arXiv subjects

Rui Xu

Publications and source records attributed to Rui Xu.

At least 217 records · Page 12Linked to original sources

arXiv2021

SSDL: Self-Supervised Dictionary Learning

The label-embedded dictionary learning (DL) algorithms generate influential dictionaries by introducing discriminative information. However, there exists a limitation: All the label-embedded DL methods rely on the labels due that this way merely achieves ideal performances in supervised learning. While in semi-supervised and unsupervised learning, it is no longer sufficient to be effective. Inspired by the concept of self-supervised learning (e.g., setting the pretext task to generate a universal model for the downstream task), we propose a Self-Supervised Dictionary Learning (SSDL) framework to address this challenge. Specifically, we first design a $p$-Laplacian Attention Hypergraph Learning (pAHL) block as the pretext task to generate pseudo soft labels for DL. Then, we adopt the pseudo labels to train a dictionary from a primary label-embedded DL method. We evaluate our SSDL on two human activity recognition datasets. The comparison results with other state-of-the-art methods have demonstrated the efficiency of SSDL.

cs.LG↗

arXiv2021

Uncertainty quantification and inverse modeling for subsurface flow in 3D heterogeneous formations using a theory-guided convolutional encoder-decoder network

We build surrogate models for dynamic 3D subsurface single-phase flow problems with multiple vertical producing wells. The surrogate model provides efficient pressure estimation of the entire formation at any timestep given a stochastic permeability field, arbitrary well locations and penetration lengths, and a timestep matrix as inputs. The well production rate or bottom hole pressure can then be determined based on Peaceman's formula. The original surrogate modeling task is transformed into an image-to-image regression problem using a convolutional encoder-decoder neural network architecture. The residual of the governing flow equation in its discretized form is incorporated into the loss function to impose theoretical guidance on the model training process. As a result, the accuracy and generalization ability of the trained surrogate models are significantly improved compared to fully data-driven models. They are also shown to have flexible extrapolation ability to permeability fields with different statistics. The surrogate models are used to conduct uncertainty quantification considering a stochastic permeability field, as well as to infer unknown permeability information based on limited well production data and observation data of formation properties. Results are shown to be in good agreement with traditional numerical simulation tools, but computational efficiency is dramatically improved.

cs.CE↗

arXiv2021

BSC: Block-based Stochastic Computing to Enable Accurate and Efficient TinyML

Along with the progress of AI democratization, machine learning (ML) has been successfully applied to edge applications, such as smart phones and automated driving. Nowadays, more applications require ML on tiny devices with extremely limited resources, like implantable cardioverter defibrillator (ICD), which is known as TinyML. Unlike ML on the edge, TinyML with a limited energy supply has higher demands on low-power execution. Stochastic computing (SC) using bitstreams for data representation is promising for TinyML since it can perform the fundamental ML operations using simple logical gates, instead of the complicated binary adder and multiplier. However, SC commonly suffers from low accuracy for ML tasks due to low data precision and inaccuracy of arithmetic units. Increasing the length of the bitstream in the existing works can mitigate the precision issue but incur higher latency. In this work, we propose a novel SC architecture, namely Block-based Stochastic Computing (BSC). BSC divides inputs into blocks, such that the latency can be reduced by exploiting high data parallelism. Moreover, optimized arithmetic units and output revision (OUR) scheme are proposed to improve accuracy. On top of it, a global optimization approach is devised to determine the number of blocks, which can make a better latency-power trade-off. Experimental results show that BSC can outperform the existing designs in achieving over 10% higher accuracy on ML tasks and over 6 times power reduction.

cs.LG↗

arXiv2021

MHFC: Multi-Head Feature Collaboration for Few-Shot Learning

Few-shot learning (FSL) aims to address the data-scarce problem. A standard FSL framework is composed of two components: (1) Pre-train. Employ the base data to generate a CNN-based feature extraction model (FEM). (2) Meta-test. Apply the trained FEM to acquire the novel data's features and recognize them. FSL relies heavily on the design of the FEM. However, various FEMs have distinct emphases. For example, several may focus more attention on the contour information, whereas others may lay particular emphasis on the texture information. The single-head feature is only a one-sided representation of the sample. Besides the negative influence of cross-domain (e.g., the trained FEM can not adapt to the novel class flawlessly), the distribution of novel data may have a certain degree of deviation compared with the ground truth distribution, which is dubbed as distribution-shift-problem (DSP). To address the DSP, we propose Multi-Head Feature Collaboration (MHFC) algorithm, which attempts to project the multi-head features (e.g., multiple features extracted from a variety of FEMs) to a unified space and fuse them to capture more discriminative information. Typically, first, we introduce a subspace learning method to transform the multi-head features to aligned low-dimensional representations. It corrects the DSP via learning the feature with more powerful discrimination and overcomes the problem of inconsistent measurement scales from different head features. Then, we design an attention block to update combination weights for each head feature automatically. It comprehensively considers the contribution of various perspectives and further improves the discrimination of features. We evaluate the proposed method on five benchmark datasets (including cross-domain experiments) and achieve significant improvements of 2.1%-7.8% compared with state-of-the-arts.

cs.CV↗

arXiv2021

Texture Memory-Augmented Deep Patch-Based Image Inpainting

Patch-based methods and deep networks have been employed to tackle image inpainting problem, with their own strengths and weaknesses. Patch-based methods are capable of restoring a missing region with high-quality texture through searching nearest neighbor patches from the unmasked regions. However, these methods bring problematic contents when recovering large missing regions. Deep networks, on the other hand, show promising results in completing large regions. Nonetheless, the results often lack faithful and sharp details that resemble the surrounding area. By bringing together the best of both paradigms, we propose a new deep inpainting framework where texture generation is guided by a texture memory of patch samples extracted from unmasked regions. The framework has a novel design that allows texture memory retrieval to be trained end-to-end with the deep inpainting network. In addition, we introduce a patch distribution loss to encourage high-quality patch synthesis. The proposed method shows superior performance both qualitatively and quantitatively on three challenging image benchmarks, i.e., Places, CelebA-HQ, and Paris Street-View datasets.

cs.CV↗

arXiv2021

Well-posedness in Sobolev spaces of the two-dimensional MHD Boundary layer equations without viscosity

We consider the two-dimensional MHD Boundary layer system without hydrodynamic viscosity, and establish the existence and uniqueness of solutions in Sobolev spaces under the assumption that the tangential component of magnetic fields dominates. This gives a complement to the previous works of Liu-Xie-Yang [Comm. Pure Appl. Math. 72 (2019)] and Liu-Wang-Xie-Yang [J. Funct. Anal. 279 (2020)], where the well-posedness theory was established for the MHD boundary layer systems with both viscosity and resistivity and with viscosity only, respectively. We use the pseudo-differential calculation, to overcome a new difficulty arising from the treatment of boundary integrals due to the absence of the diffusion property for the velocity.

math.AP↗

arXiv2021

Epitaxial fabrication of AgTe monolayer on Ag(111) and the sequential growth of Te film

Transition-metal chalcogenides (TMCs) materials have attracted increasing interest both for fundamental research and industrial applications. Among all these materials, two-dimensional (2D) compounds with honeycomb-like structure possess exotic electronic structures. Here, we report a systematic study of TMC monolayer AgTe fabricated by direct depositing Te on the surface of Ag(111) and annealing. Few intrinsic defects are observed and studied by scanning tunneling microscopy, indicating that there are two kinds of AgTe domains and they can form gliding twin-boundary. Then, the monolayer AgTe can serve as the template for the following growth of Te film. Meanwhile, some Te atoms are observed in the form of chains on the top of the bottom Te film. Our findings in this work might provide insightful guide for the epitaxial growth of 2D materials for study of novel physical properties and for future quantum devices.

cond-mat.mtrl-sci↗

arXiv2021

Toplayer-Dependent Crystallographic Orientation Imaging in the Bilayer Two-Dimensional Materials with Transverse Shear Microscopy

Nanocontact properties of two-dimensional (2D) materials are closely dependent on their unique nanomechanical systems, such as the number of atomic layers and the supporting substrate. Here, we report a direct observation of toplayer-dependent crystallographic orientation imaging of 2D materials with the transverse shear microscopy (TSM). Three typical nanomechanical systems, MoS2 on the amorphous SiO2/Si, graphene on the amorphous SiO2/Si, and MoS2 on the crystallized Al2O3, have been investigated in detail. This experimental observation reveals that puckering behaviour mainly occurs on the top layer of 2D materials, which is attributed to its direct contact adhesion with the AFM tip. Furthermore, the result of crystallographic orientation imaging of MoS2/SiO2/Si and MoS2/Al2O3 indicated that the underlying crystalline substrates almost do not contribute to the puckering effect of 2D materials. Our work directly revealed the top layer dependent puckering properties of 2D material, and demonstrate the general applications of TSM in the bilayer 2D systems.

cond-mat.mtrl-sci↗

arXiv2021

Learning Scene Structure Guidance via Cross-Task Knowledge Transfer for Single Depth Super-Resolution

Existing color-guided depth super-resolution (DSR) approaches require paired RGB-D data as training samples where the RGB image is used as structural guidance to recover the degraded depth map due to their geometrical similarity. However, the paired data may be limited or expensive to be collected in actual testing environment. Therefore, we explore for the first time to learn the cross-modality knowledge at training stage, where both RGB and depth modalities are available, but test on the target dataset, where only single depth modality exists. Our key idea is to distill the knowledge of scene structural guidance from RGB modality to the single DSR task without changing its network architecture. Specifically, we construct an auxiliary depth estimation (DE) task that takes an RGB image as input to estimate a depth map, and train both DSR task and DE task collaboratively to boost the performance of DSR. Upon this, a cross-task interaction module is proposed to realize bilateral cross task knowledge transfer. First, we design a cross-task distillation scheme that encourages DSR and DE networks to learn from each other in a teacher-student role-exchanging fashion. Then, we advance a structure prediction (SP) task that provides extra structure regularization to help both DSR and DE networks learn more informative structure representations for depth recovery. Extensive experiments demonstrate that our scheme achieves superior performance in comparison with other DSR methods.

cs.CV↗

arXiv2021

Precession of spheroids under Lorentz violation and observational consequences for neutron stars

The Standard-Model Extension (SME) is an effective-field-theoretic framework that catalogs all Lorentz-violating field operators. The anisotropic correction from the minimal gravitational SME to Newtonian gravitational energy for spheroids is studied, and the rotation of rigid spheroids is solved with perturbation method and numerical approach. The well-known forced precession solution given by Nordtvedt in the parameterized post-Newtonian formalism is recovered and applied to two observed solitary millisecond pulsars to set bounds on the coefficients for Lorentz violation in the SME framework. A different solution, which describes the rotation of an otherwise free-precessing star in the presence of Lorentz violation, is found, and its consequences on pulsar signals and continuous gravitational waves (GWs) emitted by neutron stars (NSs) are investigated. The study provides new possible tests of Lorentz violation once free-precessing NSs are firmly identified in the future.

gr-qc↗

arXiv2021

Signature of Lorentz Violation in Continuous Gravitational-Wave Spectra of Ellipsoidal Neutron Stars

We study effects of Lorentz-invariance violation on the rotation of neutron stars (NSs) in the minimal gravitational Standard-Model Extension framework, and calculate the quadrupole radiation generated by them. Aiming at testing Lorentz invariance with observations of continuous gravitational waves (GWs) from rotating NSs in the future, we compare the GW spectra of a rotating ellipsoidal NS under Lorentz-violating gravity with those of a Lorentz-invariant one. The former are found to possess frequency components higher than the second harmonic, which does not happen for the latter, indicating those higher frequency components to be potential signatures of Lorentz violation in continuous GW spectra of rotating NSs.

gr-qc↗

arXiv2021

BaPipe: Exploration of Balanced Pipeline Parallelism for DNN Training

The size of deep neural networks (DNNs) grows rapidly as the complexity of the machine learning algorithm increases. To satisfy the requirement of computation and memory of DNN training, distributed deep learning based on model parallelism has been widely recognized. We propose a new pipeline parallelism training framework, BaPipe, which can automatically explore pipeline parallelism training methods and balanced partition strategies for DNN distributed training. In BaPipe, each accelerator calculates the forward propagation and backward propagation of different parts of networks to implement the intra-batch pipeline parallelism strategy. BaPipe uses a new load balancing automatic exploration strategy that considers the parameters of DNN models and the computation, memory, and communication resources of accelerator clusters. We have trained different DNNs such as VGG-16, ResNet-50, and GNMT on GPU clusters and simulated the performance of different FPGA clusters. Compared with state-of-the-art data parallelism and pipeline parallelism frameworks, BaPipe provides up to 3.2x speedup and 4x memory reduction in various platforms.

cs.DC↗

arXiv2020

Positional Encoding as Spatial Inductive Bias in GANs

SinGAN shows impressive capability in learning internal patch distribution despite its limited effective receptive field. We are interested in knowing how such a translation-invariant convolutional generator could capture the global structure with just a spatially i.i.d. input. In this work, taking SinGAN and StyleGAN2 as examples, we show that such capability, to a large extent, is brought by the implicit positional encoding when using zero padding in the generators. Such positional encoding is indispensable for generating images with high fidelity. The same phenomenon is observed in other generative architectures such as DCGAN and PGGAN. We further show that zero padding leads to an unbalanced spatial bias with a vague relation between locations. To offer a better spatial inductive bias, we investigate alternative positional encodings and analyze their effects. Based on a more flexible positional encoding explicitly, we propose a new multi-scale training strategy and demonstrate its effectiveness in the state-of-the-art unconditional generator StyleGAN2. Besides, the explicit spatial inductive bias substantially improve SinGAN for more versatile image manipulation.

cs.CV↗

arXiv2020

CARAFE++: Unified Content-Aware ReAssembly of FEatures

Feature reassembly, i.e. feature downsampling and upsampling, is a key operation in a number of modern convolutional network architectures, e.g., residual networks and feature pyramids. Its design is critical for dense prediction tasks such as object detection and semantic/instance segmentation. In this work, we propose unified Content-Aware ReAssembly of FEatures (CARAFE++), a universal, lightweight and highly effective operator to fulfill this goal. CARAFE++ has several appealing properties: (1) Unlike conventional methods such as pooling and interpolation that only exploit sub-pixel neighborhood, CARAFE++ aggregates contextual information within a large receptive field. (2) Instead of using a fixed kernel for all samples (e.g. convolution and deconvolution), CARAFE++ generates adaptive kernels on-the-fly to enable instance-specific content-aware handling. (3) CARAFE++ introduces little computational overhead and can be readily integrated into modern network architectures. We conduct comprehensive evaluations on standard benchmarks in object detection, instance/semantic segmentation and image inpainting. CARAFE++ shows consistent and substantial gains across all the tasks (2.5% APbox, 2.1% APmask, 1.94% mIoU, 1.35 dB respectively) with negligible computational overhead. It shows great potential to serve as a strong building block for modern deep networks.

cs.CV↗

arXiv2020

Discovering Symbolic Models from Deep Learning with Inductive Biases

We develop a general approach to distill symbolic representations of a learned deep model by introducing strong inductive biases. We focus on Graph Neural Networks (GNNs). The technique works as follows: we first encourage sparse latent representations when we train a GNN in a supervised setting, then we apply symbolic regression to components of the learned model to extract explicit physical relations. We find the correct known equations, including force laws and Hamiltonians, can be extracted from the neural network. We then apply our method to a non-trivial cosmology example-a detailed dark matter simulation-and discover a new analytic formula which can predict the concentration of dark matter from the mass distribution of nearby cosmic structures. The symbolic expressions extracted from the GNN using our technique also generalized to out-of-distribution data better than the GNN itself. Our approach offers alternative directions for interpreting neural networks and discovering novel physical principles from the representations they learn.

cs.LG↗

arXiv2020

VolumeNet: A Lightweight Parallel Network for Super-Resolution of Medical Volumetric Data

Deep learning-based super-resolution (SR) techniques have generally achieved excellent performance in the computer vision field. Recently, it has been proven that three-dimensional (3D) SR for medical volumetric data delivers better visual results than conventional two-dimensional (2D) processing. However, deepening and widening 3D networks increases training difficulty significantly due to the large number of parameters and small number of training samples. Thus, we propose a 3D convolutional neural network (CNN) for SR of medical volumetric data called ParallelNet using parallel connections. We construct a parallel connection structure based on the group convolution and feature aggregation to build a 3D CNN that is as wide as possible with few parameters. As a result, the model thoroughly learns more feature maps with larger receptive fields. In addition, to further improve accuracy, we present an efficient version of ParallelNet (called VolumeNet), which reduces the number of parameters and deepens ParallelNet using a proposed lightweight building block module called the Queue module. Unlike most lightweight CNNs based on depthwise convolutions, the Queue module is primarily constructed using separable 2D cross-channel convolutions. As a result, the number of network parameters and computational complexity can be reduced significantly while maintaining accuracy due to full channel fusion. Experimental results demonstrate that the proposed VolumeNet significantly reduces the number of model parameters and achieves high precision results compared to state-of-the-art methods.

eess.IV↗

arXiv2020

SAHDL: Sparse Attention Hypergraph Regularized Dictionary Learning

In recent years, the attention mechanism contributes significantly to hypergraph based neural networks. However, these methods update the attention weights with the network propagating. That is to say, this type of attention mechanism is only suitable for deep learning-based methods while not applicable to the traditional machine learning approaches. In this paper, we propose a hypergraph based sparse attention mechanism to tackle this issue and embed it into dictionary learning. More specifically, we first construct a sparse attention hypergraph, asset attention weights to samples by employing the $\ell_1$-norm sparse regularization to mine the high-order relationship among sample features. Then, we introduce the hypergraph Laplacian operator to preserve the local structure for subspace transformation in dictionary learning. Besides, we incorporate the discriminative information into the hypergraph as the guidance to aggregate samples. Unlike previous works, our method updates attention weights independently, does not rely on the deep network. We demonstrate the efficacy of our approach on four benchmark datasets.

cs.CV↗

arXiv2020

DLDL: Dynamic Label Dictionary Learning via Hypergraph Regularization

For classification tasks, dictionary learning based methods have attracted lots of attention in recent years. One popular way to achieve this purpose is to introduce label information to generate a discriminative dictionary to represent samples. However, compared with traditional dictionary learning, this category of methods only achieves significant improvements in supervised learning, and has little positive influence on semi-supervised or unsupervised learning. To tackle this issue, we propose a Dynamic Label Dictionary Learning (DLDL) algorithm to generate the soft label matrix for unlabeled data. Specifically, we employ hypergraph manifold regularization to keep the relations among original data, transformed data, and soft labels consistent. We demonstrate the efficiency of the proposed DLDL approach on two remote sensing datasets.

cs.CV↗