-BERT can generate diverse utterances for the novel intents even with only a few utterances available. Experimental results show that CG-BERT achieves state-of-the-art performance on the GFSID task with 1-shot and 5-shot settings on two real-world datasets.

cs.CL↗