arXiv subjects

Xiang Li

Publications and source records attributed to Xiang Li.

At least 1,315 records · Page 73Linked to original sources

arXiv2019

Spatial Group-wise Enhance: Improving Semantic Feature Learning in Convolutional Networks

The Convolutional Neural Networks (CNNs) generate the feature representation of complex objects by collecting hierarchical and different parts of semantic sub-features. These sub-features can usually be distributed in grouped form in the feature vector of each layer, representing various semantic entities. However, the activation of these sub-features is often spatially affected by similar patterns and noisy backgrounds, resulting in erroneous localization and identification. We propose a Spatial Group-wise Enhance (SGE) module that can adjust the importance of each sub-feature by generating an attention factor for each spatial location in each semantic group, so that every individual group can autonomously enhance its learnt expression and suppress possible noise. The attention factors are only guided by the similarities between the global and local feature descriptors inside each group, thus the design of SGE module is extremely lightweight with \emph{almost no extra parameters and calculations}. Despite being trained with only category supervisions, the SGE component is extremely effective in highlighting multiple active areas with various high-order semantics (such as the dog's eyes, nose, etc.). When integrated with popular CNN backbones, SGE can significantly boost the performance of image recognition tasks. Specifically, based on ResNet50 backbones, SGE achieves 1.2\% Top-1 accuracy improvement on the ImageNet benchmark and 1.0$\sim$2.0\% AP gain on the COCO benchmark across a wide range of detectors (Faster/Mask/Cascade RCNN and RetinaNet). Codes and pretrained models are available at https://github.com/implus/PytorchInsight.

cs.CV↗

arXiv2019

Group-Attention Single-Shot Detector (GA-SSD): Finding Pulmonary Nodules in Large-Scale CT Images

Early diagnosis of pulmonary nodules (PNs) can improve the survival rate of patients and yet is a challenging task for radiologists due to the image noise and artifacts in computed tomography (CT) images. In this paper, we propose a novel and effective abnormality detector implementing the attention mechanism and group convolution on 3D single-shot detector (SSD) called group-attention SSD (GA-SSD). We find that group convolution is effective in extracting rich context information between continuous slices, and attention network can learn the target features automatically. We collected a large-scale dataset that contained 4146 CT scans with annotations of varying types and sizes of PNs (even PNs smaller than 3mm were annotated). To the best of our knowledge, this dataset is the largest cohort with relatively complete annotations for PNs detection. Our experimental results show that the proposed group-attention SSD outperforms the classic SSD framework as well as the state-of-the-art 3DCNN, especially on some challenging lesion types.

cs.CV↗

arXiv2019

Evolving nature of human contact networks with its impact on epidemic processes

Human contact networks are constituted by a multitude of individuals and pairwise contacts among them. However, the dynamic nature, which generates the evolution of human contact networks, of contact patterns is not known yet. Here, we analyse three empirical datasets and identify two crucial mechanisms of the evolution of temporal human contact networks, i.e. the activity state transition laws for an individual to be socially active, and the contact establishment mechanism that active individuals adopt. We consider both of the two mechanisms to propose a temporal network model, named the memory driven (MD) model, of human contact networks. Then we study the susceptible-infected (SI) spreading processes on empirical human contact networks and four corresponding temporal network models, and compare the full prevalence time of SI processes with various infection rates on the networks. The full prevalence time of SI processes in the MD model is the same as that in real-world human contact networks. Moreover, we find that the individual activity state transition promotes the spreading process, while, the contact establishment of active individuals suppress the prevalence. Apart from this, we observe that even a small percentage of individuals to explore new social ties is able to induce an explosive spreading on networks. The proposed temporal network framework could help the further study of dynamic processes in temporal human contact networks, and offer new insights to predict and control the diffusion processes on networks.

physics.soc-ph↗

arXiv2019

Automated Segmentation of Cervical Nuclei in Pap Smear Images using Deformable Multi-path Ensemble Model

Pap smear testing has been widely used for detecting cervical cancers based on the morphology properties of cell nuclei in microscopic image. An accurate nuclei segmentation could thus improve the success rate of cervical cancer screening. In this work, a method of automated cervical nuclei segmentation using Deformable Multipath Ensemble Model (D-MEM) is proposed. The approach adopts a U-shaped convolutional network as a backbone network, in which dense blocks are used to transfer feature information more effectively. To increase the flexibility of the model, we then use deformable convolution to deal with different nuclei irregular shapes and sizes. To reduce the predictive bias, we further construct multiple networks with different settings, which form an ensemble model. The proposed segmentation framework has achieved state-of-the-art accuracy on Herlev dataset with Zijdenbos similarity index (ZSI) of 0.933, and has the potential to be extended for solving other medical image segmentation tasks.

cs.CV↗

arXiv2019

Do Subsampled Newton Methods Work for High-Dimensional Data?

Subsampled Newton methods approximate Hessian matrices through subsampling techniques, alleviating the cost of forming Hessian matrices but using sufficient curvature information. However, previous results require $Ω(d)$ samples to approximate Hessians, where $d$ is the dimension of data points, making it less practically feasible for high-dimensional data. The situation is deteriorated when $d$ is comparably as large as the number of data points $n$, which requires to take the whole dataset into account, making subsampling useless. This paper theoretically justifies the effectiveness of subsampled Newton methods on high dimensional data. Specifically, we prove only $\widetildeΘ(d^γ_{\rm eff})$ samples are needed in the approximation of Hessian matrices, where $d^γ_{\rm eff}$ is the $γ$-ridge leverage and can be much smaller than $d$ as long as $nγ\gg 1$. Additionally, we extend this result so that subsampled Newton methods can work for high-dimensional data on both distributed optimization problems and non-smooth regularized problems.

stat.ML↗

arXiv2019

DmpIRFs and DmpST: DAMPE Instrument Response Functions and Science Tools for Gamma-Ray Data Analysis

GeV gamma ray is an important observation target of DArk Matter Particle Explorer (DAMPE) for indirect dark matter searching and high energy astrophysics. We present in this work a set of accurate instrument response functions of DAMPE (DmpIRFs) including the effective area, point-spread function and energy dispersion that are crucial for the gamma-ray data analysis based on the high statistics simulation data. A dedicated software named DmpST is developed to facilitate the scientific analyses of DAMPE gamma-ray data. Considering the limited number of photons and the angular resolution of DAMPE, the maximum likelihood method is adopted in the DmpST to better disentangle different source components. The basic mathematics and the framework regarding this software are also introduced in this paper.

astro-ph.HE↗

arXiv2019

Programmable Metasurface-based RF Chain-free 8PSK Wireless Transmitter

In this letter, a wireless transmitter using the new architecture of programmable metasurface is presented. The proposed transmitter does not require any filter, nor wideband mixer or wideband power amplifier, thereby making it a promising hardware architecture for cost-effective wireless communications systems in the future. Using experimental results, we demonstrate that a programmable metasurface-based 8-phase shift-keying (8PSK) transmitter with 8*32 phase adjustable unit cells can achieve 6.144 Mbps data rate over the air at 4.25 GHz with a comparable bit error rate (BER) performance as the conventional approach without channel coding, but with less hardware complexity.

eess.SP↗

arXiv2019

ConvLab: Multi-Domain End-to-End Dialog System Platform

We present ConvLab, an open-source multi-domain end-to-end dialog system platform, that enables researchers to quickly set up experiments with reusable components and compare a large set of different approaches, ranging from conventional pipeline systems to end-to-end neural models, in common environments. ConvLab offers a set of fully annotated datasets and associated pre-trained reference models. As a showcase, we extend the MultiWOZ dataset with user dialog act annotations to train all component models and demonstrate how ConvLab makes it easy and effortless to conduct complicated experiments in multi-domain end-to-end dialog settings.

cs.CL↗

arXiv2019

On Model Adaptation for Sensorimotor Control of Robots

In this article, we address the problem of computing adaptive sensorimotor models that can be used for guiding the motion of robotic systems with uncertain action-to-perception relations. The formulation of the uncalibrated sensor-based control problem is first presented, then, various computational methods for building adaptive sensorimotor models are derived and analysed. The proposed methodology is exemplified with two cases of study: (i) shape control of deformable objects with unknown properties, and (ii) soft manipulation of ultrasonic probes with uncalibrated sensors.

cs.RO↗

arXiv2019

Thermodynamics of 2+1 Flavor Polyakov-Loop Quark-Meson Model under External Magnetic Field

We study the thermodynamics of QCD system under external magnetic field via the 2+1 flavor Polyakov-loop quark-meson model. To incorporate quantum and thermal fluctuations, the functional renormalization group approach is implemented in our work. Pressure, entropy density, magnetic susceptibility and other thermodynamic quantities are calculated and analyzed to investigate the effect of magnetic field on the QCD system. The calculated results are in reasonable agreement with lattice QCD simulations and perturbation theory. We then give an intuitive picture for the response of QCD system to the magnetic field.

hep-ph↗

arXiv2019

The Impact of Information Dissemination on Vaccination in Multiplex Networks

The impact of information dissemination on epidemic control is essentially subject to individual behaviors. Unlike information-driven behaviors, vaccination is determined by many cost-related factors, whose correlation with the information dissemination should be better understood. To this end, we propose an evolutionary vaccination game model in multiplex networks by integrating an information-epidemic spreading process into the vaccination dynamics, and explore how information dissemination influences vaccination. The spreading process is described by a two-layer coupled susceptible-alert-infected-susceptible (SAIS) model, where the strength coefficient between two layers is defined to characterize the tendency and intensity of information dissemination. We find that information dissemination can increase the epidemic threshold, however, more information transmission cannot promote vaccination. Specifically, increasing information dissemination even leads to a decline of the vaccination equilibrium and raises the final infection density. Moreover, we study the impact of strength coefficient and individual sensitivity on social cost, and unveil the role of information dissemination in controlling the epidemic with numerical simulations.

physics.soc-ph↗

arXiv2019

Non-Rigid Point Set Registration Networks

Point set registration is defined as a process to determine the spatial transformation from the source point set to the target one. Existing methods often iteratively search for the optimal geometric transformation to register a given pair of point sets, driven by minimizing a predefined alignment loss function. In contrast, the proposed point registration neural network (PR-Net) actively learns the registration pattern as a parametric function from a training dataset, consequently predict the desired geometric transformation to align a pair of point sets. PR-Net can transfer the learned knowledge (i.e. registration pattern) from registering training pairs to testing ones without additional iterative optimization. Specifically, in this paper, we develop novel techniques to learn shape descriptors from point sets that help formulate a clear correlation between source and target point sets. With the defined correlation, PR-Net tends to predict the transformation so that the source and target point sets can be statistically aligned, which in turn leads to an optimal spatial geometric registration. PR-Net achieves robust and superior performance for non-rigid registration of point sets, even in presence of Gaussian noise, outliers, and missing points, but requires much less time for registering large number of pairs. More importantly, for a new pair of point sets, PR-Net is able to directly predict the desired transformation using the learned model without repetitive iterative optimization routine. Our code is available at https://github.com/Lingjing324/PR-Net.

cs.GR↗

arXiv2019

Selective Kernel Networks

In standard Convolutional Neural Networks (CNNs), the receptive fields of artificial neurons in each layer are designed to share the same size. It is well-known in the neuroscience community that the receptive field size of visual cortical neurons are modulated by the stimulus, which has been rarely considered in constructing CNNs. We propose a dynamic selection mechanism in CNNs that allows each neuron to adaptively adjust its receptive field size based on multiple scales of input information. A building block called Selective Kernel (SK) unit is designed, in which multiple branches with different kernel sizes are fused using softmax attention that is guided by the information in these branches. Different attentions on these branches yield different sizes of the effective receptive fields of neurons in the fusion layer. Multiple SK units are stacked to a deep network termed Selective Kernel Networks (SKNets). On the ImageNet and CIFAR benchmarks, we empirically show that SKNet outperforms the existing state-of-the-art architectures with lower model complexity. Detailed analyses show that the neurons in SKNet can capture target objects with different scales, which verifies the capability of neurons for adaptively adjusting their receptive field sizes according to the input. The code and models are available at https://github.com/implus/SKNet.

cs.CV↗

arXiv2019

Development of an Autonomous Sanding Robot with Structured-Light Technology

Large demand for robotics and automation has been reflected in the sanding works, as current manual operations are labor-intensive, without consistent quality, and also subject to safety and health issues. While several machines have been developed to automate one or two steps in the sanding works, the autonomous capability of existing solutions is relatively low, and the human assistance or supervision is still heavily required in the calibration of target objects or the planning of robot motion and tasks. This paper presents the development of an autonomous sanding robot, which is able to perform the sanding works on an unknown object automatically, without any prior calibration or human intervention. The developed robot works as follows. First, the target object is scanned then modeled with the structured-light camera. Second, the robot motion is planned to cover all the surfaces of the object with an optimized transition sequence. Third, the robot is controlled to perform the sanding on the object under the desired impedance model. A prototype of the sanding robot is fabricated and its performance is validated in the task of sanding a batch of wooden boxes. With sufficient degrees of freedom (DOFs) and the module design for the end effector, the developed robot is able to provide a general solution to the autonomous sanding on many other different objects.

eess.SY↗

arXiv2019

Dynamic Feature Fusion for Semantic Edge Detection

Features from multiple scales can greatly benefit the semantic edge detection task if they are well fused. However, the prevalent semantic edge detection methods apply a fixed weight fusion strategy where images with different semantics are forced to share the same weights, resulting in universal fusion weights for all images and locations regardless of their different semantics or local context. In this work, we propose a novel dynamic feature fusion strategy that assigns different fusion weights for different input images and locations adaptively. This is achieved by a proposed weight learner to infer proper fusion weights over multi-level features for each location of the feature map, conditioned on the specific input. In this way, the heterogeneity in contributions made by different locations of feature maps and input images can be better considered and thus help produce more accurate and sharper edge predictions. We show that our model with the novel dynamic feature fusion is superior to fixed weight fusion and also the naïve location-invariant weight fusion methods, via comprehensive experiments on benchmarks Cityscapes and SBD. In particular, our method outperforms all existing well established methods and achieves new state-of-the-art.

cs.CV↗

arXiv2019

TipTop: (Almost) Exact Solutions for Influence Maximization in Billion-scale Networks

In this paper, we study the Cost-aware Target Viral Marketing (CTVM) problem, a generalization of Influence Maximization (IM). CTVM asks for the most cost-effective users to influence the most relevant users. In contrast to the vast literature, we attempt to offer exact solutions. As the problem is NP-hard, thus, exact solutions are intractable, we propose TipTop, a $(1-ε)$-optimal solution for arbitrary $ε>0$ that scales to very large networks such as Twitter. At the heart of TipTop lies an innovative technique that reduces the number of samples as much as possible. This allows us to exactly solve CTVM on a much smaller space of generated samples using Integer Programming. Furthermore, TipTop lends a tool for researchers to benchmark their solutions against the optimal one in large-scale networks, which is currently not available.

cs.SI↗

arXiv2019

An Approximation Algorithm for Active Friending in Online Social Networks

Guiding users to actively expanding their online social circles is one of the primary strategies for enhancing user participation and growing online social networks. In this paper, we study the active friending problem which aims at providing users with the strategy for methodically sending invitations to successfully build a friendship with target users. We consider the prominent linear threshold model for the friending process and formulate the active friending problem as an optimization problem. The key observation is the relationship between the active friending problem and the minimum subset cover problem, based on which we present the first randomized algorithm with a data-independent approximation ratio and a controllable success probability for general graphs. The performance of the proposed algorithm is theoretically analyzed and supported by encouraging simulation results done on extensive datasets.

cs.SI↗

arXiv2019

Pelee: A Real-Time Object Detection System on Mobile Devices

An increasing need of running Convolutional Neural Network (CNN) models on mobile devices with limited computing power and memory resource encourages studies on efficient model design. A number of efficient architectures have been proposed in recent years, for example, MobileNet, ShuffleNet, and MobileNetV2. However, all these models are heavily dependent on depthwise separable convolution which lacks efficient implementation in most deep learning frameworks. In this study, we propose an efficient architecture named PeleeNet, which is built with conventional convolution instead. On ImageNet ILSVRC 2012 dataset, our proposed PeleeNet achieves a higher accuracy and over 1.8 times faster speed than MobileNet and MobileNetV2 on NVIDIA TX2. Meanwhile, PeleeNet is only 66% of the model size of MobileNet. We then propose a real-time object detection system by combining PeleeNet with Single Shot MultiBox Detector (SSD) method and optimizing the architecture for fast speed. Our proposed detection system2, named Pelee, achieves 76.4% mAP (mean average precision) on PASCAL VOC2007 and 22.4 mAP on MS COCO dataset at the speed of 23.6 FPS on iPhone 8 and 125 FPS on NVIDIA TX2. The result on COCO outperforms YOLOv2 in consideration of a higher precision, 13.6 times lower computational cost and 11.3 times smaller model size.

cs.CV↗