arXiv subjects

Yu Li

Publications and source records attributed to Yu Li.

At least 559 records · Page 31Linked to original sources

arXiv2018

ω/T scaling and magnetic quantum criticality in BaFe2(As0.7P0.3)2

We used transport and inelastic neutron scattering to study the optimally phosphorus-doped BaFe$_2$(As$_{0.7}$P$_{0.3}$)$_2$ superconductor ($T_c = 30$ K). In the normal state, we find that the previously reported linear temperature dependence of the resistivity below room temperature extends to $\sim$ 500 K. Our analysis of the temperature and energy ($E=\hbarω$) dependence of spin dynamical susceptibility at the antiferromagnetic (AF) ordering wave vector $χ^{\prime\prime}({\bf Q}_{\rm AF},ω)$ reveal an $ω/ T$ scaling within $1.1<E/k_BT<110$. These results suggest that the linear temperature dependence of the resistivity is due to the presence of a magnetic quantum critical point in the cleanest iron pnictides near optimal superconductivity. Moreover, the results reconcile the strange-metal temperature dependences with the weakly first-order nature of the quantum transition out of the AF and nematic orders.

cond-mat.supr-con↗

arXiv2018

SupportNet: solving catastrophic forgetting in class incremental learning with support data

A plain well-trained deep learning model often does not have the ability to learn new knowledge without forgetting the previously learned knowledge, which is known as catastrophic forgetting. Here we propose a novel method, SupportNet, to efficiently and effectively solve the catastrophic forgetting problem in the class incremental learning scenario. SupportNet combines the strength of deep learning and support vector machine (SVM), where SVM is used to identify the support data from the old data, which are fed to the deep learning model together with the new data for further training so that the model can review the essential information of the old data when learning the new information. Two powerful consolidation regularizers are applied to stabilize the learned representation and ensure the robustness of the learned model. We validate our method with comprehensive experiments on various tasks, which show that SupportNet drastically outperforms the state-of-the-art incremental learning methods and even reaches similar performance as the deep learning model trained from scratch on both old and new data. Our program is accessible at: https://github.com/lykaust15/SupportNet

cs.NE↗

arXiv2018

On Configurable Defense against Adversarial Example Attacks

Machine learning systems based on deep neural networks (DNNs) have gained mainstream adoption in many applications. Recently, however, DNNs are shown to be vulnerable to adversarial example attacks with slight perturbations on the inputs. Existing defense mechanisms against such attacks try to improve the overall robustness of the system, but they do not differentiate different targeted attacks even though the corresponding impacts may vary significantly. To tackle this problem, we propose a novel configurable defense mechanism in this work, wherein we are able to flexibly tune the robustness of the system against different targeted attacks to satisfy application requirements. This is achieved by refining the DNN loss function with an attack sensitive matrix to represent the impacts of different targeted attacks. Experimental results on CIFAR-10 and GTSRB data sets demonstrate the efficacy of the proposed solution.

cs.LG↗

arXiv2018

Computation Load Balancing Real-Time Model Predictive Control in Urban Traffic Networks

Owing to the rapid growth number of vehicles, urban traffic congestion has become more and more severe in the last decades. As an effective approach, Model Predictive Control (MPC) has been applied to urban traffic signal control system. However, the potentially high online computation burden may limit its further application for real scenarios. In this paper, a new approach based on online active set strategy is proposed to improve the real-time performance of MPC-based traffic controller by reducing the online computing time. This approach divides one control cycle into several sequential sampling intervals. In each interval, online active set method is applied to solve quadratic programming (QP) of traffic signal control model, by searching the optimal solution starting at the optimal solution of previous interval in the feasible region. The most appealing property of this approach lies in that it can distribute the computational complexity into several sample intervals, instead of imposing heavy computation burden at each end of control cycle. The simulation experiments show that this breakthrough approach can obviously reduce the online computational complexity, and increase the applicability of the MPC in real-life traffic networks.

eess.SY↗

arXiv2018

Sample-Efficient Policy Learning based on Completely Behavior Cloning

Direct policy search is one of the most important algorithm of reinforcement learning. However, learning from scratch needs a large amount of experience data and can be easily prone to poor local optima. In addition to that, a partially trained policy tends to perform dangerous action to agent and environment. In order to overcome these challenges, this paper proposed a policy initialization algorithm called Policy Learning based on Completely Behavior Cloning (PLCBC). PLCBC first transforms the Model Predictive Control (MPC) controller into a piecewise affine (PWA) function using multi-parametric programming, and uses a neural network to express this function. By this way, PLCBC can completely clone the MPC controller without any performance loss, and is totally training-free. The experiments show that this initialization strategy can help agent learn at the high reward state region, and converge faster and better.

cs.LG↗

arXiv2018

Style Separation and Synthesis via Generative Adversarial Networks

Style synthesis attracts great interests recently, while few works focus on its dual problem "style separation". In this paper, we propose the Style Separation and Synthesis Generative Adversarial Network (S3-GAN) to simultaneously implement style separation and style synthesis on object photographs of specific categories. Based on the assumption that the object photographs lie on a manifold, and the contents and styles are independent, we employ S3-GAN to build mappings between the manifold and a latent vector space for separating and synthesizing the contents and styles. The S3-GAN consists of an encoder network, a generator network, and an adversarial network. The encoder network performs style separation by mapping an object photograph to a latent vector. Two halves of the latent vector represent the content and style, respectively. The generator network performs style synthesis by taking a concatenated vector as input. The concatenated vector contains the style half vector of the style target image and the content half vector of the content target image. Once obtaining the images from the generator network, an adversarial network is imposed to generate more photo-realistic images. Experiments on CelebA and UT Zappos 50K datasets demonstrate that the S3-GAN has the capacity of style separation and synthesis simultaneously, and could capture various styles in a single model.

cs.CV↗

arXiv2018

IRA assisted MMC-based topology optimization method

An Iterative Reanalysis Approximation (IRA) is integrated with the Moving Morphable Components (MMCs) based topology optimization (IRA-MMC) in this study. Compared with other classical topology optimization methods, the Finite Element (FE) based solver is replaced with the suggested IRA method. In this way, the expensive computational cost can be significantly saved by several nested iterations. The optimization of linearly elastic planar structures is constructed by the MMC, the specifically geometric parameters of which are taken as design variables to acquire explicitly geometric boundary. In the suggested algorithm, a hybrid optimizer based on the Method of Moving Asymptotes (MMA) approach and the Globally Convergent version of the Method of Moving Asymptotes (GCMMA) is suggested to improve convergence ratio and avoid local optimum. The proposed approach is evaluated by some classical benchmark problems in topology optimization, where the results show significant time saving without compromising accuracy.

math.NA↗

arXiv2018

Fast and accurate algorithms for the computation of spherically symmetric nonlocal diffusion operators on lattices

We present a unified treatment of the Fourier spectra of spherically symmetric nonlocal diffusion operators. We develop numerical and analytical results for the class of kernels with weak algebraic singularity as the distance between source and target tends to $0$. Rapid algorithms are derived for their Fourier spectra with the computation of each eigenvalue independent of all others. The algorithms are trivially parallelizable, capable of leveraging more powerful compute environments, and the accuracy of the eigenvalues is individually controllable. The algorithms include a Maclaurin series and a full divergent asymptotic series valid for any $d$ spatial dimensions. Using Drummond's sequence transformation, we prove linear complexity recurrence relations for degree-graded sequences of numerators and denominators in the rational approximations to the divergent asymptotic series. These relations are important to ensure that the algorithms are efficient, and also increase the numerical stability compared with the conventional algorithm with quadratic complexity.

math.NA↗

arXiv2018

PromID: human promoter prediction by deep learning

Computational identification of promoters is notoriously difficult as human genes often have unique promoter sequences that provide regulation of transcription and interaction with transcription initiation complex. While there are many attempts to develop computational promoter identification methods, we have no reliable tool to analyze long genomic sequences. In this work we further develop our deep learning approach that was relatively successful to discriminate short promoter and non-promoter sequences. Instead of focusing on the classification accuracy, in this work we predict the exact positions of the TSS inside the genomic sequences testing every possible location. We studied human promoters to find effective regions for discrimination and built corresponding deep learning models. These models use adaptively constructed negative set which iteratively improves the models discriminative ability. The developed promoter identification models significantly outperform the previously developed promoter prediction programs by considerably reducing the number of false positive predictions. The best model we have built has recall 0.76, precision 0.77 and MCC 0.76, while the next best tool FPROM achieved precision 0.48 and MCC 0.60 for the recall of 0.75. Our method is available at http://www.cbrc.kaust.edu.sa/PromID/.

q-bio.GN↗

arXiv2018

The Freiheitssatz and automorphisms for free brace algebras

Over a field of characteristic zero, we prove that the Freiheitssatz holds for brace algebras, the word problem for the brace algebras with a single defining relation is decidable, two generated subalgebras of free brace algebras are free, and that automorphisms of two generated free brace algebras are tame.

math.RA↗

arXiv2018

Accelerating Flash Calculation through Deep Learning Methods

In the past two decades, researchers have made remarkable progress in accelerating flash calculation, which is very useful in a variety of engineering processes. In this paper, general phase splitting problem statements and flash calculation procedures using the Successive Substitution Method are reviewed, while the main shortages are pointed out. Two acceleration methods, Newton's method and the Sparse Grids Method are presented afterwards as a comparison with the deep learning model proposed in this paper. A detailed introduction from artificial neural networks to deep learning methods is provided here with the authors' own remarks. Factors in the deep learning model are investigated to show their effect on the final result. A selected model based on that has been used in a flash calculation predictor with comparison with other methods mentioned above. It is shown that results from the optimized deep learning model meet the experimental data well with the shortest CPU time. More comparison with experimental data has been conducted to show the robustness of our model.

cs.CE↗

arXiv2018

Reconstruction of Simulation-Based Physical Field by Reconstruction Neural Network Method

A variety of modeling techniques have been developed in the past decade to reduce the computational expense and improve the accuracy of modeling. In this study, a new framework of modeling is suggested. Compared with other popular methods, a distinctive characteristic is "from image based model to analysis based model (e.g. stress, strain, and deformation)". In such a framework, a reconstruction neural network (ReConNN) model designed for simulation-based physical field's reconstruction is proposed. The ReConNN contains two submodels that are convolutional neural network (CNN) and generative adversarial net-work (GAN). The CNN is employed to construct the mapping between contour images of physical field and objective function. Subsequently, the GAN is utilized to generate more images which are similar to the existing contour images. Finally, Lagrange polynomial is applied to complete the reconstruction. However, the existing CNN models are commonly applied to the classification tasks, which seem to be difficult to handle with regression tasks of images. Meanwhile, the existing GAN architectures are insufficient to generate high-accuracy "pseudo contour images". Therefore, a ReConNN model based on a Convolution in Convolution (CIC) and a Convolutional AutoEncoder based on Wasserstein Generative Adversarial Network (WGAN-CAE) is suggested. To evaluate the performance of the proposed model representatively, a classical topology optimization procedure is considered. Then the ReConNN is utilized to the reconstruction of heat transfer process of a pin fin heat sink. It demonstrates that the proposed ReConNN model is proved to be a potential capability to reconstruct physical field for multidisciplinary, such as structural optimization.

cs.LG↗

arXiv2018

Detail Preserving Depth Estimation from a Single Image Using Attention Guided Networks

Convolutional Neural Networks have demonstrated superior performance on single image depth estimation in recent years. These works usually use stacked spatial pooling or strided convolution to get high-level information which are common practices in classification task. However, depth estimation is a dense prediction problem and low-resolution feature maps usually generate blurred depth map which is undesirable in application. In order to produce high quality depth map, say clean and accurate, we propose a network consists of a Dense Feature Extractor (DFE) and a Depth Map Generator (DMG). The DFE combines ResNet and dilated convolutions. It extracts multi-scale information from input image while keeping the feature maps dense. As for DMG, we use attention mechanism to fuse multi-scale features produced in DFE. Our Network is trained end-to-end and does not need any post-processing. Hence, it runs fast and can predict depth map in about 15 fps. Experiment results show that our method is competitive with the state-of-the-art in quantitative evaluation, but can preserve better structural details of the scene depth.

cs.CV↗

arXiv2018

DLBI: Deep learning guided Bayesian inference for structure reconstruction of super-resolution fluorescence microscopy

Super-resolution fluorescence microscopy, with a resolution beyond the diffraction limit of light, has become an indispensable tool to directly visualize biological structures in living cells at a nanometer-scale resolution. Despite advances in high-density super-resolution fluorescent techniques, existing methods still have bottlenecks, including extremely long execution time, artificial thinning and thickening of structures, and lack of ability to capture latent structures. Here we propose a novel deep learning guided Bayesian inference approach, DLBI, for the time-series analysis of high-density fluorescent images. Our method combines the strength of deep learning and statistical inference, where deep learning captures the underlying distribution of the fluorophores that are consistent with the observed time-series fluorescent images by exploring local features and correlation along time-axis, and statistical inference further refines the ultrastructure extracted by deep learning and endues physical meaning to the final image. Comprehensive experimental results on both real and simulated datasets demonstrate that our method provides more accurate and realistic local patch and large-field reconstruction than the state-of-the-art method, the 3B analysis, while our method is more than two orders of magnitude faster. The main program is available at https://github.com/lykaust15/DLBI

cs.CV↗

arXiv2018

c-axis pressure induced antiferromagnetic order in optimally P-doped BaFe2(As0.70P0.30)2 superconductor

Superconductivity in BaFe2(As1-xPx)2 iron pnictides emerges when its in-plane two-dimensional (2D) orthorhombic lattice distortion associated with nematic phase at Ts and three-dimensional (3D) collinear antiferromagnetic (AF) order at TN (Ts = TN) are gradually suppressed with increasing x, reaching optimal superconductivity around x = 0.30 with Tc $\approx$ 30 K. Here we show that a moderate uniaxial pressure along the c-axis in BaFe2(As0.70P0.30)2 spontaneously induces a 3D collinear AF order with TN = Ts > 30 K, while only slightly suppresses Tc. Although a ~ 400 MPa pressure compresses the c-axis lattice while expanding the in-plane lattice and increasing the nearest-neighbor Fe-Fe distance, it barely changes the average iron-pnictogen height in BaFe2(As0.70P0.30)2. Therefore, the pressure- induced AF order must arise from a strong in-plane magnetoelastic coupling, suggesting that the 2D nematic phase is a competing state with superconductivity.

cond-mat.supr-con↗

arXiv2018

Clause Vivification by Unit Propagation in CDCL SAT Solvers

Original and learnt clauses in Conflict-Driven Clause Learning (CDCL) SAT solvers often contain redundant literals. This may have a negative impact on performance because redundant literals may deteriorate both the effectiveness of Boolean constraint propagation and the quality of subsequent learnt clauses. To overcome this drawback, we propose a clause vivification approach that eliminates redundant literals by applying unit propagation. The proposed clause vivification is activated before the SAT solver triggers some selected restarts, and only affects a subset of original and learnt clauses, which are considered to be more relevant according to metrics like the literal block distance (LBD). Moreover, we conducted an empirical investigation with instances coming from the hard combinatorial and application categories of recent SAT competitions. The results show that a remarkable number of additional instances are solved when the proposed approach is incorporated into five of the best performing CDCL SAT solvers (Glucose, TC_Glucose, COMiniSatPS, MapleCOMSPS and MapleCOMSPS_LRB). More importantly, the empirical investigation includes an in-depth analysis of the effectiveness of clause vivification. It is worth mentioning that one of the SAT solvers described here was ranked first in the main track of SAT Competition 2017 thanks to the incorporation of the proposed clause vivification. That solver was further improved in this paper and won the bronze medal in the main track of SAT Competition 2018.

cs.AI↗

arXiv2018

Local orthorhombic lattice distortions in the paramagnetic tetragonal phase of superconducting NaFe$_{1-x}$Ni$_x$As

Understanding the interplay between nematicity, magnetism and superconductivity is pivotal for elucidating the physics of iron-based superconductors. Here we use neutron scattering to probe magnetic and nematic orders throughout the phase diagram of NaFe$_{1-x}$Ni$_x$As, finding that while both static antiferromagnetic and nematic orders compete with superconductivity, the onset temperatures for these two orders remain well-separated approaching the putative quantum critical points. We uncover local orthorhombic distortions that persist well above the tetragonal-to-orthorhombic structural transition temperature $T_{\rm s}$ in underdoped samples and extend well into the overdoped regime that exhibits neither magnetic nor structural phase transitions. These unexpected local orthorhombic distortions display Curie-Weiss temperature dependence and become suppressed below the superconducting transition temperature $T_{\rm c}$, suggesting they result from a large nematic susceptibility near optimal superconductivity. Our results account for observations of rotational symmetry-breaking above $T_{\rm s}$, and attest to the presence of significant nematic fluctuations near optimal superconductivity.

cond-mat.supr-con↗

arXiv2018

Webpage Saliency Prediction with Two-stage Generative Adversarial Networks

Web page saliency prediction is a challenge problem in image transformation and computer vision. In this paper, we propose a new model combined with web page outline information to prediction people's interest region in web page. For each web page image, our model can generate the saliency map which indicates the region of interest for people. A two-stage generative adversarial networks are proposed and image outline information is introduced for better transferring. Experiment results on FIWI dataset show that our model have better performance in terms of saliency prediction.

cs.CV↗