arXiv subjects

Wei Yan

Publications and source records attributed to Wei Yan.

At least 91 records · Page 5Linked to original sources

arXiv2021

Risk Assessment for Performance-Driven Building Design with BIM-Based Parametric Methods

A growing demand for handling uncertainties and risks in performance-driven building design decision-making has challenged conventional design methods. Thus, researchers in this field lean towards viable alternatives to using deterministic design methods, e.g., probabilistic methods. This research addresses the challenges associated with conventional methods of performance-driven building design, i.e., ignoring the existing uncertainties and lacking a systematic framework to incorporate risk assessment in building performance analysis. This research introduces a framework (BIMProbE) to integrate BIM-based parametric tools with building probabilistic performance analysis to facilitate uncertainty analysis and risk assessment in performance-based building design decision-making. A hypothetical building design scenario is used to demonstrate the application of the proposed framework. The results show that the probabilistic method leads to a different performance ranking order than the deterministic method. Also, the probabilistic method allows evaluating design options based on different risk attitudes.

math.NA↗

arXiv2021

The Cauchy problem for the generalized KdV equation with rough data and random data

In this paper, we consider the Cauchy problem for the generalized KdV equation with rough data and random data. Firstly, we prove that $u(x,t)\longrightarrow u(x,0)$ as $t\longrightarrow0$ for a.e. $x\in \mathbb{R}$ with $u(x,0)\in H^{s}(\mathbb{R})(s>\frac{1}{2}-\frac{2}{k},k\geq8).$ Secondly, we prove that $u(x,t)\longrightarrow e^{-t\partial_{x}^{3}}u(x,0)$ as $t\longrightarrow0$ for a.e. $x\in \mathbb{R}$ with $u(x,0)\in H^{s}(\mathbb{R})(s>\frac{1}{2}-\frac{2}{k},k\geq8).$ Thirdly, we prove that $\lim\limits_{t\longrightarrow 0}\left\|u(x,t)-e^{-t\partial_{x}^{3}}u(x,0)\right\|_{L_{x}^{\infty}}=0$ with $u(x,0)\in H^{s}(\mathbb{R})(s>\frac{1}{2}-\frac{2}{k+1},k\geq5)$. Fourthly, by using Strichartz estimates, probabilistic Strichartz estimates, we establish the probabilistic well-posedness in $H^{s}(\mathbb{R})\left(s>{\rm max} \left\{\frac{1}{k+1}\left(\frac{1}{2}-\frac{2}{k}\right), \frac{1}{6}-\frac{2}{k}\right\}\right)$ with random data. Our result improves the result of Hwang, Kwak (Proc. Amer. Math. Soc. 146(2018), 267-280.). Fifthly, we prove that $\forall ε>0,$ $\forall ω\in Ω_{T},$ $\lim\limits_{t\longrightarrow0}\left\|u(x,t)-e^{-t\partial_{x}^{3}}u^ω(x,0)\right\|_{L_{x}^{\infty}}=0$ with $u(x,0)\in H^{s}(\mathbb{R})(s>\frac{1}{6},k\geq6),$ where ${\rm P}(Ω_{T})\geq 1- C_{1}{\rm exp} \left(-\frac{C}{T^{\fracε{48k}}\|u(x,0)\|_{H^{s}}^{2}}\right)$ and $u^ω(x,0)$ is the randomization of $u(x,0)$. Finally, we prove that $\forall ε>0,$ $\forall ω\in Ω_{T}, \lim\limits_{t\longrightarrow0}\left\|u(x,t)-u^ω(x,0)\right\|_{L_{x}^{\infty}}=0$ with ${\rm P}(Ω_{T})\geq 1- C_{1}{\rm exp} \left(-\frac{C}{T^{\fracε{48k}}\|u(x,0)\|_{H^{s}}^{2}}\right)$ and $u(x,0)\in H^{s}(\mathbb{R})(s>\frac{1}{6},k\geq6)$.

math.AP↗

arXiv2021

Convolutional versus Dense Neural Networks: Comparing the Two Neural Networks Performance in Predicting Building Operational Energy Use Based on the Building Shape

A building self-shading shape impacts substantially on the amount of direct sunlight received by the building and contributes significantly to building operational energy use, in addition to other major contributing variables, such as materials and window-to-wall ratios. Deep Learning has the potential to assist designers and engineers by efficiently predicting building energy performance. This paper assesses the applicability of two different neural networks structures, Dense Neural Network (DNN) and Convolutional Neural Network (CNN), for predicting building operational energy use with respect to building shape. The comparison between the two neural networks shows that the DNN model surpasses the CNN model in performance, simplicity, and computation time. However, image-based CNN has the benefit of utilizing architectural graphics that facilitates design communication.

cs.LG↗

arXiv2021

Probabilistic pointwise convergence problem of some dispersive equations

In this paper, we investigate the almost surely pointwise convergence problem of free KdV equation, free wave equation, free elliptic and non-elliptic Schrödinger equation respectively. We firstly establish some estimates related to the Wiener decomposition of frequency spaces which are just Lemmas 2.1-2.6 in this paper. Secondly, by using Lemmas 2.1-2.6, 3.1, we establish the probabilistic estimates of some random series which are just Lemmas 3.2-3.11 in this paper. Finally, combining the density theorem in L$^{2}$ with Lemmas 3.2-3.11, we obtain almost surely pointwise convergence of the solutions to corresponding equations with randomized initial data in $L^{2}$, which require much less regularity of the initial data than the rough data case. At the same time, we present the probabilistic density theorem, which is Lemma 3.11 in this paper.

math.AP↗

arXiv2021

Micro-scale opto-thermo-mechanical actuation in the dry adhesive regime

Realizing optical manipulation of microscopic objects is crucial in the research fields of life science, condensed matter physics and physical chemistry. In non-liquid environments, this task is commonly regarded as difficult due to strong adhesive surface force ($\simμ\rm N$) between solid interfaces that makes tiny optical driven force ($\sim\rm pN$) insignificant. Here, by recognizing the microscopic interaction mechanism between friction force -- the parallel component of surface force on the contact surface -- and thermoelastic waves induced by pulsed optical absorption, we establish a general principle enabling the actuation of micro-objects on dry frictional surfaces based on the opto-thermo-mechanical effects. Theoretically, we predict that nanosecond pulsed optical absorption with mW-scale peak power is sufficient to tame $μ\rm N$-scale friction force. Experimentally, we demonstrate that two-dimensional spiral motion of gold plates on micro-fibers driven by a nanosecond pulsed laser, and reveal the specific rules of motion control. Our results pave the way for future development of micro-scale actuators in nonliquid environments.

physics.app-ph↗

arXiv2021

Single-photon non-reciprocity with an integrated magneto-optical isolator

Non-reciprocal photonic devices are essential components of classical optical information processing. It is interesting and important to investigate their feasibility in the quantum world. In this work, the quantum properties of an on-chip silicon nitride (SiN)-based magneto-optical (MO) isolator were studied using a single-photon non-reciprocal dynamical transmission experiment. The measured isolation ratio for single photons achieved was 12.33 dB, which proved the functionality of our on-chip isolator. The quantum coherence of the passing single photons was further verified using high-visibility quantum interference. Our work will promote on-chip isolators within the integrated quantum circuits and help introduce novel phenomena in quantum information processes.

quant-ph↗

arXiv2021

Light-induced in-plane Rotation of Microobjects on Microfibers

The transfer of angular momentum carried by photons into a microobject has been widely exploited to achieve the actuation of the microobject. However, this scheme is fundamentally defective in nonliquid environments as a result of the scale gap between friction forces ($μ$N) and optical forces (pN). To bypass this challenge, the researchers have recently proposed to take advantage of elastic waves based on opto-thermo-mechanical effects [1-4]. Grounded on this insight, we here demonstrate and characterize the in-plane rotation of a gold nanoplate in its surface contacting with a microfiber, driven by nanosecond laser pulses, which has not been explored before. Furthermore, we examine the underlying physical mechanisms and highlight the essential role of the spatial gradient of optical absorption. The combined experimental and theoretical results offer new insights into the study of the light-induced actuation of the microobjects in nonliquid environments, an emerging field far from being mature in both comprehensive understanding and practical applications.

physics.optics↗

arXiv2021

Second harmonic generation from Chalcogenide metasurfaces via mode coupling engineering

Dielectric metasurfaces have shown prominent applications in nonlinear optics due to strong field enhancement and low dissipation losses at the nanoscale. Chalcogenide glasses are one of the promising materials for the observation of nonlinear effects due to their high intrinsic nonlinearities. Here, we demonstrate, experimentally and theoretically, that significant second harmonic generation can be obtained within amorphous chalcogenide based metasurfaces by relying on the coupling between lattice and particle resonances. We further show that the high quality factor resonance at the origin of the second harmonic generation can be tuned over a wide wavelength range using a simple and versatile fabrication approach. The measured second harmonic intensity is orders of magnitude higher than that from a deposited chalcogenide film, and more than three orders of magnitude higher than conventional plasmonic and Silicon-based structures. Fabricated via a simple and scalable technique, these all-dielectric architectures are ideal candidates for the design of flat non-linear optical components on flexible substrates.

physics.optics↗

arXiv2021

Application of Deep Learning in Generating Desired Design Options: Experiments Using Synthetic Training Dataset

Most design methods contain a forward framework, asking for primary specifications of a building to generate an output or assess its performance. However, architects urge for specific objectives though uncertain of the proper design parameters. Deep Learning (DL) algorithms provide an intelligent workflow in which the system can learn from sequential training experiments. This study applies a method using DL algorithms towards generating demanded design options. In this study, an object recognition problem is investigated to initially predict the label of unseen sample images based on training dataset consisting of different types of synthetic 2D shapes; later, a generative DL algorithm is applied to be trained and generate new shapes for given labels. In the next step, the algorithm is trained to generate a window/wall pattern for desired light/shadow performance based on the spatial daylight autonomy (sDA) metrics. The experiments show promising results both in predicting unseen sample shapes and generating new design options.

cs.CV↗

arXiv2021

Towards Learning Geometric Transformations through Play: An AR-powered approach

Despite the excessive developments of architectural parametric platforms, parametric design is often interpreted as an architectural style rather than a computational method. Also, the problem is still a lack of knowledge and skill about the technical application of parametric design in architectural modelling. Students often dive into utilizing complex digital modelling without having a competent pedagogical context to learn algorithmic thinking and the corresponding logic behind digital and parametric modelling. The insufficient skills and superficial knowledge often result in utilizing the modelling software through trial and error, not taking full advantage of what it has to offer. Geometric transformations as the fundamental functions of parametric modelling is explored in this study to anchor learning essential components in parametric modelling. Students need to understand the differences between variables, parameters, functions and their relations. Fologram, an Augmented Reality tool, is utilized in this study to learn geometric transformation and its components in an intuitive way. A LEGO set is used as an editable physical model to improve spatial skill through hand movement beside an instant feedback in the physical environment.

cs.HC↗

arXiv2021

BIM Hyperreality: Data Synthesis Using BIM and Hyperrealistic Rendering for Deep Learning

Deep learning is expected to offer new opportunities and a new paradigm for the field of architecture. One such opportunity is teaching neural networks to visually understand architectural elements from the built environment. However, the availability of large training datasets is one of the biggest limitations of neural networks. Also, the vast majority of training data for visual recognition tasks is annotated by humans. In order to resolve this bottleneck, we present a concept of a hybrid system using both building information modeling (BIM) and hyperrealistic (photorealistic) rendering to synthesize datasets for training a neural network for building object recognition in photos. For generating our training dataset BIMrAI, we used an existing BIM model and a corresponding photo-realistically rendered model of the same building. We created methods for using renderings to train a deep learning model, trained a generative adversarial network (GAN) model using these methods, and tested the output model on real-world photos. For the specific case study presented in this paper, our results show that a neural network trained with synthetic data; i.e., photorealistic renderings and BIM-based semantic labels, can be used to identify building objects from photos without using photos in the training data. Future work can enhance the presented methods using available BIM models and renderings for more generalized mapping and description of photographed built environments.

cs.LG↗

arXiv2021

Chandra Observations of Excess Fe K$α$ Line Emission in Galaxies with High Star Formation Rates: X-ray Reflection on Galaxy Scales?

In active galactic nuclei (AGN), fluorescent Fe K$α$ (iron) line emission is generally interpreted as originating from obscuring material around a supermassive black hole (SMBH) on the scale of a few parsecs (pc). However, recent Chandra studies indicate the existence of iron line emission extending to kpc scales in the host galaxy. The connection between iron line emission and large-scale material can be spatially resolved directly only in nearby galaxies, but could be inferred in more distant AGNs by a connection between line emission and star-forming gas and dust that is more extended than the pc-scale torus. Here we present the results from a stacking analysis and X-ray spectral fitting performed on sources in the Chandra Deep Field South (CDFS) 7 Ms observations. From the deep stacked spectra, we select sources with stellar mass $\log(M_*/M_\odot)>10$ at $0.5<z<2$, obtaining 25 sources with high infrared luminosity ($ {\rm SFR}_{\rm FIR} \geq 17\;M_{\odot}\;{\rm yr}^{-1}$) and 32 sources below this threshold. We find that the equivalent width of the iron line EW(Fe) is a factor of three higher with 3$σ$ significance for high infrared luminosity measured from Herschel observations, indicating a connection between iron line emission and star-forming material on galaxy scales. We show that there is no significant dependence in EW(Fe) on $M_*$ or X-ray luminosity, suggesting the reflection of AGN X-ray emission over large scales in their host galaxies may be widespread.

astro-ph.GA↗

arXiv2021

The Cauchy problem for the generalized Ostrovsky equation with negative dispersion

This paper is devoted to studying the Cauchy problem for the generalized Ostrovsky equation \begin{eqnarray*} u_{t}-β\partial_{x}^{3}u-γ\partial_{x}^{-1}u+\frac{1}{k+1}(u^{k+1})_{x}=0,k\geq5 \end{eqnarray*} with $βγ<0,γ>0$. Firstly, we prove that the Cauchy problem for the generalized Ostrovsky equation is locally well-posed in $H^{s}(\mathbb{R})\left(s>\frac{1}{2}-\frac{2}{k}\right)$. Then, we prove that the Cauchy problem for the generalized Ostrovsky equation is locally well-posed in $X_{s}(\mathbb{R}): =\|f\|_{H^{s}}+\left\|\mathscr{F}_{x}^{-1}\left(\frac{\mathscr{F}_{x} f(ξ)}ξ\right)\right\|_{H^{s}}\left(s>\frac{1}{2}-\frac{2}{k}\right).$ Finally, we show that the solution to the Cauchy problem for generalized Ostrovsky equation converges to the solution to the generalized KdV equation as the rotation parameter $γ$ tends to zero for data belonging to $X_{s}(\mathbb{R})(s>\frac{3}{2})$. The main difficulty is that the phase function of Ostrosvky equation with negative dispersive $βξ^{3}+\fracγξ$ possesses the zero singular point.

math.AP↗

arXiv2021

Pointwise convergence problem of Ostrovsky equation with rough data and random data

In this paper, we consider the pointwise convergence problem of free Ostrovsky equation with rough data and random data. Firstly, we show the almost everywhere pointwise convergence of free Ostrovsky equation in $H^{s}(\mathbb{R})$ with $s\geq \frac{1}{4}$ with rough data. Secondly, we present counterexamples showing that the maximal function estimate related to the free Ostrovsky equation can fail if $s<\frac{1}{4}$. Finally, for every $x\in \mathbb{R}$, we show the almost surely pointwise convergence of free Ostrovsky equation in $L^{2}(\mathbb{R})$ with random data. The main tools are the density theorem, high-low frequency idea, Wiener decomposition and Lemmas 2.1-2.6 as well as the probabilistic estimates of some random series which are just Lemmas 3.2-3.4 in this paper. The main difficulty is that zero is the singular point of the phase functions of free Ostrovsky equation. We use high-low frequency idea to conquer the difficulties.

math.AP↗

arXiv2021

Convergence problem of Schrödinger equation in Fourier-Lebesgue spaces with rough data and random data

In this paper, we consider the convergence problem of Schrödinger equation. Firstly, we show the almost everywhere pointwise convergence of Schrödinger equation in Fourier-Lebesgue spaces $\hat{H}^{\frac{1}{p},\frac{p}{2}}(\mathbb{R})(4\leq p<\infty),$ $\hat{H}^{\frac{3 s_{1}}{p},\frac{2p}{3}}(\mathbb{R}^2)(s_{1}>\frac{1}{3},3\leq p<\infty),$ $\hat{H}^{\frac{2 s_{1}}{p},p}(\mathbb{R}^n)(s_{1}>\frac{n}{2(n+1)},2\leq p<\infty,n\geq3)$ with rough data. Secondly, we show that the maximal function estimate related to one Schrödinger equation can fail with data in $\hat{H}^{s,\frac{p}{2}}(\mathbb{R})(s<\frac{1}{p})$. Finally, we show the stochastic continuity of Schrödinger equation with random data in $\hat{L}^{r}(\mathbb{R}^n)(2\leq r<\infty)$ almost surely. The main ingredients are Lemmas 2.4, 2.5, 3.2-3.4.

math.AP↗

arXiv2020

Augmented Reality Applied to LEGO Construction: AR-based Building Instructions with High Accuracy & Precision and Realistic Object-Hand Occlusions

BRICKxAR is a novel Augmented Reality (AR) instruction method for construction toys such as LEGO. With BRICKxAR, physical LEGO construction is guided by virtual bricks. Compared with the state-of-the-art, accuracy of the virtual - physical model alignment is significantly improved through a new design of marker-based registration, which can achieve an average error less than 1mm throughout the model. Realistic object occlusion is accomplished to reveal the true spatial relationship between physical and virtual bricks. LEGO players' hand detection and occlusion are realized to visualize the correct spatial relationship between real hands and virtual bricks, and allow virtual bricks to be "grasped" by real hands. The integration of these features makes AR instructions possible for small-parts assembly, validated through a working AR prototype for constructing LEGO Arc de Triomphe, quantitative measures of the accuracies of registration and occlusions, and heuristic evaluation of AR instruction features.

cs.HC↗

arXiv2020

Partially Shared Semi-supervised Deep Matrix Factorization with Multi-view Data

Since many real-world data can be described from multiple views, multi-view learning has attracted considerable attention. Various methods have been proposed and successfully applied to multi-view learning, typically based on matrix factorization models. Recently, it is extended to the deep structure to exploit the hierarchical information of multi-view data, but the view-specific features and the label information are seldom considered. To address these concerns, we present a partially shared semi-supervised deep matrix factorization model (PSDMF). By integrating the partially shared deep decomposition structure, graph regularization and the semi-supervised regression model, PSDMF can learn a compact and discriminative representation through eliminating the effects of uncorrelated information. In addition, we develop an efficient iterative updating algorithm for PSDMF. Extensive experiments on five benchmark datasets demonstrate that PSDMF can achieve better performance than the state-of-the-art multi-view learning approaches. The MATLAB source code is available at https://github.com/libertyhhn/PartiallySharedDMF.

cs.LG↗

arXiv2020

Sharp well-posedness of the Cauchy problem for the rotation-modified Kadomtsev-Petviashvili equation in anisotropic Sobolev spaces

We consider the Cauchy problem for the rotation-modified Kadomtsev-Petviashvili (RMKP) equation \begin{align*} \partial_{x}\left(u_{t}-β\partial_{x}^{3}u +\partial_{x}(u^{2})\right)+\partial_{y}^{2}u-γu=0 \end{align*} in the anisotropic Sobolev spaces $H^{s_{1},\>s_{2}}(\mathbb{R}^{2})$. When $β<0$ and $γ>0,$ we prove that the Cauchy problem is locally well-posed in $H^{s_{1},\>s_{2}}(\mathbb{R}^{2})$ with $s_{1}>-\frac{1}{2}$ and $s_{2}\geq 0$. Our result considerably improves the Theorem 1.4 of R. M. Chen, Y. Liu, P. Z. Zhang( Transactions of the American Mathematical Society, 364(2012), 3395--3425.). The key idea is that we divide the frequency space into regular region and singular region. We further prove that the Cauchy problem for RMKP equation is ill-posed in $H^{s_{1},\>0}(\mathbb{R}^{2})$ with $s_{1}<-\frac{1}{2}$ in the sense that the flow map associated to the rotation-modified Kadomtsev-Petviashvili is not $C^{3}$. When $β<0,γ>0,$ by using the $U^{p}$ and $V^{p}$ spaces, we prove that the Cauchy problem is locally well-posed in $H^{-\frac{1}{2},\>0}(\mathbb{R}^{2})$.

math.AP↗