Searcharxiv⌕ Search

arXiv subjects

Li Sun

Publications and source records attributed to Li Sun.

At least 199 records · Page 11Linked to original sources
arXiv2020

Enhancing Graph Neural Network-based Fraud Detectors against Camouflaged Fraudsters

Graph Neural Networks (GNNs) have been widely applied to fraud detection problems in recent years, revealing the suspiciousness of nodes by aggregating their neighborhood information via different relations. However, few prior works have noticed the camouflage behavior of fraudsters, which could hamper the performance of GNN-based fraud detectors during the aggregation process. In this paper, we introduce two types of camouflages based on recent empirical studies, i.e., the feature camouflage and the relation camouflage. Existing GNNs have not addressed these two camouflages, which results in their poor performance in fraud detection problems. Alternatively, we propose a new model named CAmouflage-REsistant GNN (CARE-GNN), to enhance the GNN aggregation process with three unique modules against camouflages. Concretely, we first devise a label-aware similarity measure to find informative neighboring nodes. Then, we leverage reinforcement learning (RL) to find the optimal amounts of neighbors to be selected. Finally, the selected neighbors across different relations are aggregated together. Comprehensive experiments on two real-world fraud datasets demonstrate the effectiveness of the RL algorithm. The proposed CARE-GNN also outperforms state-of-the-art GNNs and GNN-based fraud detectors. We integrate all GNN-based fraud detectors as an opensource toolbox: https://github.com/safe-graph/DGFraud. The CARE-GNN code and datasets are available at https://github.com/YingtongDou/CARE-GNN.

cs.SI↗
arXiv2020

EU Long-term Dataset with Multiple Sensors for Autonomous Driving

The field of autonomous driving has grown tremendously over the past few years, along with the rapid progress in sensor technology. One of the major purposes of using sensors is to provide environment perception for vehicle understanding, learning and reasoning, and ultimately interacting with the environment. In this paper, we first introduce a multisensor platform allowing vehicle to perceive its surroundings and locate itself in a more efficient and accurate way. The platform integrates eleven heterogeneous sensors including various cameras and lidars, a radar, an IMU (Inertial Measurement Unit), and a GPS-RTK (Global Positioning System / Real-Time Kinematic), while exploits a ROS (Robot Operating System) based software to process the sensory data. Then, we present a new dataset (https://epan-utbm.github.io/utbm_robocar_dataset/) for autonomous driving captured many new research challenges (e.g. highly dynamic environment), and especially for long-term autonomy (e.g. creating and maintaining maps), collected with our instrumented vehicle, publicly available to the community.

cs.RO↗
arXiv2020

Novel View Synthesis on Unpaired Data by Conditional Deformable Variational Auto-Encoder

Novel view synthesis often needs the paired data from both the source and target views. This paper proposes a view translation model under cVAE-GAN framework without requiring the paired data. We design a conditional deformable module (CDM) which uses the view condition vectors as the filters to convolve the feature maps of the main branch in VAE. It generates several pairs of displacement maps to deform the features, like the 2D optical flows. The results are fed into the deformed feature based normalization module (DFNM), which scales and offsets the main branch feature, given its deformed one as the input from the side branch. Taking the advantage of the CDM and DFNM, the encoder outputs a view-irrelevant posterior, while the decoder takes the code drawn from it to synthesize the reconstructed and the viewtranslated images. To further ensure the disentanglement between the views and other factors, we add adversarial training on the code. The results and ablation studies on MultiPIE and 3D chair datasets validate the effectiveness of the framework in cVAE and the designed module.

cs.CV↗
arXiv2020

Learning Posterior and Prior for Uncertainty Modeling in Person Re-Identification

Data uncertainty in practical person reID is ubiquitous, hence it requires not only learning the discriminative features, but also modeling the uncertainty based on the input. This paper proposes to learn the sample posterior and the class prior distribution in the latent space, so that not only representative features but also the uncertainty can be built by the model. The prior reflects the distribution of all data in the same class, and it is the trainable model parameters. While the posterior is the probability density of a single sample, so it is actually the feature defined on the input. We assume that both of them are in Gaussian form. To simultaneously model them, we put forward a distribution loss, which measures the KL divergence from the posterior to the priors in the manner of supervised learning. In addition, we assume that the posterior variance, which is essentially the uncertainty, is supposed to have the second-order characteristic. Therefore, a $Σ-$net is proposed to compute it by the high order representation from its input. Extensive experiments have been carried out on Market1501, DukeMTMC, MARS and noisy dataset as well.

cs.CV↗
arXiv2020

Disentangling the Spatial Structure and Style in Conditional VAE

This paper aims to disentangle the latent space in cVAE into the spatial structure and the style code, which are complementary to each other, with one of them $z_s$ being label relevant and the other $z_u$ irrelevant. The generator is built by a connected encoder-decoder and a label condition mapping network. Depending on whether the label is related with the spatial structure, the output $z_s$ from the condition mapping network is used either as a style code or a spatial structure code. The encoder provides the label irrelevant posterior from which $z_u$ is sampled. The decoder employs $z_s$ and $z_u$ in each layer by adaptive normalization like SPADE or AdaIN. Extensive experiments on two datasets with different types of labels show the effectiveness of our method.

cs.CV↗
arXiv2020

Localising Faster: Efficient and precise lidar-based robot localisation in large-scale environments

This paper proposes a novel approach for global localisation of mobile robots in large-scale environments. Our method leverages learning-based localisation and filtering-based localisation, to localise the robot efficiently and precisely through seeding Monte Carlo Localisation (MCL) with a deep-learned distribution. In particular, a fast localisation system rapidly estimates the 6-DOF pose through a deep-probabilistic model (Gaussian Process Regression with a deep kernel), then a precise recursive estimator refines the estimated robot pose according to the geometric alignment. More importantly, the Gaussian method (i.e. deep probabilistic localisation) and non-Gaussian method (i.e. MCL) can be integrated naturally via importance sampling. Consequently, the two systems can be integrated seamlessly and mutually benefit from each other. To verify the proposed framework, we provide a case study in large-scale localisation with a 3D lidar sensor. Our experiments on the Michigan NCLT long-term dataset show that the proposed method is able to localise the robot in 1.94 s on average (median of 0.8 s) with precision 0.75~m in a large-scale environment of approximately 0.5 km2.

cs.RO↗
arXiv2020

Quantum locally linear embedding for nonlinear dimensionality reduction

Reducing the dimension of nonlinear data is crucial in data processing and visualization. The locally linear embedding algorithm (LLE) is specifically a representative nonlinear dimensionality reduction method with well maintaining the original manifold structure. In this paper, we present two implementations of the quantum locally linear embedding algorithm (QLLE) to perform the nonlinear dimensionality reduction on quantum devices. One implementation, the linear-algebra-based QLLE algorithm, utilizes quantum linear algebra subroutines to reduce the dimension of the given data. The other implementation, the variational quantum locally linear embedding algorithm (VQLLE) utilizes a variational hybrid quantum-classical procedure to acquire the low-dimensional data. The classical LLE algorithm requires polynomial time complexity of $N$, where $N$ is the global number of the original high-dimensional data. Compared with the classical LLE, the linear-algebra-based QLLE achieves quadratic speedup in the number and dimension of the given data. The VQLLE can be implemented on the near term quantum devices in two different designs. In addition, the numerical experiments are presented to demonstrate that the two implementations in our work can achieve the procedure of locally linear embedding.

quant-ph↗
arXiv2020

Graph-Bert: Only Attention is Needed for Learning Graph Representations

The dominant graph neural networks (GNNs) over-rely on the graph links, several serious performance problems with which have been witnessed already, e.g., suspended animation problem and over-smoothing problem. What's more, the inherently inter-connected nature precludes parallelization within the graph, which becomes critical for large-sized graph, as memory constraints limit batching across the nodes. In this paper, we will introduce a new graph neural network, namely GRAPH-BERT (Graph based BERT), solely based on the attention mechanism without any graph convolution or aggregation operators. Instead of feeding GRAPH-BERT with the complete large input graph, we propose to train GRAPH-BERT with sampled linkless subgraphs within their local contexts. GRAPH-BERT can be learned effectively in a standalone mode. Meanwhile, a pre-trained GRAPH-BERT can also be transferred to other application tasks directly or with necessary fine-tuning if any supervised label information or certain application oriented objective is available. We have tested the effectiveness of GRAPH-BERT on several graph benchmark datasets. Based the pre-trained GRAPH-BERT with the node attribute reconstruction and structure recovery tasks, we further fine-tune GRAPH-BERT on node classification and graph clustering tasks specifically. The experimental results have demonstrated that GRAPH-BERT can out-perform the existing GNNs in both the learning effectiveness and efficiency.

cs.LG↗
arXiv2019

DNA: Dynamic Social Network Alignment

Social network alignment, aligning different social networks on their common users, is receiving dramatic attention from both academic and industry. All existing studies consider the social network to be static and neglect its inherent dynamics. In fact, the dynamics of social networks contain the discriminative pattern of an individual, which can be leveraged to facilitate social network alignment. Hence, we for the first time propose to study the problem of aligning dynamic social networks. Towards this end, we propose a novel Dynamic social Network Alignment (DNA) framework, a unified optimization approach over deep neural architectures, to unfold the fruitful dynamics to perform alignment. However, it faces tremendous challenges in both modeling and optimization: (1) To model the intra-network dynamics, we explore the local dynamics of the latent pattern in friending evolvement and the global consistency of the representation similarity with neighbors. We design a novel deep neural architecture to obtain the dual embedding capturing local dynamics and global consistency for each user. (2) To model the inter-network alignment, we exploit the underlying identity of an individual from the dual embedding in each dynamic social network. We design a unified optimization approach interplaying proposed deep neural architectures to construct a common subspace of identity embeddings. (3) To address this optimization problem, we design an effective alternating algorithm with solid theoretical guarantees.We conduct extensive experiments on real-world datasets and show that the proposed DNA framework substantially outperforms the state-of-the-art methods.

cs.SI↗
arXiv2019

CO~($J=1-0$) Observations toward Filamentary Molecular Clouds in the Galactic Region with $l = [169\arcdeg.75, 174\arcdeg.75], b = [-0\arcdeg.75, 0\arcdeg.5]$

We present observations of the CO isotopologues ($^{12}$CO, $^{13}$CO, and C$^{18}$O) toward the Galactic region with $169\arcdeg.75 \leqslant l \leqslant 174\arcdeg.75$ and $-0\arcdeg.75 \leqslant b \leqslant 0\arcdeg.5$, using the Purple Mountain Observatory 13.7~m millimeter-wavelength telescope. Based on the $^{13}$CO~($J = 1-0$) data, we find five molecular clouds within the velocity range between $-$25 and 8~km~s$^{-1}$ that are all characterized by conspicuous filamentary structures. We have identified eight filaments with a length of 6.38--28.45~pc, a mean H$_2$ column density of 0.70$\times$10$^{21}$--6.53$\times$10$^{21}$~cm$^{-2}$, and a line mass of 20.24--161.91~$M_\sun$ pc$^{-1}$, assuming a distance of $\sim$1.7~kpc. Gaussian fittings to the inner parts of the radial density profiles lead to a mean FWHM width of 1.13$\pm$0.01~pc. The velocity structures of most filaments present continuous distributions with slight velocity gradients. We find that turbulence is the dominant internal pressure to support the fragmentation of filaments instead of thermal pressure. Most filaments have virial parameters smaller than 2; thus, they are gravitationally bound. Four filaments have an LTE line mass close to the virial line mass. We further extract dense clumps using the $^{13}$CO data and find that 64$\%$ of the clumps are associated with the filaments. According to the complementary IR data, most filaments have associated Class~II young stellar objects. Class~I objects are mainly found to be located in the filaments with a virial parameter close to 1. Within two virialized filaments, $^{12}$CO outflows have been detected, indicating ongoing star-forming activity therein.

astro-ph.GA↗
arXiv2019

Customizing Student Networks From Heterogeneous Teachers via Adaptive Knowledge Amalgamation

A massive number of well-trained deep networks have been released by developers online. These networks may focus on different tasks and in many cases are optimized for different datasets. In this paper, we study how to exploit such heterogeneous pre-trained networks, known as teachers, so as to train a customized student network that tackles a set of selective tasks defined by the user. We assume no human annotations are available, and each teacher may be either single- or multi-task. To this end, we introduce a dual-step strategy that first extracts the task-specific knowledge from the heterogeneous teachers sharing the same sub-task, and then amalgamates the extracted knowledge to build the student network. To facilitate the training, we employ a selective learning scheme where, for each unlabelled sample, the student learns adaptively from only the teacher with the least prediction ambiguity. We evaluate the proposed approach on several datasets and experimental results demonstrate that the student, learned by such adaptive knowledge amalgamation, achieves performances even better than those of the teachers.

cs.CV↗
arXiv2019

LabelECG: A Web-based Tool for Distributed Electrocardiogram Annotation

Electrocardiography plays an essential role in diagnosing and screening cardiovascular diseases in daily healthcare. Deep neural networks have shown the potentials to improve the accuracies of arrhythmia detection based on electrocardiograms (ECGs). However, more ECG records with ground truth are needed to promote the development and progression of deep learning techniques in automatic ECG analysis. Here we propose a web-based tool for ECG viewing and annotating, LabelECG. With the facilitation of unified data management, LabelECG is able to distribute large cohorts of ECGs to dozens of technicians and physicians, who can simultaneously make annotations through web-browsers on PCs, tablets and cell phones. Along with the doctors from four hospitals in China, we applied LabelECG to support the annotations of about 15,000 12-lead resting ECG records in three months. These annotated ECGs have successfully supported the First China ECG intelligent Competition. La-belECG will be freely accessible on the Internet to support similar researches, and will also be upgraded through future works.

cs.DB↗
arXiv2019

Identification of Noncausal Models by Quantile Autoregressions

We propose a model selection criterion to detect purely causal from purely noncausal models in the framework of quantile autoregressions (QAR). We also present asymptotics for the i.i.d. case with regularly varying distributed innovations in QAR. This new modelling perspective is appealing for investigating the presence of bubbles in economic and financial time series, and is an alternative to approximate maximum likelihood methods. We illustrate our analysis using hyperinflation episodes in Latin American countries.

econ.EM↗
arXiv2019

Disentangling Latent Space for VAE by Label Relevant/Irrelevant Dimensions

VAE requires the standard Gaussian distribution as a prior in the latent space. Since all codes tend to follow the same prior, it often suffers the so-called "posterior collapse". To avoid this, this paper introduces the class specific distribution for the latent code. But different from CVAE, we present a method for disentangling the latent space into the label relevant and irrelevant dimensions, $\bm{\mathrm{z}}_s$ and $\bm{\mathrm{z}}_u$, for a single input. We apply two separated encoders to map the input into $\bm{\mathrm{z}}_s$ and $\bm{\mathrm{z}}_u$ respectively, and then give the concatenated code to the decoder to reconstruct the input. The label irrelevant code $\bm{\mathrm{z}}_u$ represent the common characteristics of all inputs, hence they are constrained by the standard Gaussian, and their encoder is trained in amortized variational inference way, like VAE. While $\bm{\mathrm{z}}_s$ is assumed to follow the Gaussian mixture distribution in which each component corresponds to a particular class. The parameters for the Gaussian components in $\bm{\mathrm{z}}_s$ encoder are optimized by the label supervision in a global stochastic way. In theory, we show that our method is actually equivalent to adding a KL divergence term on the joint distribution of $\bm{\mathrm{z}}_s$ and the class label $c$, and it can directly increase the mutual information between $\bm{\mathrm{z}}_s$ and the label $c$. Our model can also be extended to GAN by adding a discriminator in the pixel domain so that it produces high quality and diverse images.

cs.CV↗
arXiv2019

Resilience Evaluation Framework for Integrated Critical Infrastructure-Community Systems under Seismic Hazard

In this paper, an agent-based modeling framework to evaluate the seismic resilience of an integrated system comprising the community and its civil infrastructure systems is proposed. Specifically, an agent-based model of the recovery process of civil infrastructure systems is incorporated into a previously developed compositional supply/demand seismic resilience quantification framework. The proposed model represents the behavior of the operators of civil infrastructure systems as they strive to recover their functionality in the aftermath of an earthquake as well as their mutual interactions and the interactions with the community they provide services to. A case study of the seismic resilience of a virtual system comprising the electric power supply system, the transportation system and the community (EPSS-TS-Community system) is conducted using the proposed framework. A parametric investigation is carried out to examine the effect of different earthquake magnitude scenarios as well as different behaviors of the involved agents and their interaction on the seismic resilience of the virtual system. It was demonstrated that the proposed agent-based modeling approach is effective in representing the interactions among different participants in the recovery process. It was also revealed that timely and well-planned intervention in the recovery process can be effective in alleviating the post-earthquake lack of resilience resulting from the insufficient supply of civil infrastructure service to meet the community demands. Therefore, the proposed framework could be employed to formulate the recovery trajectory of the intertwined socio-technical system subjected to different earthquake scenarios. The interplay among different agents, as well as the interdependence between civil infrastructure systems is found to profoundly shape the recovery path for this integrated EPSS-TS-Community system.

physics.soc-ph↗
arXiv2019

Agent-Based Recovery Model for Seismic Resilience Evaluation of Electrified Communities

In this paper, an agent-based framework to quantify the seismic resilience of an Electric Power Supply System (EPSS) and the community it serves is presented. Within the framework, the loss and restoration of the EPSS power generation and delivery capacity and of the power demand from the served community are used to assess the electric power deficit during the damage absorption and recovery processes. Damage to the components of the EPSS and of the community built environment is evaluated using the seismic fragility functions. The restoration of the community electric power demand is evaluated using the seismic recovery functions. However, the post-earthquake EPSS recovery process is modeled using an agent-based model with two agents, the EPSS Operator and the Community Administrator. The resilience of the EPSS-Community system is quantified using direct, EPSS-related, and societal, community-related, indicators. Parametric studies are carried out to quantify the influence of different seismic hazard scenarios, agent characteristics, and power dispatch strategies on the EPSS-Community seismic resilience. The use of the agent-based modeling framework enabled a rational formulation of the post-earthquake recovery phase and highlighted the interaction between the EPSS and the community in the recovery process not quantified in resilience models developed to date. Furthermore, it shows that the resilience of different community sectors can be enhanced by different power dispatch strategies. The proposed agent-based EPSS-Community system resilience quantification framework can be used to develop better community and infrastructure system risk governance policies.

physics.soc-ph↗
arXiv2018

Amalgamating Knowledge towards Comprehensive Classification

With the rapid development of deep learning, there have been an unprecedentedly large number of trained deep network models available online. Reusing such trained models can significantly reduce the cost of training the new models from scratch, if not infeasible at all as the annotations used for the training original networks are often unavailable to public. We propose in this paper to study a new model-reusing task, which we term as \emph{knowledge amalgamation}. Given multiple trained teacher networks, each of which specializes in a different classification problem, the goal of knowledge amalgamation is to learn a lightweight student model capable of handling the comprehensive classification. We assume no other annotations except the outputs from the teacher models are available, and thus focus on extracting and amalgamating knowledge from the multiple teachers. To this end, we propose a pilot two-step strategy to tackle the knowledge amalgamation task, by learning first the compact feature representations from teachers and then the network parameters in a layer-wise manner so as to build the student model. We apply this approach to four public datasets and obtain very encouraging results: even without any human annotation, the obtained student model is competent to handle the comprehensive classification task and in most cases outperforms the teachers in individual sub-tasks.

cs.CV↗