arXiv subjects

Ke Xu

Publications and source records attributed to Ke Xu.

At least 379 records · Page 21Linked to original sources

arXiv2020

Behavior variations and their implications for popularity promotions: From elites to mass in Weibo

The boom in social media with regard to producing and consuming information simultaneously implies the crucial role of online user influence in determining content popularity. In particular, understanding behavior variations between the influential elites and the mass grassroots is an important issue in communication. However, how their behavior varies across user categories and content domains, and how these differences influence content popularity are rarely addressed. From a novel view of seven content-domains, a detailed picture of behavior variations among five user groups, from both views of elites and mass, is drawn in Weibo, one of the most popular Twitter-like services in China. Interestingly, elites post more diverse contents with video links while the mass possess retweeters of higher loyalty. According to these variations, user-oriented actions of enhancing content popularity are discussed and testified. The most surprising finding is that the diversity of contents do not always bring more retweets, and the mass and elites should promote content popularity by increasing their retweeter counts and loyalty, respectively. Our results for the first time demonstrate the possibility of highly individualized strategies of popularity promotions in social media, instead of a universal principle.

cs.SI↗

arXiv2020

Electronics-Free Pneumatic Logic Circuits for Localized Feedback Control of Multi-Actuator Soft Robots

The vision of creating entirely-soft robots capable of performing complex tasks will be accomplished only when the controllers required for autonomous operation can be fully implemented on soft components. Despite recent advances in compliant fluidic circuitry for mechanical signal processing, the applicability of this technology for soft robot control has been limited by complicated fabrication and tuning processes, and also the need for external signals such as clocks and digital references. We propose a method to develop pneumatic soft robots in which coordinated interactions between multiple actuators are performed using controllers implemented on components distributedly embedded in the soft structures of the system. In this approach, the notions of binary and multi-valued actuator logic states are introduced. In this way, the physical local dynamical couplings between the analog states of the actuators, established using soft valves of a new type, can be thought of as logic-gate-based mappings acting on discretized representations of the actuator states. Consequently, techniques for digital logic design can be applied to derive the architectures of the localized mechanical couplings that intelligently coordinate the oscillation patterns of the actuator responses. For the purposes of controller tuning, the soft valves are conceived so that their main physical parameters can be adjusted from the exterior of the robot through simple geometrical changes of the corresponding structural elements. To demonstrate the proposed approach, we present the development of a six-state locomoting soft robot.

cs.RO↗

arXiv2020

Self-Adversarial Learning with Comparative Discrimination for Text Generation

Conventional Generative Adversarial Networks (GANs) for text generation tend to have issues of reward sparsity and mode collapse that affect the quality and diversity of generated samples. To address the issues, we propose a novel self-adversarial learning (SAL) paradigm for improving GANs' performance in text generation. In contrast to standard GANs that use a binary classifier as its discriminator to predict whether a sample is real or generated, SAL employs a comparative discriminator which is a pairwise classifier for comparing the text quality between a pair of samples. During training, SAL rewards the generator when its currently generated sentence is found to be better than its previously generated samples. This self-improvement reward mechanism allows the model to receive credits more easily and avoid collapsing towards the limited number of real samples, which not only helps alleviate the reward sparsity issue but also reduces the risk of mode collapse. Experiments on text generation benchmark datasets show that our proposed approach substantially improves both the quality and the diversity, and yields more stable performance compared to the previous GANs for text generation.

cs.CL↗

arXiv2020

Learning to Compare for Better Training and Evaluation of Open Domain Natural Language Generation Models

Automated evaluation of open domain natural language generation (NLG) models remains a challenge and widely used metrics such as BLEU and Perplexity can be misleading in some cases. In our paper, we propose to evaluate natural language generation models by learning to compare a pair of generated sentences by fine-tuning BERT, which has been shown to have good natural language understanding ability. We also propose to evaluate the model-level quality of NLG models with sample-level comparison results with skill rating system. While able to be trained in a fully self-supervised fashion, our model can be further fine-tuned with a little amount of human preference annotation to better imitate human judgment. In addition to evaluating trained models, we propose to apply our model as a performance indicator during training for better hyperparameter tuning and early-stopping. We evaluate our approach on both story generation and chit-chat dialogue response generation. Experimental results show that our model correlates better with human preference compared with previous automated evaluation approaches. Training with the proposed metric yields better performance in human evaluation, which further demonstrates the effectiveness of the proposed model.

cs.CL↗

arXiv2019

Compactness of isospectral conformal metrics on 4-manifolds

Let a sequence of conformal Riemannian metrics $\{g_k=u_k^2g_0\}$ be isospectral to $g_0$ over a compact boundaryless smooth 4-dimension manifold $(M,g_0)$. We prove that the subsequence of conformal factors $\{u_k\}$ converges to $u$ weakly in $W^{2,p}_{loc}(M\setminus \mathcal{S})$ for some $p<2$, where $\mathcal{S}$ is a finite set of points and $u\in W^{2,p}(M,g_0)$. Moreover, if the isospectral invariant $\frac{\int_M R(g_k)dV_{g_k}}{6\sqrt{\mathrm{Vol}(M,g_k)}}$ is strictly smaller than the Yamabe constant of the standard sphere $\mathbb{S}^4$, then the subsequence of distance functions $\{d_k\}$ defined by $\{g_k\}$ uniformly converges to $d_u$ and the subsequence of metric spaces $\{(M,d_k)\}$ converges to the metric space $(M,d_u)$ in the Gromov-Hausdorff topology, where $d_u$ is the distance function defined by $u^2g_0$.

math.DG↗

arXiv2019

DeepMask: an algorithm for cloud and cloud shadow detection in optical satellite remote sensing images using deep residual network

Detecting and masking cloud and cloud shadow from satellite remote sensing images is a pervasive problem in the remote sensing community. Accurate and efficient detection of cloud and cloud shadow is an essential step to harness the value of remotely sensed data for almost all downstream analysis. DeepMask, a new algorithm for cloud and cloud shadow detection in optical satellite remote sensing imagery, is proposed in this study. DeepMask utilizes ResNet, a deep convolutional neural network, for pixel-level cloud mask generation. The algorithm is trained and evaluated on the Landsat 8 Cloud Cover Assessment Validation Dataset distributed across 8 different land types. Compared with CFMask, the most widely used cloud detection algorithm, land-type-specific DeepMask models achieve higher accuracy across all land types. The average accuracy is 93.56%, compared with 85.36% from CFMask. DeepMask also achieves 91.02% accuracy on all-land-type dataset. Compared with other CNN-based cloud mask algorithms, DeepMask benefits from the parsimonious architecture and the residual connection of ResNet. It is compatible with input of any size and shape. DeepMask still maintains high performance when using only red, green, blue, and NIR bands, indicating its potential to be applied to other satellite platforms that only have limited optical bands.

cs.CV↗

arXiv2019

Multi-Level Order-Flow Imbalance in a Limit Order Book

We study the multi-level order-flow imbalance (MLOFI), which is a vector quantity that measures the net flow of buy and sell orders at different price levels in a limit order book (LOB). Using a recent, high-quality data set for 6 liquid stocks on Nasdaq, we fit a simple, linear relationship between MLOFI and the contemporaneous change in mid-price. For all 6 stocks that we study, we find that the out-of-sample goodness-of-fit of the relationship improves with each additional price level that we include in the MLOFI vector. Our results underline how order-flow activity deep into the LOB can influence the price-formation process.

q-fin.TR↗

arXiv2019

Where Is My Mirror?

Mirrors are everywhere in our daily lives. Existing computer vision systems do not consider mirrors, and hence may get confused by the reflected content inside a mirror, resulting in a severe performance degradation. However, separating the real content outside a mirror from the reflected content inside it is non-trivial. The key challenge is that mirrors typically reflect contents similar to their surroundings, making it very difficult to differentiate the two. In this paper, we present a novel method to segment mirrors from an input image. To the best of our knowledge, this is the first work to address the mirror segmentation problem with a computational approach. We make the following contributions. First, we construct a large-scale mirror dataset that contains mirror images with corresponding manually annotated masks. This dataset covers a variety of daily life scenes, and will be made publicly available for future research. Second, we propose a novel network, called MirrorNet, for mirror segmentation, by modeling both semantical and low-level color/texture discontinuities between the contents inside and outside of the mirrors. Third, we conduct extensive experiments to evaluate the proposed method, and show that it outperforms the carefully chosen baselines from the state-of-the-art detection and segmentation methods.

cs.CV↗

arXiv2019

Remote blind state preparation with weak coherent pulses in the field

Quantum computing has seen tremendous progress in the past years. Due to the implementation complexity and cost, the future path of quantum computation is strongly believed to delegate computational tasks to powerful quantum servers on cloud. Universal blind quantum computing (UBQC) provides the protocol for the secure delegation of arbitrary quantum computations, and it has received significant attention. However, a great challenge in UBQC is how to transmit quantum state over long distance securely and reliably. Here, we solve this challenge by proposing and demonstrating a resource-efficient remote blind qubit preparation (RBQP) protocol with weak coherent pulses for the client to produce, using a compact and low-cost laser. We demonstrate the protocol in field, experimentally verifying the protocol over 100-km fiber. Our experiment uses a quantum teleportation setup in telecom wavelength and generates $1000$ secure qubits with an average fidelity of $(86.9\pm1.5)\%$, which exceeds the quantum no-cloning fidelity of equatorial qubit states. The results prove the feasibility of UBQC over long distances, and thus serving as a key milestone towards secure cloud quantum computing.

quant-ph↗

arXiv2019

DRFN: Deep Recurrent Fusion Network for Single-Image Super-Resolution with Large Factors

Recently, single-image super-resolution has made great progress owing to the development of deep convolutional neural networks (CNNs). The vast majority of CNN-based models use a pre-defined upsampling operator, such as bicubic interpolation, to upscale input low-resolution images to the desired size and learn non-linear mapping between the interpolated image and ground truth high-resolution (HR) image. However, interpolation processing can lead to visual artifacts as details are over-smoothed, particularly when the super-resolution factor is high. In this paper, we propose a Deep Recurrent Fusion Network (DRFN), which utilizes transposed convolution instead of bicubic interpolation for upsampling and integrates different-level features extracted from recurrent residual blocks to reconstruct the final HR images. We adopt a deep recurrence learning strategy and thus have a larger receptive field, which is conducive to reconstructing an image more accurately. Furthermore, we show that the multi-level fusion structure is suitable for dealing with image super-resolution problems. Extensive benchmark evaluations demonstrate that the proposed DRFN performs better than most current deep learning methods in terms of accuracy and visual effects, especially for large-scale images, while using fewer parameters.

cs.CV↗

arXiv2019

Spatial Attentive Single-Image Deraining with a High Quality Real Rain Dataset

Removing rain streaks from a single image has been drawing considerable attention as rain streaks can severely degrade the image quality and affect the performance of existing outdoor vision tasks. While recent CNN-based derainers have reported promising performances, deraining remains an open problem for two reasons. First, existing synthesized rain datasets have only limited realism, in terms of modeling real rain characteristics such as rain shape, direction and intensity. Second, there are no public benchmarks for quantitative comparisons on real rain images, which makes the current evaluation less objective. The core challenge is that real world rain/clean image pairs cannot be captured at the same time. In this paper, we address the single image rain removal problem in two ways. First, we propose a semi-automatic method that incorporates temporal priors and human supervision to generate a high-quality clean image from each input sequence of real rain images. Using this method, we construct a large-scale dataset of $\sim$$29.5K$ rain/rain-free image pairs that covers a wide range of natural rain scenes. Second, to better cover the stochastic distribution of real rain streaks, we propose a novel SPatial Attentive Network (SPANet) to remove rain streaks in a local-to-global manner. Extensive experiments demonstrate that our network performs favorably against the state-of-the-art deraining methods.

cs.CV↗

arXiv2019

Visualizing and Understanding the Effectiveness of BERT

Language model pre-training, such as BERT, has achieved remarkable results in many NLP tasks. However, it is unclear why the pre-training-then-fine-tuning paradigm can improve performance and generalization capability across different tasks. In this paper, we propose to visualize loss landscapes and optimization trajectories of fine-tuning BERT on specific datasets. First, we find that pre-training reaches a good initial point across downstream tasks, which leads to wider optima and easier optimization compared with training from scratch. We also demonstrate that the fine-tuning procedure is robust to overfitting, even though BERT is highly over-parameterized for downstream tasks. Second, the visualization results indicate that fine-tuning BERT tends to generalize better because of the flat and wide optima, and the consistency between the training loss surface and the generalization error surface. Third, the lower layers of BERT are more invariant during fine-tuning, which suggests that the layers that are close to input learn more transferable representations of language.

cs.CL↗

arXiv2019

MIN: Co-Governing Multi-Identifier Network Architecture and its Prototype on Operator's Network

IP protocol is the core of TCP/IP network layer. However, since IP address and its Domain Name are allocated and managed by a single agency, there are risks of centralization. The semantic overload of IP address also reduces its scalability and mobility, which further hinders the security. This paper proposes a co-governing Multi-Identifier Network (MIN) architecture that constructs a network layer with parallel coexistence of multiple identifiers, including identity, content, geographic information, and IP address. On the management plane, we develop an efficient management system using consortium blockchain with voting consensus, so the network can simultaneously manage and support by hundreds or thousands of nodes with high throughput. On the data plane, we propose an algorithm merging hash table and prefix tree (HTP) for FIB, which avoids the false-negative error and can inter-translate different identifiers with tens of billions of entries. Further, we propose a scheme to transport IP packets using CCN as a tunnel for supporting progressive deployment. We deployed the prototype of MIN to the largest operators' network in Mainland China, Hongkong and Macao, and demonstrated that the network can register identifier under co-governing consensus algorithm, support VoD service very well.

cs.NI↗

arXiv2019

CloudDet: Interactive Visual Analysis of Anomalous Performances in Cloud Computing Systems

Detecting and analyzing potential anomalous performances in cloud computing systems is essential for avoiding losses to customers and ensuring the efficient operation of the systems. To this end, a variety of automated techniques have been developed to identify anomalies in cloud computing performance. These techniques are usually adopted to track the performance metrics of the system (e.g., CPU, memory, and disk I/O), represented by a multivariate time series. However, given the complex characteristics of cloud computing data, the effectiveness of these automated methods is affected. Thus, substantial human judgment on the automated analysis results is required for anomaly interpretation. In this paper, we present a unified visual analytics system named CloudDet to interactively detect, inspect, and diagnose anomalies in cloud computing systems. A novel unsupervised anomaly detection algorithm is developed to identify anomalies based on the specific temporal patterns of the given metrics data (e.g., the periodic pattern), the results of which are visualized in our system to indicate the occurrences of anomalies. Rich visualization and interaction designs are used to help understand the anomalies in the spatial and temporal context. We demonstrate the effectiveness of CloudDet through a quantitative evaluation, two case studies with real-world data, and interviews with domain experts.

cs.HC↗

arXiv2019

The Prototype of Decentralized Multilateral Co-Governing Post-IP Internet Architecture and Its Testing on Operator Networks

The Internet has become the most important infrastructure of modern society, while the existing IP network is unable to provide high-quality service. The unilateralism IP network is unable to satisfy the Co-managing and Co-governing demands to Cyberspace for most Nations in the world as well. Facing this challenge, we propose a novel Decentralized Multilateral Co-Governing Post-IP Internet architecture. To verify its effectiveness, we develop the prototype on the operator's networks including China Mainland, Hong Kong, and Macao. The experiments and testing results show that this architecture is feasible for co-existing of Content-Centric Networking and IP network, and it might become a Chinese Solution to the world.

cs.NI↗

arXiv2019

Online reviews can predict long-term returns of individual stocks

Online reviews are feedback voluntarily posted by consumers about their consumption experiences. This feedback indicates customer attitudes such as affection, awareness and faith towards a brand or a firm and demonstrates inherent connections with a company's future sales, cash flow and stock pricing. However, the predicting power of online reviews for long-term returns on stocks, especially at the individual level, has received little research attention, making a comprehensive exploration necessary to resolve existing debates. In this paper, which is based exclusively on online reviews, a methodology framework for predicting long-term returns of individual stocks with competent performance is established. Specifically, 6,246 features of 13 categories inferred from more than 18 million product reviews are selected to build the prediction models. With the best classifier selected from cross-validation tests, a satisfactory increase in accuracy, 13.94%, was achieved compared to the cutting-edge solution with 10 technical indicators being features, representing an 18.28% improvement relative to the random value. The robustness of our model is further evaluated and testified in realistic scenarios. It is thus confirmed for the first time that long-term returns of individual stocks can be predicted by online reviews. This study provides new opportunities for investors with respect to long-term investments in individual stocks.

econ.GN↗

arXiv2019

High thermoelectric efficiency in monolayer PbI$_2$ from 300 K to 900 K

By using a first-principles approach, monolayer PbI$_2$ is found to have great potential in thermoelectric applications. The linear Boltzmann transport equation is applied to obtain the perturbation to the electron distribution by different scattering mechanisms. The mobility is mainly limited by the deformation-potential interaction with long-wavelength acoustic vibrations at low carrier concentrations. At high concentrations, ionized impurity scattering becomes stronger. The electrical conductivity and Seebeck coefficient are calculated accurately over various ranges of temperature and carrier concentration. The lattice thermal conductivity of PbI$_2$, 0.065 W/mK at 300 K, is the lowest among other 2D thermoelectric materials. Such ultralow thermal conductivity is attributed to large atomic mass, weak interatomic bonding, strong anharmonicity, and localized vibrations in which the vast majority of heat is trapped. These electrical and phonon transport properties enable high thermoelectric figure of merit over 1 for both p-type and n-type doping from 300 K to 900 K. A maximum $zT$ of 4.9 is achieved at 900 K with an electron concentration of 1.9$\times$10$^{12}$ cm$^{-2}$. Our work shows exceptionally good thermoelectric energy conversion efficiency in monolayer PbI$_2$, which can be integrated to the existing photovoltaic devices.

cond-mat.mtrl-sci↗

arXiv2018

Arbitrarily routed mode-division multiplexed photonic circuits for dense integration

Mode-division multiplexing (MDM) is becoming an enabling technique for large-capacity data communications via encoding the information on orthogonal guiding modes. However, the on-chip routing of a multimode waveguide occupies too large chip area due to the constraints on inter-mode cross talk and mode leakage. Very recently, many efforts have been made to shrink the footprint of individual element like bending and crossing, but the devices still occupy >10x10 um2 footprint for three-mode multiplexed signals and the high-speed signal transmission has not been demonstrated yet. In this work, we demonstrate the first MDM circuits based on digitized meta-structures which have extremely compact footprints. The radius for a three-mode bending is only 3.9 μm and the footprint of a crossing is only 8x8um2. The 3x100 Gbit/s mode-multiplexed signals are arbitrarily routed through the circuits consists of many sharp bends and compact crossing with a bit error rate under forward error correction limit. This work is a significant step towards the large-scale and dense integration of MDM photonic integrated circuits.

physics.app-ph↗