arXiv subjects

Xin Liu

Publications and source records attributed to Xin Liu.

At least 721 records · Page 40Linked to original sources

arXiv2022

Ethereum Fraud Detection with Heterogeneous Graph Neural Networks

While transactions with cryptocurrencies such as Ethereum are becoming more prevalent, fraud and other criminal transactions are not uncommon. Graph analysis algorithms and machine learning techniques detect suspicious transactions that lead to phishing in large transaction networks. Many graph neural network (GNN) models have been proposed to apply deep learning techniques to graph structures. Although there is research on phishing detection using GNN models in the Ethereum transaction network, models that address the scale of the number of vertices and edges and the imbalance of labels have not yet been studied. In this paper, we compared the model performance of GNN models on the actual Ethereum transaction network dataset and phishing reported label data to exhaustively compare and verify which GNN models and hyperparameters produce the best accuracy. Specifically, we evaluated the model performance of representative homogeneous GNN models which consider single-type nodes and edges and heterogeneous GNN models which support different types of nodes and edges. We showed that heterogeneous models had better model performance than homogeneous models. In particular, the RGCN model achieved the best performance in the overall metrics.

cs.LG↗

arXiv2022

Enhancing Local Feature Learning Using Diffusion for 3D Point Cloud Understanding

Learning point clouds is challenging due to the lack of connectivity information, i.e., edges. Although existing edge-aware methods can improve the performance by modeling edges, how edges contribute to the improvement is unclear. In this study, we propose a method that automatically learns to enhance/suppress edges while keeping the its working mechanism clear. First, we theoretically figure out how edge enhancement/suppression works. Second, we experimentally verify the edge enhancement/suppression behavior. Third, we empirically show that this behavior improves performance. In general, we observe that the proposed method achieves competitive performance in point cloud classification and segmentation tasks.

cs.CV↗

arXiv2022

Enhancing Local Geometry Learning for 3D Point Cloud via Decoupling Convolution

Modeling the local surface geometry is challenging in 3D point cloud understanding due to the lack of connectivity information. Most prior works model local geometry using various convolution operations. We observe that the convolution can be equivalently decomposed as a weighted combination of a local and a global component. With this observation, we explicitly decouple these two components so that the local one can be enhanced and facilitate the learning of local surface geometry. Specifically, we propose Laplacian Unit (LU), a simple yet effective architectural unit that can enhance the learning of local geometry. Extensive experiments demonstrate that networks equipped with LUs achieve competitive or superior performance on typical point cloud understanding tasks. Moreover, through establishing connections between the mean curvature flow, a further investigation of LU based on curvatures is made to interpret the adaptive smoothing and sharpening effect of LU. The code will be available.

cs.CV↗

arXiv2022

Very Large Array Multi-band Radio Imaging of the Triple AGN Candidate SDSS J0849+1114

Kpc-scale triple active galactic nuclei (AGNs), potential precursors of gravitationally-bound triple massive black holes (MBHs), are rarely seen objects and believed to play an important role in the evolution of MBHs and their host galaxies. In this work we present a multi-band (3.0, 6.0 10.0, and 15.0 GHz), high-resolution radio imaging of the triple AGN candidate, SDSS J0849+1114, using the Very Large Array. Two of the three nuclei (A and C) are detected at 3.0, 6.0, and 15 GHz for the first time, both exhibiting a steep spectrum over 3--15 GHz (with a spectral index $-0.90 \pm 0.05$ and $-1.03 \pm 0.04$) consistent with a synchrotron origin. Nucleus A, the strongest nucleus among the three, shows a double-sided jet, with the jet orientation changing by $\sim20^{\circ}$ between its inner 1" and the outer 5.5" (8.1 kpc) components, which may be explained as the MBH's angular momentum having been altered by merger-enhanced accretion. Nucleus C also shows a two-sided jet, with the western jet inflating into a radio lobe with an extent of 1.5" (2.2 kpc). The internal energy of the radio lobe is estimated to be $\rm 5.0 \times 10^{55}$ erg, for an equipartition magnetic field strength of $\rm \sim 160\ μG$. No significant radio emission is detected at all four frequencies for nucleus B, yielding an upper limit of 15, 15, 15, and 18 $\rm μJy\ beam^{-1}$ at 3.0, 6.0, 10.0, and 15.0 GHz, based on which we constrain the star formation rate in nucleus B to be $\lesssim 0.4~\rm M_{\odot}~yr^{-1}$.

astro-ph.GA↗

arXiv2022

Cross-Silo Federated Learning: Challenges and Opportunities

Federated learning (FL) is an emerging technology that enables the training of machine learning models from multiple clients while keeping the data distributed and private. Based on the participating clients and the model training scale, federated learning can be classified into two types: cross-device FL where clients are typically mobile devices and the client number can reach up to a scale of millions; cross-silo FL where clients are organizations or companies and the client number is usually small (e.g., within a hundred). While existing studies mainly focus on cross-device FL, this paper aims to provide an overview of the cross-silo FL. More specifically, we first discuss applications of cross-silo FL and outline its major challenges. We then provide a systematic overview of the existing approaches to the challenges in cross-silo FL by focusing on their connections and differences to cross-device FL. Finally, we discuss future directions and open issues that merit research efforts from the community.

cs.LG↗

arXiv2022

GNNSampler: Bridging the Gap between Sampling Algorithms of GNN and Hardware

Sampling is a critical operation in Graph Neural Network (GNN) training that helps reduce the cost. Previous literature has explored improving sampling algorithms via mathematical and statistical methods. However, there is a gap between sampling algorithms and hardware. Without consideration of hardware, algorithm designers merely optimize sampling at the algorithm level, missing the great potential of promoting the efficiency of existing sampling algorithms by leveraging hardware features. In this paper, we pioneer to propose a unified programming model for mainstream sampling algorithms, termed GNNSampler, covering the critical processes of sampling algorithms in various categories. Second, to leverage the hardware feature, we choose the data locality as a case study, and explore the data locality among nodes and their neighbors in a graph to alleviate irregular memory access in sampling. Third, we implement locality-aware optimizations in GNNSampler for various sampling algorithms to optimize the general sampling process. Finally, we emphatically conduct experiments on large graph datasets to analyze the relevance among training time, accuracy, and hardware-level metrics. Extensive experiments show that our method is universal to mainstream sampling algorithms and helps significantly reduce the training time, especially in large-scale graphs.

cs.LG↗

arXiv2022

Boosting Graph Structure Learning with Dummy Nodes

With the development of graph kernels and graph representation learning, many superior methods have been proposed to handle scalability and oversmoothing issues on graph structure learning. However, most of those strategies are designed based on practical experience rather than theoretical analysis. In this paper, we use a particular dummy node connecting to all existing vertices without affecting original vertex and edge properties. We further prove that such the dummy node can help build an efficient monomorphic edge-to-vertex transform and an epimorphic inverse to recover the original graph back. It also indicates that adding dummy nodes can preserve local and global structures for better graph representation learning. We extend graph kernels and graph neural networks with dummy nodes and conduct experiments on graph classification and subgraph isomorphism matching tasks. Empirical results demonstrate that taking graphs with dummy nodes as input significantly boosts graph structure learning, and using their edge-to-vertex graphs can also achieve similar results. We also discuss the gain of expressive power from the dummy in neural networks.

cs.LG↗

arXiv2022

A Multi-task Framework for Infrared Small Target Detection and Segmentation

Due to the complicated background and noise of infrared images, infrared small target detection is one of the most difficult problems in the field of computer vision. In most existing studies, semantic segmentation methods are typically used to achieve better results. The centroid of each target is calculated from the segmentation map as the detection result. In contrast, we propose a novel end-to-end framework for infrared small target detection and segmentation in this paper. First, with the use of UNet as the backbone to maintain resolution and semantic information, our model can achieve a higher detection accuracy than other state-of-the-art methods by attaching a simple anchor-free head. Then, a pyramid pool module is used to further extract features and improve the precision of target segmentation. Next, we use semantic segmentation tasks that pay more attention to pixel-level features to assist in the training process of object detection, which increases the average precision and allows the model to detect some targets that were previously not detectable. Furthermore, we develop a multi-task framework for infrared small target detection and segmentation. Our multi-task learning model reduces complexity by nearly half and speeds up inference by nearly twice compared to the composite single-task model, while maintaining accuracy. The code and models are publicly available at https://github.com/Chenastron/MTUNet.

cs.CV↗

arXiv2022

The Exactness of the $\ell_1$ Penalty Function for a Class of Mathematical Programs with Generalized Complementarity Constraints

In a Mathematical Program with Generalized Complementarity Constraints (MPGCC), complementarity relationships are imposed between each pair of variable blocks. MPGCC includes the traditional Mathematical Program with Complementarity Constraints (MPCC) as a special case. On account of the disjunctive feasible region, MPCC and MPGCC are generally difficult to handle. The $\ell_1$ penalty method, often adopted in computation, opens a way of circumventing the difficulty. Yet it remains unclear about the exactness of the $\ell_1$ penalty function, namely, whether there exists a sufficiently large penalty parameter so that the penalty problem shares the optimal solution set with the original one. In this paper, we consider a class of MPGCCs that are of multi-affine objective functions. This problem class finds applications in various fields, e.g., the multi-marginal optimal transport problems in many-body quantum physics and the pricing problem in network transportation. We first provide an instance from this class, the exactness of whose $\ell_1$ penalty function cannot be derived by existing tools. We then establish the exactness results under rather mild conditions. Our results cover those existing ones for MPCC and apply to multi-block contexts.

math.OC↗

arXiv2022

SCAMPS: Synthetics for Camera Measurement of Physiological Signals

The use of cameras and computational algorithms for noninvasive, low-cost and scalable measurement of physiological (e.g., cardiac and pulmonary) vital signs is very attractive. However, diverse data representing a range of environments, body motions, illumination conditions and physiological states is laborious, time consuming and expensive to obtain. Synthetic data have proven a valuable tool in several areas of machine learning, yet are not widely available for camera measurement of physiological states. Synthetic data offer "perfect" labels (e.g., without noise and with precise synchronization), labels that may not be possible to obtain otherwise (e.g., precise pixel level segmentation maps) and provide a high degree of control over variation and diversity in the dataset. We present SCAMPS, a dataset of synthetics containing 2,800 videos (1.68M frames) with aligned cardiac and respiratory signals and facial action intensities. The RGB frames are provided alongside segmentation maps. We provide precise descriptive statistics about the underlying waveforms, including inter-beat interval, heart rate variability, and pulse arrival time. Finally, we present baseline results training on these synthetic data and testing on real-world datasets to illustrate generalizability.

cs.CV↗

arXiv2022

On the effect of fast rotation and vertical viscosity on the lifespan of the $3D$ primitive equations

We study the effect of the fast rotation and vertical viscosity on the lifespan of solutions to the three-dimensional primitive equations (also known as the hydrostatic Navier-Stokes equations) with impermeable and stress-free boundary conditions. Firstly, for a short time interval, independent of the rate of rotation $|Ω|$, we establish the local well-posedness of solutions with initial data that is analytic in the horizontal variables and only $L^2$ in the vertical variable. Moreover, it is shown that the solutions immediately become analytic in all the variables with increasing-in-time (at least linearly) radius of analyticity in the vertical variable for as long as the solutions exist. On the other hand, the radius of analyticity in the horizontal variables might decrease with time, but as long as it remains positive the solution exists. Secondly, with fast rotation, i.e., large $|Ω|$, we show that the existence time of the solution can be prolonged, with "well-prepared" initial data. Finally, in the case of two spatial dimensions with $Ω=0$, we establish the global well-posedness provided that the initial data is small enough. The smallness condition on the initial data depends on the vertical viscosity and the initial radius of analyticity in the horizontal variables.

math.AP↗

arXiv2022

Optical Variability of Quasars with 20-Year Photometric Light Curves

We study the optical $gri$ photometric variability of a sample of 190 quasars within the SDSS Stripe 82 region that have long-term photometric coverage during $\sim 1998-2020$ with SDSS, PanSTARRS-1, the Dark Energy Survey, and dedicated follow-up monitoring with Blanco 4m/DECam. With on average $\sim 200$ nightly epochs per quasar per filter band, we improve the parameter constraints from a Damped Random Walk (DRW) model fit to the light curves over previous studies with 10-15 yr baselines and $\lesssim 100$ epochs. We find that the average damping timescale $τ_{\rm DRW}$ continues to rise with increased baseline, reaching a median value of $\sim 750$ days ($g$ band) in the rest-frame of these quasars using the 20-yr light curves. Some quasars may have gradual, long-term trends in their light curves, suggesting that either the DRW fit requires very long baselines to converge, or that the underlying variability is more complex than a single DRW process for these quasars. Using a subset of quasars with better-constrained $τ_{\rm DRW}$ (less than 20\% of the baseline), we confirm a weak wavelength dependence of $τ_{\rm DRW}\propto λ^{0.51\pm0.20}$. We further quantify optical variability of these quasars over days to decades timescales using structure function (SF) and power spectrum density (PSD) analyses. The SF and PSD measurements qualitatively confirm the measured (hundreds of days) damping timescales from the DRW fits. However, the ensemble PSD is steeper than that of a DRW on timescales less than $\sim$ a month for these luminous quasars, and this second break point correlates with the longer DRW damping timescale.

astro-ph.GA↗

arXiv2022

A force-based gradient descent method for $\mathit{\text{ab initio}}$ atomic structure relaxation

Force-based algorithms for $\mathit{\text{ab initio}}$ atomic structure relaxation, such as conjugate gradient methods, usually get stuck in the line minimization processes along search directions, where expensive $\mathit{\text{ab initio}}$ calculations are triggered frequently to test trial positions before locating the next iterate. We present a force-based gradient descent method, WANBB, that circumvents the deficiency. At each iteration, WANBB enters the line minimization process with a trial stepsize capturing the local curvature of the energy surface. The exit is controlled by an unrestrictive criterion that tends to accept early trials. These two ingredients streamline the line minimization process in WANBB. The numerical simulations on nearly 80 systems with good universality demonstrate the considerable compression of WANBB on the cost for the unaccepted trials compared with conjugate gradient methods. We also observe across the board significant and universal speedups as well as the superior robustness of WANBB over several widely used methods. The latter point is theoretically established. The implementation of WANBB is pretty simple, in that no a priori physical knowledge is required and only two parameters are present without tuning.

physics.comp-ph↗

arXiv2022

Autonomous Smart Grid Fault Detection

Smart grid plays a crucial role for the smart society and the upcoming carbon neutral society. Achieving autonomous smart grid fault detection is critical for smart grid system state awareness, maintenance and operation. This paper focuses on fault monitoring in smart grid and discusses the inherent technical challenges and solutions. In particular, we first present the basic principles of smart grid fault detection. Then, we explain the new requirements for autonomous smart grid fault detection, the technical challenges and their possible solutions. A case study is introduced, as a preliminary study for autonomous smart grid fault detection. In addition, we highlight relevant directions for future research.

cs.DC↗

arXiv2022

Exploration, Exploitation, and Engagement in Multi-Armed Bandits with Abandonment

Multi-armed bandit (MAB) is a classic model for understanding the exploration-exploitation trade-off. The traditional MAB model for recommendation systems assumes the user stays in the system for the entire learning horizon. In new online education platforms such as ALEKS or new video recommendation systems such as TikTok and YouTube Shorts, the amount of time a user spends on the app depends on how engaging the recommended contents are. Users may temporarily leave the system if the recommended items cannot engage the users. To understand the exploration, exploitation, and engagement in these systems, we propose a new model, called MAB-A where "A" stands for abandonment and the abandonment probability depends on the current recommended item and the user's past experience (called state). We propose two algorithms, ULCB and KL-ULCB, both of which do more exploration (being optimistic) when the user likes the previous recommended item and less exploration (being pessimistic) when the user does not like the previous item. We prove that both ULCB and KL-ULCB achieve logarithmic regret, $O(\log K)$, where $K$ is the number of visits (or episodes). Furthermore, the regret bound under KL-ULCB is asymptotically sharp. We also extend the proposed algorithms to the general-state setting. Simulation results confirm our theoretical analysis and show that the proposed algorithms have significantly lower regrets than the traditional UCB and KL-UCB, and Q-learning-based algorithms.

cs.LG↗

arXiv2022

FastAdaBelief: Improving Convergence Rate for Belief-based Adaptive Optimizers by Exploiting Strong Convexity

AdaBelief, one of the current best optimizers, demonstrates superior generalization ability compared to the popular Adam algorithm by viewing the exponential moving average of observed gradients. AdaBelief is theoretically appealing in that it has a data-dependent $O(\sqrt{T})$ regret bound when objective functions are convex, where $T$ is a time horizon. It remains however an open problem whether the convergence rate can be further improved without sacrificing its generalization ability. %on how to exploit strong convexity to further improve the convergence rate of AdaBelief. To this end, we make a first attempt in this work and design a novel optimization algorithm called FastAdaBelief that aims to exploit its strong convexity in order to achieve an even faster convergence rate. In particular, by adjusting the step size that better considers strong convexity and prevents fluctuation, our proposed FastAdaBelief demonstrates excellent generalization ability as well as superior convergence. As an important theoretical contribution, we prove that FastAdaBelief attains a data-dependant $O(\log T)$ regret bound, which is substantially lower than AdaBelief. On the empirical side, we validate our theoretical analysis with extensive experiments in both scenarios of strong and non-strong convexity on three popular baseline models. Experimental results are very encouraging: FastAdaBelief converges the quickest in comparison to all mainstream algorithms while maintaining an excellent generalization ability, in cases of both strong or non-strong convexity. FastAdaBelief is thus posited as a new benchmark model for the research community.

cs.LG↗

arXiv2022

Cosmo-Paleontology: Statistics of Fossil Groups in a Gravity-Only Simulation

We present a detailed study of fossil group candidates identified in "Last Journey", a gravity-only cosmological simulation covering a $(3.4\, h^{-1}\mathrm{Gpc})^3$ volume with a particle mass resolution of $m_p \approx 2.7 \times 10^9\, h^{-1}\mathrm{M}_\odot$. The simulation allows us to simultaneously capture a large number of group-scale halos and to resolve their internal structure. Historically, fossil groups have been characterized by high X-ray brightness and a large luminosity gap between the brightest and second brightest galaxy in the group. In order to identify candidate halos that host fossil groups, we use halo merger tree information to introduce two parameters: a luminous merger mass threshold ($M_\mathrm{LM}$) and a last luminous merger redshift cut-off ($z_\mathrm{LLM}$). The final parameter choices are informed by observational data and allow us to identify a plausible fossil group sample from the simulation. The candidate halos are characterized by reduced substructure and are therefore less likely to host bright galaxies beyond the brightest central galaxy. We carry out detailed studies of this sample, including analysis of halo properties and clustering. We find that our simple assumptions lead to fossil group candidates that form early, have higher concentrations, and are more relaxed compared to other halos in the same mass range.

astro-ph.CO↗

arXiv2022

An Inexact Augmented Lagrangian Algorithm for Training Leaky ReLU Neural Network with Group Sparsity

The leaky ReLU network with a group sparse regularization term has been widely used in the recent years. However, training such a network yields a nonsmooth nonconvex optimization problem and there exists a lack of approaches to compute a stationary point deterministically. In this paper, we first resolve the multi-layer composite term in the original optimization problem by introducing auxiliary variables and additional constraints. We show the new model has a nonempty and bounded solution set and its feasible set satisfies the Mangasarian-Fromovitz constraint qualification. Moreover, we show the relationship between the new model and the original problem. Remarkably, we propose an inexact augmented Lagrangian algorithm for solving the new model and show the convergence of the algorithm to a KKT point. Numerical experiments demonstrate that our algorithm is more efficient for training sparse leaky ReLU neural networks than some well-known algorithms.

math.OC↗