arXiv subjects

He Huang

Publications and source records attributed to He Huang.

At least 109 records · Page 6Linked to original sources

arXiv2020

Travel Time Prediction using Tree-Based Ensembles

In this paper, we consider the task of predicting travel times between two arbitrary points in an urban scenario. We view this problem from two temporal perspectives: long-term forecasting with a horizon of several days and short-term forecasting with a horizon of one hour. Both of these perspectives are relevant for planning tasks in the context of urban mobility and transportation services. We utilize tree-based ensemble methods that we train and evaluate on a dataset of taxi trip records from New York City. Through extensive data analysis, we identify relevant temporal and spatial features. We also engineer additional features based on weather and routing data. The latter is obtained via a routing solver operating on the road network. The computational results show that the addition of this routing data can be beneficial to the model performance. Moreover, employing different models for short and long-term prediction is useful as short-term models are better suited to mirror current traffic conditions. In fact, we show that accurate short-term predictions may be obtained with only little training data.

stat.ML↗

arXiv2020

Penetration of a supersonic particle at the interface in a binary complex plasma

The penetration of a supersonic particle at the interface was studied in a binary complex plasma. Inspired by the experiments performed in the PK-3 Plus Laboratory on board the International Space Station, Langevin dynamics simulations were carried out. The evolution of Mach cone at the interface was observed, where a kink of the lateral wake front was observed at the interface. By comparing the evolution of axial and radial velocity, we show that the interface solitary wave is non-linear. The dependence of the background particle dynamics in the vicinity of the interface on the penetration direction reveals that the disparity of the mobility may be the cause of various interface effects.

physics.plasm-ph↗

arXiv2020

FedVision: An Online Visual Object Detection Platform Powered by Federated Learning

Visual object detection is a computer vision-based artificial intelligence (AI) technique which has many practical applications (e.g., fire hazard monitoring). However, due to privacy concerns and the high cost of transmitting video data, it is highly challenging to build object detection models on centrally stored large training datasets following the current approach. Federated learning (FL) is a promising approach to resolve this challenge. Nevertheless, there currently lacks an easy to use tool to enable computer vision application developers who are not experts in federated learning to conveniently leverage this technology and apply it in their systems. In this paper, we report FedVision - a machine learning engineering platform to support the development of federated learning powered computer vision applications. The platform has been deployed through a collaboration between WeBank and Extreme Vision to help customers develop computer vision-based safety monitoring solutions in smart city applications. Over four months of usage, it has achieved significant efficiency improvement and cost reduction while removing the need to transmit sensitive data for three major corporate customers. To the best of our knowledge, this is the first real application of FL in computer vision-based tasks.

cs.LG↗

arXiv2019

Text-to-Viz: Automatic Generation of Infographics from Proportion-Related Natural Language Statements

Combining data content with visual embellishments, infographics can effectively deliver messages in an engaging and memorable manner. Various authoring tools have been proposed to facilitate the creation of infographics. However, creating a professional infographic with these authoring tools is still not an easy task, requiring much time and design expertise. Therefore, these tools are generally not attractive to casual users, who are either unwilling to take time to learn the tools or lacking in proper design expertise to create a professional infographic. In this paper, we explore an alternative approach: to automatically generate infographics from natural language statements. We first conducted a preliminary study to explore the design space of infographics. Based on the preliminary study, we built a proof-of-concept system that automatically converts statements about simple proportion-related statistics to a set of infographics with pre-designed styles. Finally, we demonstrated the usability and usefulness of the system through sample results, exhibits, and expert reviews.

cs.HC↗

arXiv2019

Generative Dual Adversarial Network for Generalized Zero-shot Learning

This paper studies the problem of generalized zero-shot learning which requires the model to train on image-label pairs from some seen classes and test on the task of classifying new images from both seen and unseen classes. Most previous models try to learn a fixed one-directional mapping between visual and semantic space, while some recently proposed generative methods try to generate image features for unseen classes so that the zero-shot learning problem becomes a traditional fully-supervised classification problem. In this paper, we propose a novel model that provides a unified framework for three different approaches: visual-> semantic mapping, semantic->visual mapping, and metric learning. Specifically, our proposed model consists of a feature generator that can generate various visual features given class embeddings as input, a regressor that maps each visual feature back to its corresponding class embedding, and a discriminator that learns to evaluate the closeness of an image feature and a class embedding. All three components are trained under the combination of cyclic consistency loss and dual adversarial loss. Experimental results show that our model not only preserves higher accuracy in classifying images from seen classes, but also performs better than existing state-of-the-art models in in classifying images from unseen classes.

cs.CV↗

arXiv2019

Slow Vortex Creep Induced by Strong Grain Boundary Pinning in Advanced Ba122 Superconducting Tapes

We report the temperature, magnetic field and time dependences of magnetization in advanced Ba122 superconducting tapes. The sample exhibits a peculiar vortex creep behavior. Below 10 K, the normalized magnetization relaxation rate S=dln(-M)/dln(t) shows a temperature insensitive plateau with a value comparable to that of the low temperature superconductors, which can be explained within the framework of the collective creep theory. It then enters into a second collective creep regime when the temperature increases. Interestingly, the relaxation rate below 20 K tends to saturate with the increasing field. However, it changes to a power law dependence on field at a higher temperature. A vortex phase diagram composed of the collective and the plastic creep regions is concluded. Benefit from the strong grain boundary pinning, the advanced Ba122 superconducting tape has promising potential to be applied not only in liquid helium but also in liquid hydrogen or at the temperature accessible with cryocoolers.

cond-mat.supr-con↗

arXiv2019

First performance test of 30 mm iron-based superconductor single pancake coil under 24 T background field

The iron-based superconductor (IBS) single pancake coil (SPC) with 30 mm inner diameter was firstly fabricated and tested under 24 T background field. This SPC was successfully made using the 7-filamentary Ba1-xKxFe2As2 (Ba122) tape by wind-and-react method. This IBS coil show the highest Ic value at magnetic field reported so far. For example, the transport critical current of this Ba122 SPC achieved 35 A at 4.2 K and 10 T, which is about half of that of short sample. This indicates that the non-insulation winding process together with the stainless-steel tape is suitable to the iron-based superconductor. Even more encouraging is the fact that the Ic of this SPC is still as high as 26 A under 24 T background field, which is still about 40% of that at zero external magnetic field. These results clearly demonstrate that the iron-based superconductors are very promising for high-field magnet applications.

cond-mat.supr-con↗

arXiv2018

On-Demand Video Dispatch Networks: A Scalable End-to-End Learning Approach

We design a dispatch system to improve the peak service quality of video on demand (VOD). Our system predicts the hot videos during the peak hours of the next day based on the historical requests, and dispatches to the content delivery networks (CDNs) at the previous off-peak time. In order to scale to billions of videos, we build the system with two neural networks, one for video clustering and the other for dispatch policy developing. The clustering network employs autoencoder layers and reduces the video number to a fixed value. The policy network employs fully connected layers and ranks the clustered videos with dispatch probabilities. The two networks are coupled with weight-sharing temporal layers, which analyze the video request sequences with convolutional and recurrent modules. Therefore, the clustering and dispatch tasks are trained in an end-to-end mechanism. The real-world results show that our approach achieves an average prediction accuracy of 17%, compared with 3% from the present baseline method, for the same amount of dispatches.

cs.NI↗

arXiv2018

An Introduction to Image Synthesis with Generative Adversarial Nets

There has been a drastic growth of research in Generative Adversarial Nets (GANs) in the past few years. Proposed in 2014, GAN has been applied to various applications such as computer vision and natural language processing, and achieves impressive performance. Among the many applications of GAN, image synthesis is the most well-studied one, and research in this area has already demonstrated the great potential of using GAN in image synthesis. In this paper, we provide a taxonomy of methods used in image synthesis, review different models for text-to-image synthesis and image-to-image translation, and discuss some evaluation metrics as well as possible future research directions in image synthesis with GAN.

cs.CV↗

arXiv2018

Chemical stability and superconductivity in Ag-sheathed CaKFe4As4 superconducting tapes

Ag-sheathed CaKFe4As4 superconducting tapes have been fabricated via the ex-situ powder-in-tube method. Thermal and X-ray diffraction analyses suggest that the CaKFe4As4 phase is unstable at high temperatures. It decomposes into the CaAgAs phase which reacts strongly with the silver sheath. We therefore sintered the tape at 500C and obtain a transport critical current density Jc(4.2 K, 0 T)~ 2.7x10^4 A/cm2. The pinning potential derived from magnetoresistance measurements is one order of magnitude lower than that of the (Ba/Sr)1-xKxFe2As2 tapes. Combining with the scanning electron microscopy and magneto-optical imaging results, we suggest that bad connectivity between superconducting grains caused by the low sintering temperature is the main factor responsible for the low Jc. However, this system is still a promising candidate for superconducting wires and tapes if we further optimize the post-annealing process to achieve better grain connectivity.

cond-mat.supr-con↗

arXiv2018

dpMood: Exploiting Local and Periodic Typing Dynamics for Personalized Mood Prediction

Mood disorders are common and associated with significant morbidity and mortality. Early diagnosis has the potential to greatly alleviate the burden of mental illness and the ever increasing costs to families and society. Mobile devices provide us a promising opportunity to detect the users' mood in an unobtrusive manner. In this study, we use a custom keyboard which collects keystrokes' meta-data and accelerometer values. Based on the collected time series data in multiple modalities, we propose a deep personalized mood prediction approach, called {\pro}, by integrating convolutional and recurrent deep architectures as well as exploring each individual's circadian rhythm. Experimental results not only demonstrate the feasibility and effectiveness of using smart-phone meta-data to predict the presence and severity of mood disturbances in bipolar subjects, but also show the potential of personalized medical treatment for mood disorders.

cs.HC↗

arXiv2018

NavigationNet: A Large-scale Interactive Indoor Navigation Dataset

Indoor navigation aims at performing navigation within buildings. In scenes like home and factory, most intelligent mobile devices require an functionality of routing to guide itself precisely through indoor scenes to complete various tasks in order to serve human. In most scenarios, we expected an intelligent device capable of navigating itself in unseen environment. Although several solutions have been proposed to deal with this issue, they usually require pre-installed beacons or a map pre-built with SLAM, which means that they are not capable of working in novel environments. To address this, we proposed NavigationNet, a computer vision dataset and benchmark to allow the utilization of deep reinforcement learning on scene-understanding-based indoor navigation. We also proposed and formalized several typical indoor routing problems that are suitable for deep reinforcement learning.

cs.CV↗

arXiv2018

Catalogue of Topological Electronic Materials

Topological electronic materials are new quantum states of matter hosting novel linear responses in the bulk and anomalous gapless states at the boundary, and are for scientific and applied reasons under intensive research in physics and in materials sciences. The detection for such materials has so far been hindered by the level of complication involved in the calculation of the so-called topological invariants, and is hence considered a specialized task that requires both experience with materials and expertise with advanced theoretical tools. Here we introduce an effective, efficient and fully automated algorithm in obtaining the topological invariants for all non-magnetic materials that are known to human, based on recently developed principles that allow for exhaustive mappings between the symmetry representation of occupied bands and the topological invariants. Our algorithm requires as input only the occupied-band information (energy and wavefunction) at a handful (up to eight) of high-symmetry points in the Brillouin zone, which is readily calculable with any first-principles software. In return, it is capable of providing a detailed topological classification of all non-magnetic materials. Equipped with this method we have scanned through a total of 39519 materials available in structural databases, and found that as many as 8056 of them are actually topological (8889 if spin-orbital coupling is neglected). These are further catalogued into classes of 5005 topological semimetals,1814 topological insulators and 1237 topological crystalline insulators, most of which are new to human knowledge. All the results are available and searchable at http://materiae.iphy.ac.cn/ .

cond-mat.mtrl-sci↗

arXiv2018

Performance Analysis of Energy Detection over Composite kappa-miu Shadowed Fading Channels

Energy detection is a reliable non-coherent signal processing technology of spectrum sensing of cognitive radio networks, which thanks to its low complexity, no requirement of priori received information and fast sensing ability etc. Since the excellent performance of energy detection would be actually affected by physical multipath fading, this paper is concentrating on characteristics analysis of energy detection over composite shadowed fading channels. The small-scale and line-of-sight fading distribution consists of particular examples such as Rayleigh, Hoyt, Nakagami-m and one sided Gaussian distributions. Based on this, we derive the probability density function of signal envelope and signal-to-noise ratio of the composite shadowed fading channels, which could accurately present the line-of-sight shadowed fading characterization. Subsequently the exact close-form expressions with infinite series formulation for the appropriate detection probability have been firstly extended to estimate detection capacity of the above-mentioned model by adopting Inverse Gaussian asymptotic distribution. In addition, the absolute truncation error is deduced for evaluating minimum detection efficiency. The established model can be also applied in detection estimation with non-integral fading parameters. Last but not least, the analytical results and quantification performance are approved by numerically evaluation with MATHEMATICA and MATLAB as the power variables of dominant components changes.

cs.IT↗

arXiv2018

MARS: Memory Attention-Aware Recommender System

In this paper, we study the problem of modeling users' diverse interests. Previous methods usually learn a fixed user representation, which has a limited ability to represent distinct interests of a user. In order to model users' various interests, we propose a Memory Attention-aware Recommender System (MARS). MARS utilizes a memory component and a novel attentional mechanism to learn deep \textit{adaptive user representations}. Trained in an end-to-end fashion, MARS adaptively summarizes users' interests. In the experiments, MARS outperforms seven state-of-the-art methods on three real-world datasets in terms of recall and mean average precision. We also demonstrate that MARS has a great interpretability to explain its recommendation results, which is important in many recommendation scenarios.

cs.IR↗

arXiv2018

Superconducting joint between iron-based superconductor tapes

Superconducting joints are essential for iron-based superconductor's applications in future. In this study, a process for fabricating superconducting joints between Sr1-xKxFe2As2 (Sr-122) tapes is developed for the first time. The Ag sheath was peeled off from one side of each sample. The exposed superconducting parts of the two tapes were joined and wrapped again with Ag foil. The diffusion bonding of the iron-based superconducting joint was achieved by hot-pressing process in Argon atmosphere. The superconducting properties, microstructures and the elements distribution of the joint regions had been investigated. The pressure and pressing times were optimized in order to enhance the transport current of the joints. At 4.2 K and 10 T, a transport critical current Ic of 57 A for the joint was obtained, which is approximately 63.3% of the current capacity of the tapes themselves. Furthermore, the joint resistances dV/dI were estimated from the V-I curve of the joints and the calculated joint resistances values are below 10^-9 Ohm. These results demonstrate that the hot pressing was useful for fabricating the superconducting joint samples.

cond-mat.supr-con↗

arXiv2018

Functional Variable Selection for EMG-based Control of a Robotic Hand Prosthetic

State-of-the-art robotic hand prosthetics generate finger and wrist movement through pattern recognition (PR) algorithms using features of forearm electromyogram (EMG) signals, but re- quires extensive training and is prone to poor predictions for conditions outside the training data (Peerdeman et al., 2011; Scheme et al., 2010). We propose a novel approach to develop a dynamic robotic limb by utilizing the recent history of EMG signals in a model that accounts for physiological features of hand movement which are ignored by PR algorithms. We do this by viewing EMG signals as functional covariates and develop a functional linear model that quantifies the effect of the EMG signals on finger/wrist velocity through a bivariate coefficient function that is allowed to vary with current finger/wrist position. The model is made par- simonious and interpretable through a two-step variable selection procedure, called Sequential Adaptive Functional Empirical group LASSO (SAFE-gLASSO). Numerical studies show excel- lent selection and prediction properties of SAFE-gLASSO compared to popular alternatives. For our motivating dataset, the method correctly identifies the few EMG signals that are known to be important for an able-bodied subject with negligible false positives and the model can be directly implemented in a robotic prosthetic.

stat.AP↗

arXiv2018

Amino-acid network clique analysis of protein mutation correlation effects: a case study of lysozme

Optimizing amino-acid mutations has been a most challenging task in modern bio- industrial enzyme designing. It is well known that many successful designs often hinge on extensive correlations among mutations at different sites within the enzyme, however, the underpinning mechanism for these correlations is far from clear. Here, we present a topology-based model to quantitively characterize correlation effects between mutations. The method is based on the molecular dynamic simulations and the amino-acid network clique analysis that simply examines if two single mutation sites belong to some 3-clique. We analyzed 13 dual mutations of T4 phage lysozyme and found that the clique-based model successfully distinguishes highly correlated or non-additive double-site mutations from those with less correlation or additive mutations. We also applied the model to the protein Eglin c whose topology is significantly distinct from that of T4 phage lysozyme, and found that the model can, to some extension, still identify non-additive mutations from additive ones. Our calculations showed that mutation correlation effects may heavily depend on topology relationship among mutation sites, which can be quantitatively characterized using amino-acid network k-cliques. We also showed that double-site mutation correlations can be significantly altered by exerting a third mutation, indicating that more detailed physico-chemistry interactions might be considered with the network model for better understanding of the elusive mutation-correlation principle.

q-bio.QM↗