arXiv subjects

Qi Wang

Publications and source records attributed to Qi Wang.

At least 577 records · Page 32Linked to original sources

arXiv2021

Long-range spin-wave propagation in transversely magnetized nano-scaled conduits

Magnonics attracts increasing attention in the view of novel low-energy computation technologies based on spin waves. Recently, spin-wave propagation in longitudinally magnetized nano-scaled spin-wave conduits was demonstrated, proving the fundamental feasibility of magnonics at the sub-100 nm scale. Transversely magnetized nano-conduits, which are of great interest in this regard as they offer a large group velocity and a potentially chirality-based protected transport of energy, have not yet been investigated due to their complex internal magnetic field distribution. Here, we present a study of propagating spin waves in a transversely magnetized nanoscopic yttrium iron garnet conduit of 50 nm width. Space and time-resolved micro-focused Brillouin-light-scattering spectroscopy is employed to measure the spin-wave group velocity and decay length. A long-range spin-wave propagation is observed with a decay length of up to (8.0+-1.5) μm and a large spin-wave lifetime of up to (44.7+-9.1) ns. The results are supported with micromagnetic simulations, revealing a single-mode dispersion relation in contrast to the common formation of localized edge modes for microscopic systems. Furthermore, a frequency non-reciprocity for counter-propagating spin waves is observed in the simulations and the experiment, caused by the trapezoidal cross-section of the structure. The revealed long-distance spin-wave propagation on the nanoscale is particularly interesting for an application in spin-wave devices, allowing for long-distance transport of information in magnonic circuits, as well as novel low-energy device architectures.

cond-mat.mes-hall↗

arXiv2021

Neuron Linear Transformation: Modeling the Domain Shift for Crowd Counting

Cross-domain crowd counting (CDCC) is a hot topic due to its importance in public safety. The purpose of CDCC is to alleviate the domain shift between the source and target domain. Recently, typical methods attempt to extract domain-invariant features via image translation and adversarial learning. When it comes to specific tasks, we find that the domain shifts are reflected on model parameters' differences. To describe the domain gap directly at the parameter-level, we propose a Neuron Linear Transformation (NLT) method, exploiting domain factor and bias weights to learn the domain shift. Specifically, for a specific neuron of a source model, NLT exploits few labeled target data to learn domain shift parameters. Finally, the target neuron is generated via a linear transformation. Extensive experiments and analysis on six real-world datasets validate that NLT achieves top performance compared with other domain adaptation methods. An ablation study also shows that the NLT is robust and more effective than supervised and fine-tune training. Code is available at: \url{https://github.com/taohan10200/NLT}.

cs.CV↗

arXiv2020

An Optimal Mass Transport Method for Random Genetic Drift

We propose and analyze an optimal mass transport method for a random genetic drift problem driven by a Moran process under weak-selection. The continuum limit, formulated as a reaction-advection-diffusion equation known as the Kimura equation, inherits degenerate diffusion from the discrete stochastic process that conveys to the blow-up into Dirac-delta singularities hence brings great challenges to both the analytical and numerical studies. The proposed numerical method can quantitatively capture to the fullest possible extent the development of Dirac-delta singularities for genetic segregation on one hand, and preserves several sets of biologically relevant and computationally favored properties of the random genetic drift on the other. Moreover, the numerical scheme exponentially converges to the unique numerical stationary state in time at a rate independent of the mesh size up to a mesh error. Numerical evidence is given to illustrate and support these properties, and to demonstrate the spatio-temporal dynamics of random generic drift.

math.AP↗

arXiv2020

Universal properties of penetrative turbulent Rayleigh--Bénard convection in cold water near $4^\circ\rm{C}$

Penetrative turbulence, which occurs in a convectively unstable fluid layer and penetrates into an adjacent, originally stably stratified layer, is numerically and theoretically analyzed. We chose the most relevant example, namely thermally driven flow of water with a temperature around $T_m\approx 4^\circ\rm{C}$, where it has its density maximum. We pick the Rayleigh-Bénard geometry with the bottom plate temperature $T_b > 4^\circ\rm{C}$ and the top plate temperature $T_t \le 4^\circ\rm{C}$. Next to the overall thermal driving strength set by the temperature difference $Δ= T_b - T_t$ (the Rayleigh number $Ra$ in dimensionless form), the crucial new control parameter as compared to standard Rayleigh-Bénard convection is the density inversion parameter $θ_m \equiv (T_m - T_t ) / Δ$. The crucial response parameters are the relative mean mid-height temperature $θ_c$ and the overall heat transfer (i.e., the Nusselt number $Nu$). We theoretically derive the universal (i.e., $Ra$-independent) dependence $θ_c (θ_m) =(1+θ_m^2)/2$, which holds for $θ_m$ below a $Ra$-dependent critical value, beyond which $θ_c (θ_m)$ sharply decreases and drops down to $θ_c=1/2$ at $θ_m=θ_{m,c}$. Our direct numerical simulations with $Ra$ up to $10^{10}$ are consistent with these results. The critical density inversion parameter $θ_{m,c}$ can be precisely predicted by a linear stability analysis. The heat flux $Nu(θ_m)$ monotonically decreases with increasing $θ_m$ and we can theoretically derive a universal relation for the relative heat flux $Nu(θ_m)/Nu(0)$. Finally, we numerically identify and discuss rare transitions between different turbulent flow states for large $θ_m$.

physics.flu-dyn↗

arXiv2020

Inverse-design magnonic devices

The field of magnonics offers a new type of low-power information processing, in which magnons, the quanta of spin waves, carry and process data instead of electrons. Many magnonic devices were demonstrated recently, but the development of each of them requires specialized investigations and, usually, one device design is suitable for one function only. Here, we introduce the method of inverse-design magnonics, in which any functionality can be specified first, and a feedback-based computational algorithm is used to obtain the device design. Our proof-of-concept prototype is based on a rectangular ferromagnetic area which can be patterned using square shaped voids. To demonstrate the universality of this approach, we explore linear, nonlinear and nonreciprocal magnonic functionalities and use the same algorithm to create a magnonic (de-)multiplexer, a nonlinear switch and a circulator. Thus, inverse-design magnonics can be used to develop highly efficient rf applications as well as Boolean and neuromorphic computing building blocks.

physics.app-ph↗

arXiv2020

Regime transitions in thermally driven high-Rayleigh number vertical convection

Vertical convection is investigated using direct numerical simulations over a wide range of Rayleigh numbers $10^7\le Ra\le10^{14}$ with fixed Prandtl number $Pr=10$, in a two-dimensional convection cell with unit aspect ratio. It is found that the dependence of the mean vertical centre temperature gradient $S$ on $Ra$ shows three different regimes: In regime I ($Ra \lesssim 5\times10^{10}$), $S$ is almost independent of $Ra$; In the newly identified regime II ($5\times10^{10} \lesssim Ra \lesssim 10^{13}$), $S$ first increases with increasing $Ra$ (regime ${\rm{II}}_a$), reaches its maximum and then decreases again (regime ${\rm{II}}_b$); In regime III ($Ra\gtrsim10^{13}$), $S$ again becomes only weakly dependent on $Ra$, being slightly smaller than in regime I. The transitions between diffeereent regimes are discussd. In the three different regimes, significantly different flow organizations are identified: In regime I and regime ${\rm{II}}_a$, the location of the maximal horizontal velocity is close to the top and bottom walls; However, in regime ${\rm{II}}_b$ and regime III, banded zonal flow structures develop and the maximal horizontal velocity now is in the bulk region. The different flow organizations in the three regimes are also reflected in the scaling exponents in the effective power law scalings $Nu\sim Ra^β$ and $Re\sim Ra^γ$. In regime I, the fitted scaling exponents ($β\approx0.26$ and $γ\approx0.51$) are in excellent agreement with the theoretical predication of $β=1/4$ and $γ=1/2$ for laminar VC (Shishkina, {\it{Phys. Rev. E.}} 2016, 93, 051102). However, in regimes II and III, $β$ increases to a value close to 1/3 and $γ$ decreases to a value close to 4/9. The stronger $Ra$ dependence of $Nu$ is related to the ejection of plumes and larger local heat flux at the walls.

physics.flu-dyn↗

arXiv2020

Two-layer Thermally Driven Turbulence: Mechanisms for Interface Breakup

It is commonly accepted that the breakup criteria of drops or bubbles in turbulence is governed by surface tension and inertia. However, also {\it{buoyancy}} can play an important role at breakup. In order to better understand this role, here we numerically study Rayleigh-Bénard convection for two immiscible fluid layers, in order to identify the effects of buoyancy on interface breakup. We explore the parameter space spanned by the Weber number $5\leq We \leq 5000$ (the ratio of inertia to surface tension) and the density ratio between the two fluids $0.001 \leq Λ\leq 1$, at fixed Rayleigh number $Ra=10^8$ and Prandtl number $Pr=1$. At low $We$, the interface undulates due to plumes. When $We$ is larger than a critical value, the interface eventually breaks up. Depending on $Λ$, two breakup types are observed: The first type occurs at small $Λ\ll 1$ (e.g. air-water systems) when local filament thicknesses exceed the Hinze length scale. The second, strikingly different, type occurs at large $Λ$ with roughly $0.5 < Λ\le 1$ (e.g. oil-water systems): The layers undergo a periodic overturning caused by buoyancy overwhelming surface tension. For both types the breakup criteria can be derived from force balance arguments and show good agreement with the numerical results.

physics.flu-dyn↗

arXiv2020

Real-time Prediction of Soft Tissue Deformations Using Data-driven Nonlinear Presurgical Simulations

Imaging modalities provide clinicians with real-time visualization of anatomical regions of interest (ROI) for the purpose of minimally invasive surgery. During the procedure, low-resolution image data are acquired and registered with high-resolution preoperative 3D reconstruction to guide the execution of surgical preplan. Unfortunately, due to the potential large strain and nonlinearities in the deformation of soft biological tissues, significant mismatch may be observed between ROI shapes during pre- and intra-operative imaging stages, making the surgical preplan prone to failure. In an effort to bridge the gap between the two imaging stages, this paper presents a data-driven approach based on artificial neural network for predicting the ROI deformation in real-time with sparsely registered fiducial markers. For a head-and-neck tumor model with an average maximum displacement of 30 mm, the maximum surface offsets between benchmarks and predictions using the proposed approach for 98% of the test cases are under 1.0 mm, which is the typical resolution of high-quality interventional ultrasound. Each of the prediction processes takes less than 0.5 s. With the resulting prediction accuracy and computational efficiency, the proposed approach demonstrates its potential to be clinically relevant.

physics.med-ph↗

arXiv2020

Doubly Stochastic Variational Inference for Neural Processes with Hierarchical Latent Variables

Neural processes (NPs) constitute a family of variational approximate models for stochastic processes with promising properties in computational efficiency and uncertainty quantification. These processes use neural networks with latent variable inputs to induce predictive distributions. However, the expressiveness of vanilla NPs is limited as they only use a global latent variable, while target specific local variation may be crucial sometimes. To address this challenge, we investigate NPs systematically and present a new variant of NP model that we call Doubly Stochastic Variational Neural Process (DSVNP). This model combines the global latent variable and local latent variables for prediction. We evaluate this model in several experiments, and our results demonstrate competitive prediction performance in multi-output regression and uncertainty estimation in classification.

cs.LG↗

arXiv2020

A nonlinear magnonic nano-ring resonator

The field of magnonics, which aims at using spin waves as carriers in data processing devices, has attracted increasing interest in recent years. We present and study micromagnetically a nonlinear nanoscale magnonic ring resonator device for enabling implementations of magnonic logic gates and neuromorphic magnonic circuits. In the linear regime, this device efficiently suppresses spin-wave transmission using the phenomenon of critical resonant coupling, thus exhibiting the behavior of a notch filter. By increasing the spin-wave input power, the resonance frequency is shifted leading to transmission curves, depending on the frequency, reminiscent of the activation functions of neurons or showing the characteristics of a power limiter. An analytical theory is developed to describe the transmission curve of magnonic ring resonators in the linear and nonlinear regimes and validated by a comprehensive micromagnetic study. The proposed magnonic ring resonator provides a multi-functional nonlinear building block for unconventional magnonic circuits.

physics.app-ph↗

arXiv2020

Feature-aware Adaptation and Density Alignment for Crowd Counting in Video Surveillance

With the development of deep neural networks, the performance of crowd counting and pixel-wise density estimation are continually being refreshed. Despite this, there are still two challenging problems in this field: 1) current supervised learning needs a large amount of training data, but collecting and annotating them is difficult; 2) existing methods can not generalize well to the unseen domain. A recently released synthetic crowd dataset alleviates these two problems. However, the domain gap between the real-world data and synthetic images decreases the models' performance. To reduce the gap, in this paper, we propose a domain-adaptation-style crowd counting method, which can effectively adapt the model from synthetic data to the specific real-world scenes. It consists of Multi-level Featureaware Adaptation (MFA) and Structured Density map Alignment (SDA). To be specific, MFA boosts the model to extract domain-invariant features from multiple layers. SDA guarantees the network outputs fine density maps with a reasonable distribution on the real domain. Finally, we evaluate the proposed method on four mainstream surveillance crowd datasets, Shanghai Tech Part B, WorldExpo'10, Mall and UCSD. Extensive experiments evidence that our approach outperforms the state-of-the-art methods for the same cross-domain counting problem.

cs.CV↗

arXiv2020

Unsupervised Semantic Aggregation and Deformable Template Matching for Semi-Supervised Learning

Unlabeled data learning has attracted considerable attention recently. However, it is still elusive to extract the expected high-level semantic feature with mere unsupervised learning. In the meantime, semi-supervised learning (SSL) demonstrates a promising future in leveraging few samples. In this paper, we combine both to propose an Unsupervised Semantic Aggregation and Deformable Template Matching (USADTM) framework for SSL, which strives to improve the classification performance with few labeled data and then reduce the cost in data annotating. Specifically, unsupervised semantic aggregation based on Triplet Mutual Information (T-MI) loss is explored to generate semantic labels for unlabeled data. Then the semantic labels are aligned to the actual class by the supervision of labeled data. Furthermore, a feature pool that stores the labeled samples is dynamically updated to assign proxy labels for unlabeled data, which are used as targets for cross-entropy minimization. Extensive experiments and analysis across four standard semi-supervised learning benchmarks validate that USADTM achieves top performance (e.g., 90.46$\%$ accuracy on CIFAR-10 with 40 labels and 95.20$\%$ accuracy with 250 labels). The code is released at https://github.com/taohan10200/USADTM.

cs.CV↗

arXiv2020

Scaling in internally heated convection: a unifying theory

We offer a unifying theory for turbulent purely internally heated convection, generalizing the unifying theories of Grossmann and Lohse (2000, 2001) for Rayleigh--Bénard turbulence and of Shishkina, Grossmann and Lohse (2016) for turbulent horizontal convection, which are both based on the splitting of the kinetic and thermal dissipation rates in respective boundary and bulk contributions. We obtain the mean temperature of the system and the Reynolds number (which are the response parameters) as function of the control parameters, namely the internal thermal driving strength (called, when nondimensionalized, the Rayleigh--Roberts number) and the Prandtl number. The results of the theory are consistent with our direct numerical simulations.

physics.flu-dyn↗

arXiv2020

Observation of contact angle hysteresis due to inhomogeneous electric fields

Static contact angle hysteresis (CAH) is widely attributed to surface roughness and chemical contamination. In the latter case, chemical defects create free-energy barriers that prevent the contact line motion. Electrowetting studies have demonstrated the similar ability of electric fields to alter the surface free-energy landscape. Yet, the increase of apparent static CAH by electric fields remains unseen. Here, we report the observation and theoretical analysis of electrowetting hysteresis. This phenomenon enables the continuous and dynamic control of CAH, not only for fundamental studies but also to manufacture sticky-on-demand surfaces for sample collection.

physics.flu-dyn↗

arXiv2020

Electronic correlations and flattened band in magnetic Weyl semimetal candidate Co3Sn2S2

The interplay between electronic correlations and topological protection may offer a rich avenue for discovering emergent quantum phenomena in condensed matter. However, electronic correlations have so far been little investigated in Weyl semimetals (WSMs) by experiments. Here, we report a combined optical spectroscopy and theoretical calculation study on the strength of electronic correlations in a kagome magnet Co3Sn2S2 and the influence of electronic correlations on its WSM state expected within a single-particle picture. The electronic kinetic energy estimated from our optical data is about half of that obtained from single-particle ab initio calculations, which indicates intermediate-strength electronic correlations in this system. Furthermore, comparing the energy ratios between the interband-transition peaks at high energies in the experimental and single-particle-ab-initio-calculation derived optical conductivity spectra with the electronic bandwidth renormalization factors obtained by many-body calculations enables us to estimate the Coulomb-interaction strength (U ~ 4 eV) of electronic correlations in Co3Sn2S2. Our many-body calculations with U ~ 4 eV show that a WSM state, which is characterized by bulk Weyl cones and surface Fermi arcs, survives in this correlated electron system. Besides, a sharp experimental optical conductivity peak at low energy, which is absent in the single-particle-ab-initio-calculation-derived optical conductivity spectrum but is consistent with the optical conductivity peaks obtained by many-body calculations, indicates that an electronic band connecting the two Weyl cones is flattened by electronic correlations and emerges near the Fermi energy in Co3Sn2S2. Our work paves the way for exploring flat-band-generated quantum phenomena in WSMs.

cond-mat.mtrl-sci↗

arXiv2020

Detecting AI Trojans Using Meta Neural Analysis

In machine learning Trojan attacks, an adversary trains a corrupted model that obtains good performance on normal data but behaves maliciously on data samples with certain trigger patterns. Several approaches have been proposed to detect such attacks, but they make undesirable assumptions about the attack strategies or require direct access to the trained models, which restricts their utility in practice. This paper addresses these challenges by introducing a Meta Neural Trojan Detection (MNTD) pipeline that does not make assumptions on the attack strategies and only needs black-box access to models. The strategy is to train a meta-classifier that predicts whether a given target model is Trojaned. To train the meta-model without knowledge of the attack strategy, we introduce a technique called jumbo learning that samples a set of Trojaned models following a general distribution. We then dynamically optimize a query set together with the meta-classifier to distinguish between Trojaned and benign models. We evaluate MNTD with experiments on vision, speech, tabular data and natural language text datasets, and against different Trojan attacks such as data poisoning attack, model manipulation attack, and latent attack. We show that MNTD achieves 97% detection AUC score and significantly outperforms existing detection approaches. In addition, MNTD generalizes well and achieves high detection performance against unforeseen attacks. We also propose a robust MNTD pipeline which achieves 90% detection AUC even when the attacker aims to evade the detection with full knowledge of the system.

cs.AI↗

arXiv2020

A Flow Base Bi-path Network for Cross-scene Video Crowd Understanding in Aerial View

Drones shooting can be applied in dynamic traffic monitoring, object detecting and tracking, and other vision tasks. The variability of the shooting location adds some intractable challenges to these missions, such as varying scale, unstable exposure, and scene migration. In this paper, we strive to tackle the above challenges and automatically understand the crowd from the visual data collected from drones. First, to alleviate the background noise generated in cross-scene testing, a double-stream crowd counting model is proposed, which extracts optical flow and frame difference information as an additional branch. Besides, to improve the model's generalization ability at different scales and time, we randomly combine a variety of data transformation methods to simulate some unseen environments. To tackle the crowd density estimation problem under extreme dark environments, we introduce synthetic data generated by game Grand Theft Auto V(GTAV). Experiment results show the effectiveness of the virtual data. Our method wins the challenge with a mean absolute error (MAE) of 12.70. Moreover, a comprehensive ablation study is conducted to explore each component's contribution.

cs.CV↗

arXiv2020

Excitation function of initial temperature of heavy flavor quarkonium emission source in high energy collisions

The transverse momentum spectra of $J/ψ$, $ψ(2S)$, and $Υ(nS, n=1,2,3)$ produced in proton-proton ($p$+$p$), proton-antiproton ($p$+$\bar{p}$), proton-lead ($p$+Pb), gold-gold (Au+Au), and lead-lead (Pb+Pb) collisions over a wide energy range are analyzed by the (two-component) Erlang distribution, the Hagedorn function (the inverse power-law), and the Tsallis-Levy function. The initial temperature is obtained from the color string percolation model due to the fit by the (two-component) Erlang distribution in the framework of multisource thermal model. The excitation functions of some parameters such as the mean transverse momentum and initial temperature increase from dozens of GeV to above 10 TeV. The mean transverse momentum and initial temperature decrease (increase slightly or do not change obviously) with the increase of rapidity (centrality). Meanwhile, the mean transverse momentum of $Υ(nS, n=1,2,3)$ is larger than that of $J/ψ$ and $ψ(2S)$, and the initial temperature for $Υ(nS, n=1,2,3)$ emission is higher than that for $J/ψ$ and $ψ(2S)$ emission, which shows a mass-dependent behavior.

hep-ph↗