Searcharxiv⌕ Search

arXiv subjects

Weiwei Zhang

Publications and source records attributed to Weiwei Zhang.

At least 109 records · Page 6Linked to original sources
arXiv2019

Recurrent Aggregation Learning for Multi-View Echocardiographic Sequences Segmentation

Multi-view echocardiographic sequences segmentation is crucial for clinical diagnosis. However, this task is challenging due to limited labeled data, huge noise, and large gaps across views. Here we propose a recurrent aggregation learning method to tackle this challenging task. By pyramid ConvBlocks, multi-level and multi-scale features are extracted efficiently. Hierarchical ConvLSTMs next fuse these features and capture spatial-temporal information in multi-level and multi-scale space. We further introduce a double-branch aggregation mechanism for segmentation and classification which are mutually promoted by deep aggregation of multi-level and multi-scale features. The segmentation branch provides information to guide the classification while the classification branch affords multi-view regularization to refine segmentations and further lessen gaps across views. Our method is built as an end-to-end framework for segmentation and classification. Adequate experiments on our multi-view dataset (9000 labeled images) and the CAMUS dataset (1800 labeled images) corroborate that our method achieves not only superior segmentation and classification accuracy but also prominent temporal stability.

eess.IV↗
arXiv2019

Super Poincar'e inequality for a dynamic model for the two-parameter Dirichlet process

In this paper, we establish the super Poincaré inequality for the two-parameter Dirichlet process when the partition number of the state space is finite. Furthermore, if the partition number is infinite, the super Poincar'e inequality doesn't hold. To overcome the difficulty caused by the degenerency of the diffusion coefficient on the boundary of the domain, localization method and perturbation argument in [14] are effective.

math.PR↗
arXiv2019

Comparisons of Two Reduced-Order Models for Linearized Unsteady Aerodynamic Identification

This paper compares the performance of two unsteady aerodynamic reduced-order models (ROMs), namely linear Volterra series and the autoregressive with exogenous input (ARX) model, on modeling dynamically linear aerodynamic behaviors. The difference between these two methods is that the latter model has an autoregressive term while the former model has only the input-related term. The first system is a plunging cylinder in a low-Reynolds number flow, where the flow stable (Re < 47). Although the training data can be fitted well with both methods, the linear Volterra method requires a higher model order than the ARX model for the same accuracy. Comparison of the frequency response indicates that the ARX model approximates the frequency response more closely, while the frequency response at high Reynolds number is over-fitted by Volterra series. The second aerodynamic system is a flow over a pitching NACA0012 airfoil, including subsonic and transonic states. The convergence of the model with respect to delay orders, at different Mach numbers and mean angles of attack, is studied in detail. As the Mach number or the mean angle of attack increases, the required delay order will increase. But the ARX model still models this system with a small number of terms at the same level of accuracy. All results indicate that the ARX model outperforms the linear Volterra series in most of cases, especially when the flow is close to the unstable state.

physics.comp-ph↗
arXiv2019

Provably size-guaranteed mesh generation with superconvergence

The properties and applications of superconvergence on size-guaranteed Delaunay triangulation generated by bubble placement method (BPM), are studied in this paper. First, we derive a mesh condition that the difference between the actual side length and the desired length $h$ is as small as ${\cal O}(h^{1+α})$ $(α>0)$. Second, the superconvergence estimations are analyzed on linear and quadratic finite element for elliptic boundary value problem based on the above mesh condition. In particular, the mesh condition is suitable for many known superconvergence estimations of different equations. Numerical tests are provided to verify the theoretical findings and to exhibit the superconvergence property on BPM-based grids.

math.NA↗
arXiv2019

Laboratory photo-chemistry of covalently bonded fluorene clusters: observation of an interesting PAH bowl-forming mechanism

The fullerene C$_{60}$, one of the largest molecules identified in the interstellar medium (ISM), has been proposed to form top-down through the photo-chemical processing of large (more than 60 C-atoms) polycyclic aromatic hydrocarbon (PAH) molecules. In this article, we focus on the opposite process, investigating the possibility that fullerenes form from small PAHs, in which bowl-forming plays a central role. We combine laboratory experiments and quantum chemical calculations to study the formation of larger PAHs from charged fluorene clusters. The experiments show that with visible laser irradiation, the fluorene dimer cation - [C$_{13}$H$_{9}$$-$C$_{13}$H$_{9}$]$^+$ - and the fluorene trimer cation - [C$_{13}$H$_{9}$$-$C$_{13}$H$_{8}$$-$C$_{13}$H$_{9}$]$^+$ - undergo photo-dehydrogenation and photo-isomerization resulting in bowl structured aromatic cluster-ions, C$_{26}$H$_{12}$$^+$ and C$_{39}$H$_{20}$$^+$, respectively. To study the details of this chemical process, we employ quantum chemistry that allows us to determine the structures of the newly formed cluster-ions, to calculate the hydrogen loss dissociation energies, and to derive the underlying reaction pathways. These results demonstrate that smaller PAH clusters (with less than 60 C-atoms) can convert to larger bowled geometries that might act as building blocks for fullerenes, as the bowl-forming mechanism greatly facilitates the conversion from dehydrogenated PAHs to cages. Moreover, the bowl-forming induces a permanent dipole moment that - in principle - allows to search for such species using radio astronomy.

astro-ph.IM↗
arXiv2018

Machine learning methods for turbulence modeling in subsonic flows over airfoils

Reynolds-Averaged Navier-Stokes(RANS) method will still play a vital role in the following several decade in aerospace engineering. Although RANS models are widely used, empiricism and large discrepancies between models reduce the reliability of simulating complex flows. Therefore, in recent years, data-driven turbulence model has aroused widespread concern in fluid mechanics. Based on the experimental/numerical simulation results, this approach aims to modify or construct the turbulence model for specific purposes by machine learning technologies. In this paper, we take the results calculated by SA model as training data. Different from low Reynolds number turbulent flows, the data from high Reynolds number flows shows an apparent scaling effect, thus leading to difficulties in the data-driven modeling. In order to improve the fitting accuracy, we divided the flow field into near-wall region, wake region, and far-field region, and built individual model for every region. In this paper, we adopted the radial basis function neural network (RBFNN) and some auxiliary optimization algorithms to reconstruct a mapping function between mean variables and the eddy viscosity. Since this model reflects the relationship between local flow characteristics and turbulent eddy viscosity, it is independent on the airfoil shape and flow condition. The training data in this paper is generated from only three subsonic flow calculations of NACA0012 airfoil. By coupling the proposed approach with N-S equations, we calculated various flow cases as well as two different airfoils and showed the eddy viscosity contours, velocity profiles along the normal direction of wall and skin friction coefficient distributions, etc. Compared with the SA model, the results show a reasonable accuracy and better efficiency, which indicates the positive prospect of data-driven methods in turbulence modeling.

physics.flu-dyn↗
arXiv2018

A novel unsteady aerodynamic Reduced-Order Modeling method for transonic aeroelastic optimization

In aircraft design, structural optimization and uncertainty quantification concerning transonic aeroelastic issues are computationally impractical, because the iterative process requires great number of aeroelastic analysis. Emerging Reduced-Order Model (ROM) method is convenient for transonic aeroelastic analysis. However, current ROMs cannot be reused during iteration, thus time cost is still way too large. This study proposed an improved ROM suitable for Arbitrary Mode Shapes (ROM-AMS), which is reusable regardless the variation of design variables. By adopting Principal Component Analysis, ROM-AMS method can significantly reduce the number of basis mode shapes and improve the accuracy of flutter analysis. In an optimization case, the weight of cropped delta wing is reduced by 28.46%, and the efficiency is 900 times higher than traditional ROM method, which demonstrates the feasibility of this method in iterative design process.

physics.flu-dyn↗
arXiv2018

Mode Multigrid - A novel convergence acceleration method

This paper proposes a mode multigrid (MMG) method, and applies it to accelerate the convergence of the steady state flow on unstructured grids. The dynamic mode decomposition (DMD) technique is used to analyze the convergence process of steady flow field according to the solution vectors from the previous time steps. Unlike the traditional multigrid method, we project the flowfield solutions from the physical space into the modal space, and truncate all the high-frequency modes but only the first-order mode are retained based on the DMD analysis. The real solutions in the physical space can be obtained simply by the inverse transformation from the modal space. The developed MMG method ingeniously avoids the complicated process of coarsening computational mesh, and does not need to make any change for the grid in physical space. Therefore, it is very convenient to be applied to any numerical schemes with just little change for the flow solver, which is also suitable for unstructured grids and easy for parallel computing. Several typical test cases have been used to verify the effectiveness of the proposed method, which demonstrates that the MMG can dramatically reduce the number of iterative steps for the different mesh types, different accuracy of spatial discretization and different time-marching schemes. The method is 3 to 6 times faster than the original method while ensuring the computational accuracy.

physics.comp-ph↗
arXiv2018

Nash inequality for Diffusion Processes Associated with Dirichlet Distributions

For any $N\ge 2$ and $α=(α_1,\cdots, α_{N+1})\in (0,\infty)^{N+1}$, let $μ^{(N)}_α$ be the Dirichlet distribution with parameter $α$ on the set $Δ^{ (N)}:= \{ x \in [0,1]^N:\ \sum_{1\le i\le N}x_i \le 1 \}.$ The multivariate Dirichlet diffusion is associated with the Dirichlet form $${\scr E}_α^{(N)}(f,f):= \sum_{n=1}^N \int_{ Δ^{(N)}} \bigg(1-\sum_{1\le i\le N}x_i\bigg) x_n(\partial_n f)^2(x)\,μ^{(N)}_α(d x)$$ with Domain ${\scr D}({\scr E}_α^{(N)})$ being the closure of $C^1(Δ^{(N)})$. We prove the Nash inequality $$μ_α^{(N)}(f^2)\le C {\scr E}_α^{(N)}(f,f)^{\frac p{p+1} }μ_α^{(N)} (|f|)^{\frac 2 {p+1}},\ \ f\in {\scr D}({\scr E}_α^{(N)}), μ_α^{(N)}(f)=0$$ for some constant $C>0$ and $p= (α_{N+1}-1)^+ +\sum_{i=1}^N 1\lor (2α_i),$ where the constant $p$ is sharp when $\max_{1\le i\le N} α_i \le 1/2$ and $α_{N+1}\ge 1$. This Nash inequality also holds for the corresponding Fleming-Viot process.

math.PR↗
arXiv2017

Optical pump-rejection filter based on silicon sub-wavelength engineered photonic structures

The high index contrast of the silicon-on-insulator (SOI) platform allows the realization of ultra-compact photonic circuits. However, this high contrast hinders the implementation of narrow-band Bragg filters. These typically require corrugations widths of a few nanometers or double-etch geometries, hampering device fabrication. Here we report, for the first time, on the realization of SOI Bragg filters based on sub-wavelength index engineering in a differential corrugation width configuration. The proposed double periodicity structure allows narrow-band rejection with a single etch step and relaxed width constraints. Based on this concept, we experimentally demonstrate a single-etch, $\mathbf{220\,nm}$ thick, Si Bragg filter featuring a corrugation width of $\mathbf{150\,nm}$, a rejection bandwidth of $\mathbf{1.1\,nm}$ and an extinction ratio exceeding $\mathbf{40\,dB}$. This represents a ten-fold width increase compared to conventional single-periodicity, single-etch counterparts with similar bandwidths.

physics.optics↗
arXiv2017

High-quality photonic entanglement for wavelength-multiplexed quantum communication based on a silicon chip

We report an efficient energy-time entangled photon-pair source based on four-wave mixing in a CMOS-compatible silicon photonics ring resonator. Thanks to suitable optimization, the source shows a large spectral brightness of 400\,pairs of entangled photons /s/MHz for $\rm 500\,μW$ pump power. Additionally, the resonator has been engineered so as to generate a frequency comb structure compatible with standard telecom dense wavelength division multiplexers. We demonstrate high-purity energy-time entanglement, i.e., free of photonic noise, with near perfect raw visibilities ($>$~98\%) between various channel pairs in the telecom C-band. Such a compact source stands as a path towards more complex quantum photonic circuits dedicated to quantum communication systems.

quant-ph↗
arXiv2015

Aqueous Proton Transfer Across Single Layer Graphene

Proton transfer across single layer graphene is associated with large computed energy barriers and is therefore thought to be unfavorable at room temperature unless nanoscale holes or dopants are introduced, or a potential bias is applied. Here, we subject single layer graphene supported on fused silica to cycles of high and low pH and show that protons transfer reversibly from the aqueous phase through the graphene to the other side where they undergo acid-base chemistry with the silica hydroxyl groups. After ruling out diffusion through macroscopic pinholes, the protons are found to transfer through rare, naturally occurring atomic defects. Computer simulations reveal low energy barriers of 0.68 to 0.75 eV for aqueous proton transfer across hydroxyl-terminated atomic defects that participate in a Grotthuss-type relay, while pyrylium-like ether terminations shut down proton exchange. Unfavorable energy barriers to helium and hydrogen transfer indicate the transfer process is selective for aqueous protons.

cond-mat.mtrl-sci↗
arXiv2014

Stability of numerous novel potassium chlorides at high pressure

K-Cl is a simple system displaying all four main types of bonding, as it contains (i) metallic potassium, (ii) elemental chlorine made of covalently bonded Cl2 molecules held together by van der Waals forces, and (iii) an archetypal ionic compound KCl. Well-known octet rule (and its special case, the charge balance rule, assigning classical charges of +1 to K and -1 to Cl) predicts that no compounds other than KCl are possible. However, our quantum-mechanical variable-composition evolutionary simulations predict new stable compounds. The phase diagram of the K-Cl system turned out to be extremely complicated, featuring new thermodynamically compounds K3Cl, K2Cl, K3Cl2, K4Cl3, K5Cl4, K3Cl5, KCl3, KCl7. Of particular interest are 2D-metallic homologs Kn+1Cln, the presence of positively charged Cl atoms in KCl7, and the predicted stability of KCl3 already at the atmospheric pressure.

cond-mat.mtrl-sci↗
arXiv2013

Unexpected stable stoichiometries of sodium chlorides

Sodium chloride (NaCl), or rocksalt, is well characterized at ambient pressure. Due to the large electronegativity difference between Na and Cl atoms, it has highly ionic chemical bonding, with stoichiometry 1:1 dictated by charge balance, and B1-type crystal structure. Here, by combining theoretical predictions and diamond anvil cell experiments we show that new materials with different stoichiometries emerge at pressure as low as 20 GPa. Compounds such us Na3Cl, Na2Cl, Na3Cl2, NaCl3 and NaCl7 are theoretically stable and have unusual bonding and electronic properties. To test this prediction, at 55-80 GPa we synthesized cubic and orthorhombic NaCl3 at 55-70 GPa and 2D-metallic tetragonal Na3Cl. This proves that novel compounds, violating chemical intuition, can be thermodynamically stable even in simplest systems at non-ambient conditions.

cond-mat.mtrl-sci↗
arXiv2013

Predicting genome-wide DNA methylation using methylation marks, genomic position, and DNA regulatory elements

Background: Recent assays for individual-specific genome-wide DNA methylation profiles have enabled epigenome-wide association studies to identify specific CpG sites associated with a phenotype. Computational prediction of CpG site-specific methylation levels is important, but current approaches tackle average methylation within a genomic locus and are often limited to specific genomic regions. Results: We characterize genome-wide DNA methylation patterns, and show that correlation among CpG sites decays rapidly, making predictions solely based on neighboring sites challenging. We built a random forest classifier to predict CpG site methylation levels using as features neighboring CpG site methylation levels and genomic distance, and co-localization with coding regions, CGIs, and regulatory elements from the ENCODE project, among others. Our approach achieves 91% -- 94% prediction accuracy of genome-wide methylation levels at single CpG site precision. The accuracy increases to 98% when restricted to CpG sites within CGIs. Our classifier outperforms state-of-the-art methylation classifiers and identifies features that contribute to prediction accuracy: neighboring CpG site methylation status, CpG island status, co-localized DNase I hypersensitive sites, and specific transcription factor binding sites were found to be most predictive of methylation levels. Conclusions: Our observations of DNA methylation patterns led us to develop a classifier to predict site-specific methylation levels that achieves the best DNA methylation predictive accuracy to date. Furthermore, our method identified genomic features that interact with DNA methylation, elucidating mechanisms involved in DNA methylation modification and regulation, and linking different epigenetic processes.

q-bio.GN↗
arXiv2012

Unexpected stable stoichiometries of sodium chlorides

At ambient pressure, sodium, chlorine, and their only known compound NaCl, have well-understood crystal structures and chemical bonding. Sodium is a nearly-free-electron metal with the bcc structure. Chlorine is a molecular crystal, consisting of Cl2 molecules. Sodium chloride, due to the large electronegativity difference between Na and Cl atoms, has highly ionic chemical bonding, with stoichiometry 1:1 dictated by charge balance, and rocksalt (B1-type) crystal structure in accordance with Pauling's rules. Up to now, Na-Cl was thought to be an ultimately simple textbook system. Here, we show that under pressure the stability of compounds in the Na-Cl system changes and new materials with different stoichiometries emerge at pressure as low as 25 GPa. In addition to NaCl, our theoretical calculations predict the stability of Na3Cl, Na2Cl, Na3Cl2, NaCl3 and NaCl7 compounds with unusual bonding and electronic properties. The bandgap is closed for the majority of these materials. Guided by these predictions, we have synthesized cubic NaCl3 at 55-60 GPa in the laser-heated diamond anvil cell at temperatures above 2000 K.

cond-mat.mtrl-sci↗
arXiv2011

Exciton polarization, fine structure splitting and quantum dot asymmetry under uniaxial stress

We derive a general relation between the fine structure splitting (FSS) and the exciton polarization angle of self-assembled quantum dots (QDs) under uniaxial stress. We show that the FSS lower bound under external stress can be predicted by the exciton polarization angle and FSS under zero stress. The critical stress can also be determined by monitoring the change in exciton polarization angle. We confirm the theory by performing atomistic pseudopotential calculations for the InAs/GaAs QDs. The work provides a deep insight into the dots asymmetry and their optical properties, and a useful guide in selecting QDs with smallest FSS which are crucial in entangled photon sources applications.

cond-mat.mtrl-sci↗