arXiv subjects

Wei Zeng

Publications and source records attributed to Wei Zeng.

At least 109 records · Page 6Linked to original sources

arXiv2020

Hybrid Dynamic-static Context-aware Attention Network for Action Assessment in Long Videos

The objective of action quality assessment is to score sports videos. However, most existing works focus only on video dynamic information (i.e., motion information) but ignore the specific postures that an athlete is performing in a video, which is important for action assessment in long videos. In this work, we present a novel hybrid dynAmic-static Context-aware attenTION NETwork (ACTION-NET) for action assessment in long videos. To learn more discriminative representations for videos, we not only learn the video dynamic information but also focus on the static postures of the detected athletes in specific frames, which represent the action quality at certain moments, along with the help of the proposed hybrid dynamic-static architecture. Moreover, we leverage a context-aware attention module consisting of a temporal instance-wise graph convolutional network unit and an attention unit for both streams to extract more robust stream features, where the former is for exploring the relations between instances and the latter for assigning a proper weight to each instance. Finally, we combine the features of the two streams to regress the final video score, supervised by ground-truth scores given by experts. Additionally, we have collected and annotated the new Rhythmic Gymnastics dataset, which contains videos of four different types of gymnastics routines, for evaluation of action quality assessment in long videos. Extensive experimental results validate the efficacy of our proposed method, which outperforms related approaches. The codes and dataset are available at \url{https://github.com/lingan1996/ACTION-NET}.

cs.CV↗

arXiv2020

Composition and Configuration Patterns in Multiple-View Visualizations

Multiple-view visualization (MV) is a layout design technique often employed to help users see a large number of data attributes and values in a single cohesive representation. Because of its generalizability, the MV design has been widely adopted by the visualization community to help users examine and interact with large, complex, and high-dimensional data. However, although ubiquitous, there has been little work to categorize and analyze MVs in order to better understand its design space. As a result, there has been little to no guideline in how to use the MV design effectively. In this paper, we present an in-depth study of how MVs are designed in practice. We focus on two fundamental measures of multiple-view patterns: composition, which quantifies what view types and how many are there; and configuration, which characterizes spatial arrangement of view layouts in the display space. We build a new dataset containing 360 images of MVs collected from IEEE VIS, EuroVis, and PacificVis publications 2011 to 2019, and make fine-grained annotations of view types and layouts for these visualization images. From this data we conduct composition and configuration analyses using quantitative metrics of term frequency and layout topology. We identify common practices around MVs, including relationship of view types, popular view layouts, and correlation between view types and layouts. We combine the findings into a MV recommendation system, providing interactive tools to explore the design space, and support example-based design.

cs.HC↗

arXiv2020

Variance Reduction for Deep Q-Learning using Stochastic Recursive Gradient

Deep Q-learning algorithms often suffer from poor gradient estimations with an excessive variance, resulting in unstable training and poor sampling efficiency. Stochastic variance-reduced gradient methods such as SVRG have been applied to reduce the estimation variance (Zhao et al. 2019). However, due to the online instance generation nature of reinforcement learning, directly applying SVRG to deep Q-learning is facing the problem of the inaccurate estimation of the anchor points, which dramatically limits the potentials of SVRG. To address this issue and inspired by the recursive gradient variance reduction algorithm SARAH (Nguyen et al. 2017), this paper proposes to introduce the recursive framework for updating the stochastic gradient estimates in deep Q-learning, achieving a novel algorithm called SRG-DQN. Unlike the SVRG-based algorithms, SRG-DQN designs a recursive update of the stochastic gradient estimate. The parameter update is along an accumulated direction using the past stochastic gradient information, and therefore can get rid of the estimation of the full gradients as the anchors. Additionally, SRG-DQN involves the Adam process for further accelerating the training process. Theoretical analysis and the experimental results on well-known reinforcement learning tasks demonstrate the efficiency and effectiveness of the proposed SRG-DQN algorithm.

cs.LG↗

arXiv2020

Cooperative Adaptive Learning Control for A Group of Nonholonomic UGVs by Output Feedback

A high-gain observer-based cooperative deterministic learning (CDL) control algorithm is proposed in this chapter for a group of identical unicycle-type unmanned ground vehicles (UGVs) to track over desired reference trajectories. For the vehicle states, the positions of the vehicles can be measured, while the velocities are estimated using the high-gain observer. For the trajectory tracking controller, the radial basis function (RBF) neural network (NN) is used to online estimate the unknown dynamics of the vehicle, and the NN weight convergence and estimation accuracy is guaranteed by CDL. The major challenge and novelty of this chapter is to track the reference trajectory using this observer-based CDL algorithm without the full knowledge of the vehicle state and vehicle model. In addition, any vehicle in the system is able to learn the knowledge of unmodeled dynamics along the union of trajectories experienced by all vehicle agents, such that the learned knowledge can be re-used to follow any reference trajectory defined in the learning phase. The learning-based tracking convergence and consensus learning results, as well as using learned knowledge for tracking experienced trajectories, are shown using the Lyapunov method. Simulation is given to show the effectiveness of this algorithm.

eess.SY↗

arXiv2020

On the Injection of Relativistic Electrons in the Jet of 3C 279

The acceleration of electrons in 3C 279 is investigated through analyzing the injected electron energy distribution (EED) in a time-dependent synchrotron self-Compton + external Compton emission model. In this model, it is assumed that relativistic electrons are continuously injected into the emission region, and the injected EED [$Q_e^\prime(γ^\prime)$] follows a single power-law form with low- and high-energy cutoffs $\rm γ_{min}'$ and $\rm γ_{max}'$, respectively, and the spectral index $n$, i.e, $Q_e^\prime(γ^\prime)\proptoγ^{\prime-n}$. This model is applied to 14 quasi-simultaneous spectral energy distributions (SEDs) of 3C 279. The Markov Chain Monte Carlo fitting technique is performed to obtain the best-fitting parameters and the uncertainties on the parameters. The results show that the injected EED is well constrained in each state. The value of $n$ is in the range of 2.5 to 3.8, which is larger than that expected by the classic non-relativistic shock acceleration. However, the large value of $n$ can be explained by the relativistic oblique shock acceleration. The flaring activity seems to be related to an increased acceleration efficiency, reflected in an increased $γ'_{\rm min}$ and electron injection power.

astro-ph.HE↗

arXiv2019

Prediction of Physical Load Level by Machine Learning Analysis of Heart Activity after Exercises

The assessment of energy expenditure in real life is of great importance for monitoring the current physical state of people, especially in work, sport, elderly care, health care, and everyday life even. This work reports about application of some machine learning methods (linear regression, linear discriminant analysis, k-nearest neighbors, decision tree, random forest, Gaussian naive Bayes, support-vector machine) for monitoring energy expenditures in athletes. The classification problem was to predict the known level of the in-exercise loads (in three categories by calories) by the heart rate activity features measured during the short period of time (1 minute only) after training, i.e by features of the post-exercise load. The results obtained shown that the post-exercise heart activity features preserve the information of the in-exercise training loads and allow us to predict their actual in-exercise levels. The best performance can be obtained by the random forest classifier with all 8 heart rate features (micro-averaged area under curve value AUCmicro = 0.87 and macro-averaged one AUCmacro = 0.88) and the k-nearest neighbors classifier with 4 most important heart rate features (AUCmicro = 0.91 and AUCmacro = 0.89). The limitations and perspectives of the ML methods used are outlined, and some practical advices are proposed as to their improvement and implementation for the better prediction of in-exercise energy expenditures.

cs.LG↗

arXiv2019

Off-policy Learning for Multiple Loggers

It is well known that the historical logs are used for evaluating and learning policies in interactive systems, e.g. recommendation, search, and online advertising. Since direct online policy learning usually harms user experiences, it is more crucial to apply off-policy learning in real-world applications instead. Though there have been some existing works, most are focusing on learning with one single historical policy. However, in practice, usually a number of parallel experiments, e.g. multiple AB tests, are performed simultaneously. To make full use of such historical data, learning policies from multiple loggers becomes necessary. Motivated by this, in this paper, we investigate off-policy learning when the training data coming from multiple historical policies. Specifically, policies, e.g. neural networks, can be learned directly from multi-logger data, with counterfactual estimators. In order to understand the generalization ability of such estimator better, we conduct generalization error analysis for the empirical risk minimization problem. We then introduce the generalization error bound as the new risk function, which can be reduced to a constrained optimization problem. Finally, we give the corresponding learning algorithm for the new constrained problem, where we can appeal to the minimax problems to control the constraints. Extensive experiments on benchmark datasets demonstrate that the proposed methods achieve better performances than the state-of-the-arts.

stat.ML↗

arXiv2019

$n$-VDD: Location Privacy Protection Based on Voronoi-Delaunay Duality

To date, location privacy protection is a critical issue in Location-Based Services (LBS). In this work, we propose a novel geometric framework based on the classical discrete geometric structure, the Voronoi-Delaunay duality (VDD). We utilize the fact that the user location cannot be recovered if only given an irregular $n$-sided Voronoi cell around it, and the anonymity zone is the intersection of all the parallel strips perpendicular to and bounded by $n$ Voronoi edges. The irregular Voronoi cell and its variations can be used as the concealing space to hide the user location or the region of interest and submitted to the LBS server. Within this framework, we propose multiple typical anonymizing models by introducing irregularity to the convex regular VDD structure by shifting the interior Voronoi cell, exterior Delaunay polygon, sector rays, or their combinations. The proposed methods are efficient by taking advantage of the VDD principle where main computations are linear line-line intersections. Experiments with various parameters demonstrate the efficiency and efficacy of the proposed $n$-VDD framework.

cs.CG↗

arXiv2018

3DContextNet: K-d Tree Guided Hierarchical Learning of Point Clouds Using Local and Global Contextual Cues

Classification and segmentation of 3D point clouds are important tasks in computer vision. Because of the irregular nature of point clouds, most of the existing methods convert point clouds into regular 3D voxel grids before they are used as input for ConvNets. Unfortunately, voxel representations are highly insensitive to the geometrical nature of 3D data. More recent methods encode point clouds to higher dimensional features to cover the global 3D space. However, these models are not able to sufficiently capture the local structures of point clouds. Therefore, in this paper, we propose a method that exploits both local and global contextual cues imposed by the k-d tree. The method is designed to learn representation vectors progressively along the tree structure. Experiments on challenging benchmarks show that the proposed model provides discriminative point set features. For the task of 3D scene semantic segmentation, our method significantly outperforms the state-of-the-art on the Stanford Large-Scale 3D Indoor Spaces Dataset (S3DIS).

cs.CV↗

arXiv2018

Design Rule Violation Hotspot Prediction Based on Neural Network Ensembles

Design rule check is a critical step in the physical design of integrated circuits to ensure manufacturability. However, it can be done only after a time-consuming detailed routing procedure, which adds drastically to the time of design iterations. With advanced technology nodes, the outcomes of global routing and detailed routing become less correlated, which adds to the difficulty of predicting design rule violations from earlier stages. In this paper, a framework based on neural network ensembles is proposed to predict design rule violation hotspots using information from placement and global routing. A soft voting structure and a PCA-based subset selection scheme are developed on top of a baseline neural network from a recent work. Experimental results show that the proposed architecture achieves significant improvement in model performance compared to the baseline case. For half of test cases, the performance is even better than random forest, a commonly-used ensemble learning model.

cs.LG↗

arXiv2018

Spherical Parameterization Balancing Angle and Area Distortions

This work presents a novel framework for spherical mesh parameterization. An efficient angle-preserving spherical parameterization algorithm is introduced, which is based on dynamic Yamabe flow and the conformal welding method with solid theoretic foundation. An area-preserving spherical parameterization is also discussed, which is based on discrete optimal mass transport theory. Furthermore, a spherical parameterization algorithm, which is based on the polar decomposition method, balancing angle distortion and area distortion is presented. The algorithms are tested on 3D geometric data and the experiments demonstrate the efficiency and efficacy of the proposed methods.

cs.GR↗

arXiv2018

Open Source Dataset and Machine Learning Techniques for Automatic Recognition of Historical Graffiti

Machine learning techniques are presented for automatic recognition of the historical letters (XI-XVIII centuries) carved on the stoned walls of St.Sophia cathedral in Kyiv (Ukraine). A new image dataset of these carved Glagolitic and Cyrillic letters (CGCL) was assembled and pre-processed for recognition and prediction by machine learning methods. The dataset consists of more than 4000 images for 34 types of letters. The explanatory data analysis of CGCL and notMNIST datasets shown that the carved letters can hardly be differentiated by dimensionality reduction methods, for example, by t-distributed stochastic neighbor embedding (tSNE) due to the worse letter representation by stone carving in comparison to hand writing. The multinomial logistic regression (MLR) and a 2D convolutional neural network (CNN) models were applied. The MLR model demonstrated the area under curve (AUC) values for receiver operating characteristic (ROC) are not lower than 0.92 and 0.60 for notMNIST and CGCL, respectively. The CNN model gave AUC values close to 0.99 for both notMNIST and CGCL (despite the much smaller size and quality of CGCL in comparison to notMNIST) under condition of the high lossy data augmentation. CGCL dataset was published to be available for the data science community as an open source resource.

cs.LG↗

arXiv2018

Parallel Statistical and Machine Learning Methods for Estimation of Physical Load

Several statistical and machine learning methods are proposed to estimate the type and intensity of physical load and accumulated fatigue . They are based on the statistical analysis of accumulated and moving window data subsets with construction of a kurtosis-skewness diagram. This approach was applied to the data gathered by the wearable heart monitor for various types and levels of physical activities, and for people with various physical conditions. The different levels of physical activities, loads, and fitness can be distinguished from the kurtosis-skewness diagram, and their evolution can be monitored. Several metrics for estimation of the instant effect and accumulated effect (physical fatigue) of physical loads were proposed. The data and results presented allow to extend application of these methods for modeling and characterization of complex human activity patterns, for example, to estimate the actual and accumulated physical load and fatigue, model the potential dangerous development, and give cautions and advice in real time.

cs.LG↗

arXiv2018

MQGrad: Reinforcement Learning of Gradient Quantization in Parameter Server

One of the most significant bottleneck in training large scale machine learning models on parameter server (PS) is the communication overhead, because it needs to frequently exchange the model gradients between the workers and servers during the training iterations. Gradient quantization has been proposed as an effective approach to reducing the communication volume. One key issue in gradient quantization is setting the number of bits for quantizing the gradients. Small number of bits can significantly reduce the communication overhead while hurts the gradient accuracies, and vise versa. An ideal quantization method would dynamically balance the communication overhead and model accuracy, through adjusting the number bits according to the knowledge learned from the immediate past training iterations. Existing methods, however, quantize the gradients either with fixed number of bits, or with predefined heuristic rules. In this paper we propose a novel adaptive quantization method within the framework of reinforcement learning. The method, referred to as MQGrad, formalizes the selection of quantization bits as actions in a Markov decision process (MDP) where the MDP states records the information collected from the past optimization iterations (e.g., the sequence of the loss function values). During the training iterations of a machine learning algorithm, MQGrad continuously updates the MDP state according to the changes of the loss function. Based on the information, MDP learns to select the optimal actions (number of bits) to quantize the gradients. Experimental results based on a benchmark dataset showed that MQGrad can accelerate the learning of a large scale deep neural network while keeping its prediction accuracies.

cs.LG↗

arXiv2018

Chest X-Ray Analysis of Tuberculosis by Deep Learning with Segmentation and Augmentation

The results of chest X-ray (CXR) analysis of 2D images to get the statistically reliable predictions (availability of tuberculosis) by computer-aided diagnosis (CADx) on the basis of deep learning are presented. They demonstrate the efficiency of lung segmentation, lossless and lossy data augmentation for CADx of tuberculosis by deep convolutional neural network (CNN) applied to the small and not well-balanced dataset even. CNN demonstrates ability to train (despite overfitting) on the pre-processed dataset obtained after lung segmentation in contrast to the original not-segmented dataset. Lossless data augmentation of the segmented dataset leads to the lowest validation loss (without overfitting) and nearly the same accuracy (within the limits of standard deviation) in comparison to the original and other pre-processed datasets after lossy data augmentation. The additional limited lossy data augmentation results in the lower validation loss, but with a decrease of the validation accuracy. In conclusion, besides the more complex deep CNNs and bigger datasets, the better progress of CADx for the small and not well-balanced datasets even could be obtained by better segmentation, data augmentation, dataset stratification, and exclusion of non-evident outliers.

cs.LG↗

arXiv2018

The role of jet properties in the spectral evolution of the powerful blazars

In the work, we explore the role of jet properties in the spectral evolution for a sample of Fermi-LAT bright blazars composed primarily by flat spectrum radio quasars (FSRQs). We introduce a near-equipartition log-parabola (NELP) model to fit the quasi-simultaneous multi-waveband spectral energy distributions (SEDs). The Markov Chain Monte Carlo (MCMC) method is employed to determine the best values of spectral parameters and its errors. The correlations of synchrotron peak luminosity $L_{pk}^*$ and its peak frequency $ν_{pk}^*$, SSC-dominant factor $ξ_s$ and $ν_{pk}^*$, and their implications as to the spectral evolution were studied. The statistical analysis and the comparison with theoretical approximation indicate that: (1)the spectral evolution in $L_{pk}^*-ν_{pk}^*$ plane may not only ascribed to variation of the characteristic energy $γ_{br}^\prime$ of the employed electron energy distribution (EED), but that other quantities must be varying as well.(2) for the spectral evolution in $ξ_s-ν_{pk}^*$ plane the magnetic field $B^\prime$ may play the dominant role, and $γ_{br}^\prime$ may be considered to be the subdominant role. (3) Variation of the beaming factor $δ_b$ may be mainly responsible for the absence of a possible correlation, since $B^\prime$ is strongly anti-correlated with the size of radiation zone $R_b^\prime$. The relation is agreement with that derived from the analytical theory, and should be self-consistent with the underling acceleration mechanism of the employed EED. By assuming a conical jet with opening angle $θ_{op}\sim1/Γ_b$, we find that the $γ$-ray emission site being located at a distance $\sim10^{17}-10^{19}$ cm.

astro-ph.HE↗

arXiv2018

Dimensionality Reduction in Deep Learning for Chest X-Ray Analysis of Lung Cancer

Efficiency of some dimensionality reduction techniques, like lung segmentation, bone shadow exclusion, and t-distributed stochastic neighbor embedding (t-SNE) for exclusion of outliers, is estimated for analysis of chest X-ray (CXR) 2D images by deep learning approach to help radiologists identify marks of lung cancer in CXR. Training and validation of the simple convolutional neural network (CNN) was performed on the open JSRT dataset (dataset #01), the JSRT after bone shadow exclusion - BSE-JSRT (dataset #02), JSRT after lung segmentation (dataset #03), BSE-JSRT after lung segmentation (dataset #04), and segmented BSE-JSRT after exclusion of outliers by t-SNE method (dataset #05). The results demonstrate that the pre-processed dataset obtained after lung segmentation, bone shadow exclusion, and filtering out the outliers by t-SNE (dataset #05) demonstrates the highest training rate and best accuracy in comparison to the other pre-processed datasets.

cs.LG↗

arXiv2017

Deep Learning with Lung Segmentation and Bone Shadow Exclusion Techniques for Chest X-Ray Analysis of Lung Cancer

The recent progress of computing, machine learning, and especially deep learning, for image recognition brings a meaningful effect for automatic detection of various diseases from chest X-ray images (CXRs). Here efficiency of lung segmentation and bone shadow exclusion techniques is demonstrated for analysis of 2D CXRs by deep learning approach to help radiologists identify suspicious lesions and nodules in lung cancer patients. Training and validation was performed on the original JSRT dataset (dataset #01), BSE-JSRT dataset, i.e. the same JSRT dataset, but without clavicle and rib shadows (dataset #02), original JSRT dataset after segmentation (dataset #03), and BSE-JSRT dataset after segmentation (dataset #04). The results demonstrate the high efficiency and usefulness of the considered pre-processing techniques in the simplified configuration even. The pre-processed dataset without bones (dataset #02) demonstrates the much better accuracy and loss results in comparison to the other pre-processed datasets after lung segmentation (datasets #02 and #03).

cs.LG↗