arXiv subjects

Weiwei Zhang

Publications and source records attributed to Weiwei Zhang.

At least 91 records · Page 5Linked to original sources

arXiv2022

Image Magnification Network for Vessel Segmentation in OCTA Images

Optical coherence tomography angiography (OCTA) is a novel non-invasive imaging modality that allows micron-level resolution to visualize the retinal microvasculature. The retinal vessel segmentation in OCTA images is still an open problem, and especially the thin and dense structure of the capillary plexus is an important challenge of this problem. In this work, we propose a novel image magnification network (IMN) for vessel segmentation in OCTA images. Contrary to the U-Net structure with a down-sampling encoder and up-sampling decoder, the proposed IMN adopts the design of up-sampling encoding and then down-sampling decoding. This design is to capture more low-level image details to reduce the omission of small structures. The experimental results on three open OCTA datasets show that the proposed IMN with an average dice score of 90.2% achieves the best performance in vessel segmentation of OCTA images. Besides, we also demonstrate the superior performance of IMN in cross-field image vessel segmentation and vessel skeleton extraction.

eess.IV↗

arXiv2021

Testability-Aware Low Power Controller Design with Evolutionary Learning

XORNet-based low power controller is a popular technique to reduce circuit transitions in scan-based testing. However, existing solutions construct the XORNet evenly for scan chain control, and it may result in sub-optimal solutions without any design guidance. In this paper, we propose a novel testability-aware low power controller with evolutionary learning. The XORNet generated from the proposed genetic algorithm (GA) enables adaptive control for scan chains according to their usages, thereby significantly improving XORNet encoding capacity, reducing the number of failure cases with ATPG and decreasing test data volume. Experimental results indicate that under the same control bits, our GA-guided XORNet design can improve the fault coverage by up to 2.11%. The proposed GA-guided XORNets also allows reducing the number of control bits, and the total testing time decreases by 20.78% on average and up to 47.09% compared to the existing design without sacrificing test coverage.

cs.LG↗

arXiv2021

High Reynolds number airfoil turbulence modeling method based on machine learning technique

In this paper, a turbulence model based on deep neural network is developed for turbulent flow around airfoil at high Reynolds numbers. According to the data got from the Spalart-Allmaras (SA) turbulence model, we build a neural network model that maps flow features to eddy viscosity. The model is then used to replace the SA turbulence model to mutually couple with the CFD solver. We build this suitable data-driven turbulence model mainly from the inputs, outputs features and loss function of the model. A feature selection method based on feature importance is also implemented. The results show that this feature selection method can effectively remove redundant features. The model based on the new input features has better accuracy and stability in mutual coupling with the CFD solver. The force coefficient obtained from solution can match the sample data well. The developed model also shows strong generalization at different inflow condition (angle of attack, Mach number, Reynolds number and airfoil).

physics.flu-dyn↗

arXiv2021

Challenging Machine Learning-based Clone Detectors via Semantic-preserving Code Transformations

Software clone detection identifies similar code snippets. It has been an active research topic that attracts extensive attention over the last two decades. In recent years, machine learning (ML) based detectors, especially deep learning-based ones, have demonstrated impressive capability on clone detection. It seems that this longstanding problem has already been tamed owing to the advances in ML techniques. In this work, we would like to challenge the robustness of the recent ML-based clone detectors through code semantic-preserving transformations. We first utilize fifteen simple code transformation operators combined with commonly-used heuristics (i.e., Random Search, Genetic Algorithm, and Markov Chain Monte Carlo) to perform equivalent program transformation. Furthermore, we propose a deep reinforcement learning-based sequence generation (DRLSG) strategy to effectively guide the search process of generating clones that could escape from the detection. We then evaluate the ML-based detectors with the pairs of original and generated clones. We realize our method in a framework named CloneGen. CloneGen In evaluation, we challenge the two state-of-the-art ML-based detectors and four traditional detectors with the code clones after semantic-preserving transformations via the aid of CloneGen. Surprisingly, our experiments show that, despite the notable successes achieved by existing clone detectors, the ML models inside these detectors still cannot distinguish numerous clones produced by the code transformations in CloneGen. In addition, adversarial training of ML-based clone detectors using clones generated by CloneGen can improve their robustness and accuracy. CloneGen Meanwhile, compared with the commonly-used heuristics, the DRLSG strategy has shown the best effectiveness in generating code clones to decrease the detection accuracy of the ML-based detectors.

cs.SE↗

arXiv2021

Machine Learning of Partial Differential Equations from Noise Data

Machine learning of partial differential equations from data is a potential breakthrough to solve the lack of physical equations in complex dynamic systems, but because numerical differentiation is ill-posed to noise data, noise has become the biggest obstacle in the application of partial differential equation identification method. To overcome this problem, we propose Frequency Domain Identification method based on Fourier transforms, which effectively eliminates the influence of noise by using the low frequency component of frequency domain data to identify partial differential equations in frequency domain. We also propose a new sparse identification criterion, which can accurately identify the terms in the equation from low signal-to-noise ratio data. Through identifying a variety of canonical equations spanning a number of scientific domains, the proposed method is proved to have high accuracy and robustness for equation structure and parameters identification for low signal-to-noise ratio data. The method provides a promising technique to discover potential partial differential equations from noisy experimental data.

eess.SP↗

arXiv2021

Analysis on numerical stability and convergence of RANS turbulence models from the perspective of coupling modes

Reynolds-averaged Navier-Stokes simulations are still the main method to study complex flows in engineering. However, traditional turbulence models cannot accurately predict flow fields with separations. In such situation, machine learning methods provide an effective way to build new data-driven turbulence closure models. Nevertheless, a bottleneck that the data-driven turbulence models encounter is how to ensure the stability and convergence of the RANS equations in posterior iterations. This paper studies the effects of different coupling modes on the convergence and stability between the RANS equations and turbulence models. Numerical results demonstrate that the frozen coupling mode, commonly used in machine learning turbulence models, may lead to divergence and instability in posterior iterations; while the mutual coupling mode can maintain good convergence and stability in the process of iterations. This research can provide a new perspective to the coupling mode for machine learning turbulence models with RANS equations in posterior iterations.

physics.flu-dyn↗

arXiv2021

Deep Learning for Multi-Fidelity Aerodynamic Distribution Modeling from Experimental and Simulation Data

The wind-tunnel experiment plays a critical role in the design and development phases of modern aircraft, which is limited by prohibitive cost. In contrast, numerical simulation, as an important alternative paradigm, mimics complex flow behaviors but is less accurate compared to experiment. This leads to the recent development and emerging interest in applying data fusion for aerodynamic prediction. In particular, the accurate prediction of aerodynamic with lower computational cost can be achieved by fusing experimental (high-fidelity) and computational (low-fidelity) aerodynamic data. Currently, existing works on aerodynamic model using data fusion mainly concern integral data (lift, drag, etc.). In this paper, a multi-fidelity aerodynamic model based on Deep Neural Network (DNN), where both numerical and experimental data are introduced in the loss function with proper weighting factor to balance the overall accuracy, is developed for aerodynamic distribution over the wing surface. Specially, the proposed approach is illustrated by modeling the surface pressure distribution of ONERA M6 wing in transonic flow, including different secnarios where the flow condition varies or there is less high-fidelity data. The results demonstrate that the proposed approach with decent number of low-fidelity data and a few high-fidelity data can accurately predict the surface pressure distribution on transonic wing. The outperformance of the proposed model over other DNNs from only high fidelity or low-fidelity, has been reported.

physics.flu-dyn↗

arXiv2021

Experimental study on Vortex-Induced Vibration of Cylinder at Subcritical Reynolds Number

Numerical simulation results in recent years show that vortex-induced vibration (VIV) can occur at a subcritical Reynolds number. And the VIV has been observed numerically at Reynolds numbers as low as Re = 20. The current study presents an experimental evidence for the subcritical VIV of a cylinder. We designed and built a rotating channel that makes it possible to perform VIV experiments at subcritical Reynolds numbers. Based on the rotating channel, two sets of tests were carried out for fixed natural frequency with variable incoming flow speed and fixed incoming flow speed with variable natural frequency. In both sets of experiments, subcritical VIV were observed and the VIV can be observed at a Reynolds number as low as 23, which is close to the numerical results of Mittal.

physics.flu-dyn↗

arXiv2021

Robust exploratory mean-variance problem with drift uncertainty

We solve a min-max problem in a robust exploratory mean-variance problem with drift uncertainty in this paper. It is verified that robust investors choose the Sharpe ratio with minimal $L^2$ norm in an admissible set. A reinforcement learning framework in the mean-variance problem provides an exploration-exploitation trade-off mechanism; if we additionally consider model uncertainty, the robust strategy essentially weights more on exploitation rather than exploration and thus reflects a more conservative optimization scheme. Finally, we use financial data to backtest the performance of the robust exploratory investment and find that the robust strategy can outperform the purely exploratory strategy and resist the downside risk in a bear market.

math.OC↗

arXiv2021

UCNN: A Convolutional Strategy on Unstructured Mesh

In machine learning for fluid mechanics, fully-connected neural network (FNN) only uses the local features for modelling, while the convolutional neural network (CNN) cannot be applied to data on structured/unstructured mesh. In order to overcome the limitations of FNN and CNN, the unstructured convolutional neural network (UCNN) is proposed, which aggregates and effectively exploits the features of neighbour nodes through the weight function. Adjoint vector modelling is taken as the task to study the performance of UCNN. The mapping function from flow-field features to adjoint vector is constructed through efficient parallel implementation on GPU. The modelling capability of UCNN is compared with that of FNN on validation set and in aerodynamic shape optimization at test case. The influence of mesh changing on the modelling capability of UCNN is further studied. The results indicate that UCNN is more accurate in modelling process.

physics.flu-dyn↗

arXiv2020

Machine learning for adjoint vector in aerodynamic shape optimization

Adjoint method is widely used in aerodynamic design because only once solution of flow field is required for adjoint method to obtain the gradients of all design variables. However, the calculation cost of adjoint vector is approximately equal to that of flow computation. In order to accelerate the solution of adjoint vector and improve the adjoint-based optimization efficiency, machine learning for adjoint vector modeling is presented. Deep neural network (DNN) is employed to construct the mapping between the adjoint vector and the local flow variables. DNN can efficiently predict adjoint vector and its generalization is examined by a transonic drag reduction about NACA0012 airfoil. The results indicate that with negligible calculation cost of the adjoint vector, the proposed DNN-based adjoint method can achieve the same optimization results as the traditional adjoint method.

physics.flu-dyn↗

arXiv2020

Encoding Syntactic Knowledge in Transformer Encoder for Intent Detection and Slot Filling

We propose a novel Transformer encoder-based architecture with syntactical knowledge encoded for intent detection and slot filling. Specifically, we encode syntactic knowledge into the Transformer encoder by jointly training it to predict syntactic parse ancestors and part-of-speech of each token via multi-task learning. Our model is based on self-attention and feed-forward layers and does not require external syntactic information to be available at inference time. Experiments show that on two benchmark datasets, our models with only two Transformer encoder layers achieve state-of-the-art results. Compared to the previously best performed model without pre-training, our models achieve absolute F1 score and accuracy improvement of 1.59% and 0.85% for slot filling and intent detection on the SNIPS dataset, respectively. Our models also achieve absolute F1 score and accuracy improvement of 0.1% and 0.34% for slot filling and intent detection on the ATIS dataset, respectively, over the previously best performed model. Furthermore, the visualization of the self-attention weights illustrates the benefits of incorporating syntactic information during training.

cs.AI↗

arXiv2020

CovidNet: To Bring Data Transparency in the Era of COVID-19

Timely, creditable, and fine-granular case information is vital for local communities and individual citizens to make rational and data-driven responses to the COVID-19 pandemic. This paper presents CovidNet, a COVID-19 tracking project associated with a large scale epidemic dataset, which was initiated by 1Point3Acres. To the best of our knowledge, the project is the only platform providing real-time global case information of more than 4,124 sub-divisions from over 27 countries worldwide with multi-language supports. The platform also offers interactive visualization tools to analyze the full historical case curves in each region. Initially launched as a voluntary project to bridge the data transparency gap in North America in January 2020, this project by far has become one of the major independent sources worldwide and has been consumed by many other tracking platforms. The accuracy and freshness of the dataset is a result of the painstaking efforts from our voluntary teamwork, crowd-sourcing channels, and automated data pipelines. As of May 18, 2020, the project website has been visited more than 200 million times and the CovidNet dataset has empowered over 522 institutions and organizations worldwide in policy-making and academic researches. All datasets are openly accessible for non-commercial purposes at https://coronavirus.1point3acres.com via a formal request through our APIs.

cs.CY↗

arXiv2020

A new data assimilation method of recovering turbulent flow field at high-Reynolds numbers for turbulence machine learning

This paper proposes a new data assimilation method for recovering high fidelity turbulent flow field around airfoil at high Reynolds numbers based on experimental data, which is called Proper Orthogonal Decomposition Inversion (POD-Inversion) data assimilation method. Aiming at the flows including shock wave discontinuities or separated flows at high angle of attack, the proposed method can reconstruct high-fidelity turbulent flow field combining with experimental distributed force coefficients. We firstly perform the POD analysis to the turbulent eddy viscosity fields computed by SA model and obtain the base POD modes. Then optimized the POD coefficients by global optimization algorithm coupling with the Navier-Stokes equations solver. The high-fidelity turbulent flied are recovered by several main modes, which can dramatically reduce the dimensions of the system. The effectiveness of the method is verified by the cases of transonic flow around the RAE2822 airfoil at high Reynolds numbers and the separated flow at high angles of attack. The results demonstrate that the proposed assimilation method can recover the turbulent flow field which optimally match the experimental data, and significantly reduce the error of pressure coefficients. The proposed data assimilation method can offer high-fidelity field data for turbulent model based on machine learning.

physics.flu-dyn↗

arXiv2020

Transonic aeroelasticity: a new perspective from the fluid mode

Within the transonic regime, the aeroelastic problems exhibit many unique characteristics compared with subsonic and supersonic cases. Although a lot of research has been carried out in this field, the underlying mechanisms of these complex phenomena are not clearly understood yet, resulting in a challenge in the design and use of modern aircraft. This review summarizes the recent investigations on nonclassical transonic aeroelastic problems, including transonic buzz, reduction of transonic buffet onset, transonic buffeting response and frequency lock-in phenomenon in transonic buffet flow. After introducing the research methods in unsteady aerodynamics and aeroelastic problems, the dynamical characteristics as well as the physical mechanisms of these phenomena are discussed from the perspective of the fluid mode. In the framework of the ROM (reduced order model) -based model, the dominant fluid mode (or the eigenvalue) and its coupling process with the structural model can be clearly captured. The flow nonlinearity was believed to be the cause of the complexity of transonic aeroelasticity. In fact, this review indicates that the complexity lies in the decrease of the flow stability in the transonic regime. In this condition, the fluid mode becomes a principal part of the coupling process, which results in the instability of the fluid mode itself or the structural mode, and thus, it is the root cause of different transonic aeroelastic phenomena.

physics.flu-dyn↗

arXiv2019

Layer Pruning for Accelerating Very Deep Neural Networks

In this paper, we propose an adaptive pruning method. This method can cut off the channel and layer adaptively. The proportion of the layer and the channel to be cut is learned adaptively. The pruning method proposed in this paper can reduce half of the parameters, and the accuracy will not decrease or even be higher than baseline.

cs.LG↗

arXiv2019

Laboratory formation and photo-chemistry of fullerene/anthracene cluster cations

Besides buckminsterfullerene (C60), other fullerenes and their derivatives may also reside in space. In this work, we study the formation and photo-dissociation processes of astronomically relevant fullerene/anthracene (C14H10) cluster cations in the gas phase. Experiments are carried out using a quadrupole ion trap (QIT) in combination with time-of-flight (TOF) mass spectrometry. The results show that fullerene (C60, and C70)/anthracene (i.e., [(C14H10)nC60]+ and [(C14H10)nC70]+), fullerene (C56 and C58)/anthracene (i.e., [(C14H10)nC56]+ and [(C14H10)nC58]+) and fullerene (C66 and C68)/anthracene (i.e., [(C14H10)nC66]+ and [(C14H10)nC68]+) cluster cations, are formed in the gas phase through an ion-molecule reaction pathway. With irradiation, all the fullerene/anthracene cluster cations dissociate into mono$-$anthracene and fullerene species without dehydrogenation. The structure of newly formed fullerene/anthracene cluster cations and the bonding energy for these reaction pathways are investigated with quantum chemistry calculations. Our results provide a growth route towards large fullerene derivatives in a bottom-up process and insight in their photo-evolution behavior in the ISM, and clearly, when conditions are favorable, fullerene/PAH clusters can form efficiently. In addition, these clusters (from 80 to 154 atoms or ~ 2 nm in size) offer a good model for understanding the physical-chemical processes involved in the formation and evolution of carbon dust grains in space, and provide candidates of interest for the DIBs that could motivate spectroscopic studies.

physics.atm-clus↗

arXiv2019

High Dielectric Ternary Oxides from Crystal Structure Prediction and High-throughput Screening

The development of new high dielectric materials is essential for advancement in modern electronics. Oxides are generally regarded as the most promising class of high dielectric materials for industrial applications as they possess both high dielectric constants and large band gaps. Most previous researches on high dielectrics were limited to already known materials. In this study, we conducted an extensive search for high dielectrics over a set of ternary oxides by combining crystal structure prediction and density functional perturbation theory calculations. From this search, we adopted multiple stage screening to identify 440 new low-energy high dielectric materials. Among these materials, 33 were identified as potential high dielectrics favorable for modern device applications. Our research has opened an avenue to explore novel high dielectric materials by combining crystal structure prediction and high throughput screening.

cond-mat.mtrl-sci↗