arXiv subjects

Yan Yang

Publications and source records attributed to Yan Yang.

At least 145 records · Page 8Linked to original sources

arXiv2021

Prediction of properties of metal alloy materials based on machine learning

Density functional theory and its optimization algorithm are the main methods to calculate the properties in the field of materials. Although the calculation results are accurate, it costs a lot of time and money. In order to alleviate this problem, we intend to use machine learning to predict material properties. In this paper, we conduct experiments on atomic volume, atomic energy and atomic formation energy of metal alloys, using the open quantum material database. Through the traditional machine learning models, deep learning network and automated machine learning, we verify the feasibility of machine learning in material property prediction. The experimental results show that the machine learning can predict the material properties accurately.

cond-mat.mtrl-sci↗

arXiv2021

Some new results on bar visibility of digraphs

Visibility representation of digraphs was introduced by Axenovich, Beveridge, Hutch\-inson, and West (\emph{SIAM J. Discrete Math.} {\bf 27}(3) (2013) 1429--1449) as a natural generalization of $t$-bar visibility representation of undirected graphs. A {\it $t$-bar visibility representation} of a digraph $G$ assigns each vertex at most $t$ horizontal bars in the plane so that there is an arc $xy$ in the digraph if and only if some bar for $x$ "sees" some bar for $y$ above it along an unblocked vertical strip with positive width. The {\it visibility number} $b(G)$ is the least $t$ such that $G$ has a $t$-bar visibility representation. In this paper, we solve several problems about $b(G)$ posed by Axenovich et al.\ and prove that determining whether the bar visibility number of a digraph is $2$ is NP-complete.

math.CO↗

arXiv2021

Nearest Neighborhood-Based Deep Clustering for Source Data-absent Unsupervised Domain Adaptation

In the classic setting of unsupervised domain adaptation (UDA), the labeled source data are available in the training phase. However, in many real-world scenarios, owing to some reasons such as privacy protection and information security, the source data is inaccessible, and only a model trained on the source domain is available. This paper proposes a novel deep clustering method for this challenging task. Aiming at the dynamical clustering at feature-level, we introduce extra constraints hidden in the geometric structure between data to assist the process. Concretely, we propose a geometry-based constraint, named semantic consistency on the nearest neighborhood (SCNNH), and use it to encourage robust clustering. To reach this goal, we construct the nearest neighborhood for every target data and take it as the fundamental clustering unit by building our objective on the geometry. Also, we develop a more SCNNH-compliant structure with an additional semantic credibility constraint, named semantic hyper-nearest neighborhood (SHNNH). After that, we extend our method to this new geometry. Extensive experiments on three challenging UDA datasets indicate that our method achieves state-of-the-art results. The proposed method has significant improvement on all datasets (as we adopt SHNNH, the average accuracy increases by over 3.0% on the large-scaled dataset). Code is available at https://github.com/tntek/N2DCX.

cs.CV↗

arXiv2021

A generalised multipath delayed-choice experiment on a large-scale quantum nanophotonic chip

Famous double-slit or double-path experiments, implemented in a Young's or Mach-Zehnder interferometer, have confirmed the dual nature of quantum matter, When a stream of photons, neutrons, atoms, or molecules, passes through two slits, either wave-like interference fringes build up on a screen, or particle-like which-path distribution can be ascertained. These quantum objects exhibit both wave and particle properties but exclusively, depending on the way they are measured. In an equivalent Mach-Zehnder configuration, the object displays either wave or particle nature in the presence or absence of a beamsplitter, respectively, that represents the choice of which-measurement. Wheeler further proposed a gedanken experiment, in which the choice of which-measurement is delayed, i.e. determined after the object has already entered the interferometer, so as to exclude the possibility of predicting which-measurement it will confront. The delayed-choice experiments have enabled significant demonstrations of genuine two-path duality of different quantum objects. Recently, a quantum controlled version of delayed-choice was proposed by Ionicioiu and Terno, by introducing a quantum-controlled beamsplitter that is in a coherent superposition of presence and absence. It represents a controllable experiment platform that can not only reveal wave and particle characters, but also their superposition. Moreover, a quantitative description of two-slit duality relation was initialized in Wootters and Zurek's seminal work and formalized by Greenberger,et. al. as D2+V2<=1, where D is the distinguishability of whichpath information, and V is the contrast visibility of interference. In this regard, getting which-path information exclusively reduces the interference visibility, and vice versa. This double-path duality relation has been tested in pioneer experiments and recently in delayed-choice measurements.

quant-ph↗

arXiv2021

SLAKE: A Semantically-Labeled Knowledge-Enhanced Dataset for Medical Visual Question Answering

Medical visual question answering (Med-VQA) has tremendous potential in healthcare. However, the development of this technology is hindered by the lacking of publicly-available and high-quality labeled datasets for training and evaluation. In this paper, we present a large bilingual dataset, SLAKE, with comprehensive semantic labels annotated by experienced physicians and a new structural medical knowledge base for Med-VQA. Besides, SLAKE includes richer modalities and covers more human body parts than the currently available dataset. We show that SLAKE can be used to facilitate the development and evaluation of Med-VQA systems. The dataset can be downloaded from http://www.med-vqa.com/slake.

cs.CV↗

arXiv2020

Upper bounds on the signed edge domination number of a graph

A signed edge domination function (or SEDF) of a simple graph $G=(V,E)$ is a function $f: E\rightarrow \{1,-1\}$ such that $\sum_{e'\in N[e]}f(e')\ge 1$ holds for each edge $e\in E$, where $N[e]$ is the set of edges in $G$ that share at least one endpoint with $e$. Let $γ_s'(G)$ denote the minimum value of $f(G)$ among all SEDFs $f$, where $f(G)=\sum_{e\in E}f(e)$.In 2005, Xu conjectured that $γ_s'(G)\le n-1$, where $n$ is the order of $G$. This conjecture has been proved for the two cases $v_{odd}(G)=0$ and $v_{even}(G)=0$, where $v_{odd}(G)$ (resp. $v_{even}(G)$) is the number of odd (resp. even) vertices in $G$. This article proves Xu's conjecture for $v_{even}(G)\in \{1, 2\}$. We also show that for any simple graph $G$ of order $n$, $γ_s'(G)\le n+v_{odd}(G)/2$ and $γ_s'(G)\le n-2+v_{even}(G)$ when $v_{even}(G)>0$, and thus $γ_s'(G)\le (4n-2)/3$. Our result improves the best current upper bound of $γ_s'(G)\le \lceil 3n/2\rceil$.

math.CO↗

arXiv2020

RealSmileNet: A Deep End-To-End Network for Spontaneous and Posed Smile Recognition

Smiles play a vital role in the understanding of social interactions within different communities, and reveal the physical state of mind of people in both real and deceptive ways. Several methods have been proposed to recognize spontaneous and posed smiles. All follow a feature-engineering based pipeline requiring costly pre-processing steps such as manual annotation of face landmarks, tracking, segmentation of smile phases, and hand-crafted features. The resulting computation is expensive, and strongly dependent on pre-processing steps. We investigate an end-to-end deep learning model to address these problems, the first end-to-end model for spontaneous and posed smile recognition. Our fully automated model is fast and learns the feature extraction processes by training a series of convolution and ConvLSTM layer from scratch. Our experiments on four datasets demonstrate the robustness and generalization of the proposed model by achieving state-of-the-art performances.

cs.CV↗

arXiv2020

DeepSTCL: A Deep Spatio-temporal ConvLSTM for Travel Demand Prediction

Urban resource scheduling is an important part of the development of a smart city, and transportation resources are the main components of urban resources. Currently, a series of problems with transportation resources such as unbalanced distribution and road congestion disrupt the scheduling discipline. Therefore, it is significant to predict travel demand for urban resource dispatching. Previously, the traditional time series models were used to forecast travel demand, such as AR, ARIMA and so on. However, the prediction efficiency of these methods is poor and the training time is too long. In order to improve the performance, deep learning is used to assist prediction. But most of the deep learning methods only utilize temporal dependence or spatial dependence of data in the forecasting process. To address these limitations, a novel deep learning traffic demand forecasting framework which based on Deep Spatio-Temporal ConvLSTM is proposed in this paper. In order to evaluate the performance of the framework, an end-to-end deep learning system is designed and a real dataset is used. Furthermore, the proposed method can capture temporal dependence and spatial dependence simultaneously. The closeness, period and trend components of spatio-temporal data are used in three predicted branches. These branches have the same network structures, but do not share weights. Then a linear fusion method is used to get the final result. Finally, the experimental results on DIDI order dataset of Chengdu demonstrate that our method outperforms traditional models with accuracy and speed.

cs.LG↗

arXiv2020

Statistics of Kinetic Dissipation in Earth's Magnetosheath -- MMS Observations

A familiar problem in space and astrophysical plasmas is to understand how dissipation and heating occurs. These effects are often attributed to the cascade of broadband turbulence which transports energy from large scale reservoirs to small scale kinetic degrees of freedom. When collisions are infrequent, local thermodynamic equilibrium is not established. In this case the final stage of energy conversion becomes more complex than in the fluid case, and both pressure-dilatation and pressure strain interactions (Pi-D $\equiv -Π_{ij} D_{ij}$) become relevant and potentially important. Pi-D in plasma turbulence has been studied so far primarily using simulations. The present study provides a statistical analysis of Pi-D in the Earth's magnetosheath using the unique measurement capabilities of the Magnetospheric Multiscale (MMS) mission. We find that the statistics of Pi-D in this naturally occurring plasma environment exhibit strong resemblance to previously established fully kinetic simulations results. The conversion of energy is concentrated in space and occurs near intense current sheets, but not within them. This supports recent suggestions that the chain of energy transfer channels involves regional, rather than pointwise, correlations.

physics.plasm-ph↗

arXiv2020

High order mixed finite elements with mass lumping for elasticity on triangular grids

A family of conforming mixed finite elements with mass lumping on triangular grids are presented for linear elasticity. The stress field is approximated by symmetric $H({\rm div})-P_k (k\geq 3)$ polynomial tensors enriched with higher order bubbles so as to allow mass lumping, which can be viewed as the Hu-Zhang elements enriched with higher order interior bubble functions. The displacement field is approximated by $C^{-1}-P_{k-1}$ polynomial vectors enriched with higher order terms to ensure the stability condition. For both the proposed mixed elements and their mass lumping schemes, optimal error estimates are derived for the stress with $H(\rm div)$ norm and the displacement with $L^2$ norm. Numerical results confirm the theoretical analysis.

math.NA↗

arXiv2020

In situ Measurement of Curvature of Magnetic Field in Turbulent Space Plasmas: A Statistical Study

Using in situ data, accumulated in the turbulent magnetosheath by the Magnetospheric Multiscale (MMS) Mission, we report a statistical study of magnetic field curvature and discuss its role in the turbulent space plasmas. Consistent with previous simulation results, the Probability Distribution Function (PDF) of the curvature is shown to have distinct power-law tails for both high and low value limits. We find that the magnetic-field-line curvature is intermittently distributed in space. High curvature values reside near weak magnetic-field regions, while low curvature values are correlated with small magnitude of the force acting normal to the field lines. A simple statistical treatment provides an explanation for the observed curvature distribution. This novel statistical characterization of magnetic curvature in space plasma provides a starting point for assessing, in a turbulence context, the applicability and impact of particle energization processes, such as curvature drift, that rely on this fundamental quantity.

physics.space-ph↗

arXiv2020

A novel semi-supervised multi-view clustering framework for screening Parkinson's disease

In recent years, there are many research cases for the diagnosis of Parkinson's disease (PD) with the brain magnetic resonance imaging (MRI) by utilizing the traditional unsupervised machine learning methods and the supervised deep learning models. However, unsupervised learning methods are not good at extracting accurate features among MRIs and it is difficult to collect enough data in the field of PD to satisfy the need of training deep learning models. Moreover, most of the existing studies are based on single-view MRI data, of which data characteristics are not sufficient enough. In this paper, therefore, in order to tackle the drawbacks mentioned above, we propose a novel semi-supervised learning framework called Semi-supervised Multi-view learning Clustering architecture technology (SMC). The model firstly introduces the sliding window method to grasp different features, and then uses the dimensionality reduction algorithms of Linear Discriminant Analysis (LDA) to process the data with different features. Finally, the traditional single-view clustering and multi-view clustering methods are employed on multiple feature views to obtain the results. Experiments show that our proposed method is superior to the state-of-art unsupervised learning models on the clustering effect. As a result, it may be noted that, our work could contribute to improving the effectiveness of identifying PD by previous labeled and subsequent unlabeled medical MRI data in the realistic medical environment.

eess.IV↗

arXiv2019

Deep Air Quality Forecasting Using Hybrid Deep Learning Framework

Air quality forecasting has been regarded as the key problem of air pollution early warning and control management. In this paper, we propose a novel deep learning model for air quality (mainly PM2.5) forecasting, which learns the spatial-temporal correlation features and interdependence of multivariate air quality related time series data by hybrid deep learning architecture. Due to the nonlinear and dynamic characteristics of multivariate air quality time series data, the base modules of our model include one-dimensional Convolutional Neural Networks (1D-CNNs) and Bi-directional Long Short-term Memory networks (Bi-LSTM). The former is to extract the local trend features and spatial correlation features, and the latter is to learn spatial-temporal dependencies. Then we design a jointly hybrid deep learning framework based on one-dimensional CNNs and Bi-LSTM for shared representation features learning of multivariate air quality related time series data. We conduct extensive experimental evaluations using two real-world datasets, and the results show that our model is capable of dealing with PM2.5 air pollution forecasting with satisfied accuracy.

cs.LG↗

arXiv2019

Fast Calculation of Probabilistic Power Flow: A Model-based Deep Learning Approach

Probabilistic power flow (PPF) plays a critical role in power system analysis. However, the high computational burden makes it challenging for the practical implementation of PPF. This paper proposes a model-based deep learning approach to overcome the computational challenge. A deep neural network (DNN) is used to approximate the power flow calculation and is trained according to the physical power flow equations to improve its learning ability. The training process consists of several steps: 1) the branch flows are added into the objective function of the DNN as a penalty term, which improves the approximation accuracy of the DNN; 2) the gradients used in the back propagation process are simplified according to the physical characteristics of the transmission grid, which accelerates the training speed while maintaining effective guidance of the physical model; and 3) an improved initialization method for the DNN parameters is proposed to improve the convergence speed. The simulation results demonstrate the accuracy and efficiency of the proposed method in standard IEEE and utility benchmark systems.

eess.SP↗

arXiv2019

Learning with Noisy Labels for Sentence-level Sentiment Classification

Deep neural networks (DNNs) can fit (or even over-fit) the training data very well. If a DNN model is trained using data with noisy labels and tested on data with clean labels, the model may perform poorly. This paper studies the problem of learning with noisy labels for sentence-level sentiment classification. We propose a novel DNN model called NetAb (as shorthand for convolutional neural Networks with Ab-networks) to handle noisy labels during training. NetAb consists of two convolutional neural networks, one with a noise transition layer for dealing with the input noisy labels and the other for predicting 'clean' labels. We train the two networks using their respective loss functions in a mutual reinforcement manner. Experimental results demonstrate the effectiveness of the proposed model.

cs.CL↗

arXiv2019

Fast Calculation of Probabilistic Optimal Power Flow: A Deep Learning Approach

Probabilistic optimal power flow (POPF) is an important analytical tool to ensure the secure and economic operation of power systems. POPF needs to solve enormous nonlinear and nonconvex optimization problems. The huge computational burden has become the major bottleneck for the practical application. This paper presents a deep learning approach to solve the POPF problem efficiently and accurately. Taking advantage of the deep structure and reconstructive strategy of stacked denoising auto encoders (SDAE), a SDAE-based optimal power flow (OPF) is developed to extract the high-level nonlinear correlations between the system operating condition and the OPF solution. A training process is designed to learn the feature of POPF. The trained SDAE network can be utilized to conveniently calculate the OPF solution of random samples generated by Monte-Carlo simulation (MCS) without the need of optimization. A modified IEEE 118-bus power system is simulated to demonstrate the effectiveness of the proposed method.

eess.SP↗

arXiv2019

A New Solution to Market Definition: An Approach Based on Multi-dimensional Substitutability Statistics

Market definition is an important component in the premerger investigation, but the models used in the market definition have not developed much in the past three decades since the Critical Loss Analysis (CLA) was proposed in 1989. The CLA helps the Hypothetical Monopolist Test to determine whether the hypothetical monopolist is going to profit from the small but significant and non-transitory increase in price (SSNIP). However, the CLA has long been criticized by academic scholars for its tendency to conclude a narrow market. Although the CLA was adopted by the 2010 Horizontal Merger Guidelines (the 2010 Guidelines), the criticisms are likely still valid. In this dissertation, we discussed the mathematical deduction of CLA, the data used, and the SSNIP defined by the Agencies. Based on our research, we concluded that the narrow market conclusion was due to the incorrect implementation of the CLA; not the model itself. On the other hand, there are other unresolvable problems in the CLA and the Hypothetical Monopolist Test. The SSNIP test and the CLA are bright resolutions for market definition problem during their time, but we have more advanced tools to solve the task nowadays. In this dissertation, we propose a model which is based directly on the multi-dimensional substitutability between the products and is capable of maximizing the substitutability of product features within each group. Since the 2010 Guidelines does not exclude the use of models other than the ones mentioned by the Guidelines, our method can hopefully supplement the current models to show a better picture of the substitutive relations and provide a more stable definition of the market.

econ.GN↗

arXiv2019

Forward and Backward Knowledge Transfer for Sentiment Classification

This paper studies the problem of learning a sequence of sentiment classification tasks. The learned knowledge from each task is retained and used to help future or subsequent task learning. This learning paradigm is called Lifelong Learning (LL). However, existing LL methods either only transfer knowledge forward to help future learning and do not go back to improve the model of a previous task or require the training data of the previous task to retrain its model to exploit backward/reverse knowledge transfer. This paper studies reverse knowledge transfer of LL in the context of naive Bayesian (NB) classification. It aims to improve the model of a previous task by leveraging future knowledge without retraining using its training data. This is done by exploiting a key characteristic of the generative model of NB. That is, it is possible to improve the NB classifier for a task by improving its model parameters directly by using the retained knowledge from other tasks. Experimental results show that the proposed method markedly outperforms existing LL baselines.

cs.AI↗