arXiv subjects

Yao Wei

Publications and source records attributed to Yao Wei.

24 records · Page 2Linked to original sources

arXiv2022

Flow-based GAN for 3D Point Cloud Generation from a Single Image

Generating a 3D point cloud from a single 2D image is of great importance for 3D scene understanding applications. To reconstruct the whole 3D shape of the object shown in the image, the existing deep learning based approaches use either explicit or implicit generative modeling of point clouds, which, however, suffer from limited quality. In this work, we aim to alleviate this issue by introducing a hybrid explicit-implicit generative modeling scheme, which inherits the flow-based explicit generative models for sampling point clouds with arbitrary resolutions while improving the detailed 3D structures of point clouds by leveraging the implicit generative adversarial networks (GANs). We evaluate on the large-scale synthetic dataset ShapeNet, with the experimental results demonstrating the superior performance of the proposed method. In addition, the generalization ability of our method is demonstrated by performing on cross-category synthetic images as well as by testing on real images from PASCAL3D+ dataset.

cs.CV↗

arXiv2022

Exploring the Effect of the Number of Hydrogen Atoms on the Properties of Lanthanide Hydrides by DMFT

Lanthanide hydrogen-rich materials have long been considered as one of the candidates with high-temperature superconducting properties in condensed matter physics, and have attracted great interest. Attempts to investigate the effects of different compositions of lanthanide hydrogen-rich materials are ongoing, with predictions and experimental studies in recent years having shown that substances such as LaH 10 , CeH 9 , and LaH 16 exhibit extremely high superconducting temperatures between 150-250 GPa. In particular, researchers have noted that in those materials an increase in the f character at the Fermi level leads to an increase in the superconducting temperature. Here, we further elaborate on the effect of the ratios of lanthanide to hydrogen in these substances with the aim to bring more clarity to the study of superhydrides in these extreme cases by comparing a variety of lanthanide hydrogen-rich materials with different ratios using the DMFT method, and provide ideas for later structural predictions and material property studies.

cond-mat.supr-con↗

arXiv2022

High temperature superconductivity in the lanthanide hydrides at extreme pressures

Hydrogen-rich superhydrides are promising high-Tc superconductors, with superconductivity experimentally observed near room temperature, as shown in recently discovered lanthanide superhydrides at very high pressures, e.g. LaH10 at 170 GPa and CeH9 at 150 GPa. Superconductivity is believed to be closely related with the high vibrational modes of the bound hydrogen ions. Here we study the limit of extreme pressures (above 200 GPa) where lanthanide hydrides with large hydrogen content have been reported. We focus on LaH16 and CeH16, two prototype candidates for achieving a large electronic contribution from hydrogen in the electron-phonon coupling. In this work, we propose a first principles calculation platform with the inclusion of many-body corrections to evaluate the detailed physical properties of the Ce-H and La-H systems and to understand the structure, stability and superconductivity of these systems at ultra-high pressure. We provide a practical approach to further investigate conventional superconductivity in hydrogen rich superhydrides. We report that density functional theory provides accurate structure and phonon frequencies, but many-body corrections lead to an increase of the critical temperature, that is associated with spectral weight transfer of the f-states.

cond-mat.supr-con↗

arXiv2021

Computational Materials Discovery for Lanthanide Hydrides at high pressure: predicting High Temperature superconductivity

Hydrogen-rich superhydrides are believed to be very promising high-T$_c$ superconductors, with experimentally observed critical temperatures near room temperature, as shown in recently discovered lanthanide superhydrides at very high pressures, e.g. LaH$_{10}$ at 170 GPa and CeH$_9$ at 150 GPa. With the motivation of discovering new hydrogen-rich high-T$_c$ superconductors at lowest possible pressure, quantitative theoretical predictions are needed. In these promising compounds, superconductivity is mediated by the highly energetic lattice vibrations associated with hydrogen and their interplay with the electronic structure, requiring fine descriptions of the electronic properties, notoriously challenging for correlated $f$ systems. In this work, we propose a first-principles calculation platform with the inclusion of many-body corrections to evaluate the detailed physical properties of the Ce-H system and to understand the structure, stability and superconductivity of CeH$_9$ at high pressure. We report how the prediction of T$_c$ is affected by the hierarchy of many-body corrections, and obtain a compelling increase of T$_c$ at the highest level of theory, which goes in the direction of experimental observations. Our findings shed a significant light on the search for superhydrides in close similarity with atomic hydrogen within a feasible pressure range. We provide a practical platform to further investigate and understand conventional superconductivity in hydrogen rich superhydrides.

cond-mat.supr-con↗

arXiv2020

Scribble-based Weakly Supervised Deep Learning for Road Surface Extraction from Remote Sensing Images

Road surface extraction from remote sensing images using deep learning methods has achieved good performance, while most of the existing methods are based on fully supervised learning, which requires a large amount of training data with laborious per-pixel annotation. In this paper, we propose a scribble-based weakly supervised road surface extraction method named ScRoadExtractor, which learns from easily accessible scribbles such as centerlines instead of densely annotated road surface ground-truths. To propagate semantic information from sparse scribbles to unlabeled pixels, we introduce a road label propagation algorithm which considers both the buffer-based properties of road networks and the color and spatial information of super-pixels. The proposal masks generated from the road label propagation algorithm are utilized to train a dual-branch encoder-decoder network we designed, which consists of a semantic segmentation branch and an auxiliary boundary detection branch. We perform experiments on three diverse road datasets that are comprised of highresolution remote sensing satellite and aerial images across the world. The results demonstrate that ScRoadExtractor exceed the classic scribble-supervised segmentation method by 20% for the intersection over union (IoU) indicator and outperform the state-of-the-art scribble-based weakly supervised methods at least 4%.

cs.CV↗

arXiv2006

Grid-independent Issue in Numerical Heat Transfer

Grid independent is associated with the accuracy or even rationality of numerical results. This paper takes two-dimensional steady heat transfer for example to reveal the effect of grid resolution on numerical results. The law of grid dependence is obtained and a simple mathematical formula is presented. The production acquired here can be used as the guidance in choosing grid density in numerical simulation and get exact grid independent value without using infinite fine grid. Through analyzing grid independent, we can find the minimum number of grid cells that is needed to get grid-independent results. Such strategy can save computational resource while ensure a rational computational result.

math-ph↗