Search arXiv⌕ Search

arXiv subjects

Di Wu

Publications and source records attributed to Di Wu.

At least 631 records · Page 35Linked to original sources
arXiv2020

SlotRefine: A Fast Non-Autoregressive Model for Joint Intent Detection and Slot Filling

Slot filling and intent detection are two main tasks in spoken language understanding (SLU) system. In this paper, we propose a novel non-autoregressive model named SlotRefine for joint intent detection and slot filling. Besides, we design a novel two-pass iteration mechanism to handle the uncoordinated slots problem caused by conditional independence of non-autoregressive model. Experiments demonstrate that our model significantly outperforms previous models in slot filling task, while considerably speeding up the decoding (up to X 10.77). In-depth analyses show that 1) pretraining schemes could further enhance our model; 2) two-pass mechanism indeed remedy the uncoordinated slots.

cs.CL↗
arXiv2020

Attention Deep Model with Multi-Scale Deep Supervision for Person Re-Identification

In recent years, person re-identification (PReID) has become a hot topic in computer vision duo to it is an important part in intelligent surveillance. Many state-of-the-art PReID methods are attention-based or multi-scale feature learning deep models. However, introducing attention mechanism may lead to some important feature information losing issue. Besides, most of the multi-scale models embedding the multi-scale feature learning block into the feature extraction deep network, which reduces the efficiency of inference network. To address these issue, in this study, we introduce an attention deep architecture with multi-scale deep supervision for PReID. Technically, we contribute a reverse attention block to complement the attention block, and a novel multi-scale layer with deep supervision operator for training the backbone network. The proposed block and operator are only used for training, and discard in test phase. Experiments have been performed on Market-1501, DukeMTMC-reID and CUHK03 datasets. All the experiment results show that the proposed model significantly outperforms the other competitive state-of-the-art methods.

cs.CV↗
arXiv2020

From Distributed Machine Learning To Federated Learning: In The View Of Data Privacy And Security

Federated learning is an improved version of distributed machine learning that further offloads operations which would usually be performed by a central server. The server becomes more like an assistant coordinating clients to work together rather than micro-managing the workforce as in traditional DML. One of the greatest advantages of federated learning is the additional privacy and security guarantees it affords. Federated learning architecture relies on smart devices, such as smartphones and IoT sensors, that collect and process their own data, so sensitive information never has to leave the client device. Rather, clients train a sub-model locally and send an encrypted update to the central server for aggregation into the global model. These strong privacy guarantees make federated learning an attractive choice in a world where data breaches and information theft are common and serious threats. This survey outlines the landscape and latest developments in data privacy and security for federated learning. We identify the different mechanisms used to provide privacy and security, such as differential privacy, secure multi-party computation and secure aggregation. We also survey the current attack models, identifying the areas of vulnerability and the strategies adversaries use to penetrate federated systems. The survey concludes with a discussion on the open challenges and potential directions of future work in this increasingly popular learning paradigm.

cs.DC↗
arXiv2020

When Deep Reinforcement Learning Meets Federated Learning: Intelligent Multi-Timescale Resource Management for Multi-access Edge Computing in 5G Ultra Dense Network

Ultra-dense edge computing (UDEC) has great potential, especially in the 5G era, but it still faces challenges in its current solutions, such as the lack of: i) efficient utilization of multiple 5G resources (e.g., computation, communication, storage and service resources); ii) low overhead offloading decision making and resource allocation strategies; and iii) privacy and security protection schemes. Thus, we first propose an intelligent ultra-dense edge computing (I-UDEC) framework, which integrates blockchain and Artificial Intelligence (AI) into 5G ultra-dense edge computing networks. First, we show the architecture of the framework. Then, in order to achieve real-time and low overhead computation offloading decisions and resource allocation strategies, we design a novel two-timescale deep reinforcement learning (\textit{2Ts-DRL}) approach, consisting of a fast-timescale and a slow-timescale learning process, respectively. The primary objective is to minimize the total offloading delay and network resource usage by jointly optimizing computation offloading, resource allocation and service caching placement. We also leverage federated learning (FL) to train the \textit{2Ts-DRL} model in a distributed manner, aiming to protect the edge devices' data privacy. Simulation results corroborate the effectiveness of both the \textit{2Ts-DRL} and FL in the I-UDEC framework and prove that our proposed algorithm can reduce task execution time up to 31.87%.

cs.NI↗
arXiv2020

A homogenized limit for the 2D Euler equations in a perforated domain

We study the motion of an ideal incompressible fluid in a perforated domain. The porous medium is composed of inclusions of size $a$ separated by distances $\tilde d$ and the fluid fills the exterior. We analyse the asymptotic behavior of the fluid when $(a,\tilde d) \to (0,0)$. If the inclusions are distributed on the unit square, this issue is studied recently when $\frac{\tilde d}a$ tends to zero or infinity, leaving aside the critical case where the volume fraction of the porous medium is below its possible maximal value but non-zero. In this paper, we provide the first result in this regime. In contrast with former results, we obtain an Euler type equation where a homogenized term appears in the elliptic problem relating the velocity and the vorticity. Our analysis is based on the so-called method of reflections whose convergence provides novel estimates on the solutions to the div-curl problem which is involved in the 2D-Euler equations.

math.AP↗
arXiv2020

Are ultra-spinning Kerr-Sen-AdS$_4$ black holes always super-entropic ?

We study thermodynamics of the four-dimensional Kerr-Sen-AdS black hole and its ultra-spinning counterpart, and verify that both black holes fullfil the first law and Bekenstein-Smarr mass formulae of black hole thermodynamics. Furthermore, we derive new Christodoulou-Ruffini-like squared-mass formulae for the usual and ultra-spinning Kerr-Sen-AdS$_4$ solutions. We show that this ultra-spinning Kerr-Sen-AdS$_4$ black hole does not always violate the Reverse Isoperimetric Inequality (RII) since the value of the isoperimetric ratio can be larger/smaller than, or equal to unity, depending upon where the solution parameters lie in the parameters space. This property is obviously different from that of the Kerr-Newman-AdS$_4$ super-entropic black hole, which always strictly violates the RII, although both of them have some similar properties in other aspects, such as horizon geometry and conformal boundary. In addition, it is found that while there exists the same lower bound on mass ($m_e \geqslant 8l/\sqrt{27}$ with $l$ being the cosmological scale) both for the extremal ultra-spinning Kerr-Sen-AdS$_4$ black hole and for the extremal super-entropic Kerr-Newman-AdS$_4$ case, the former has a maximal horizon radius: $r_{\rm\, HP} = l/\sqrt{3}$ which is the minimum of the latter. Therefore, these two different kinds of four-dimensional ultra-spinning charged AdS black holes exhibit some significant physical differences .

gr-qc↗
arXiv2020

Solving Bayesian Risk Optimization via Nested Stochastic Gradient Estimation

In this paper, we aim to solve Bayesian Risk Optimization (BRO), which is a recently proposed framework that formulates simulation optimization under input uncertainty. In order to efficiently solve the BRO problem, we derive nested stochastic gradient estimators and propose corresponding stochastic approximation algorithms. We show that our gradient estimators are asymptotically unbiased and consistent, and that the algorithms converge asymptotically. We demonstrate the empirical performance of the algorithms on a two-sided market model. Our estimators are of independent interest in extending the literature of stochastic gradient estimation to the case of nested risk functions.

math.OC↗
arXiv2020

An Edge Computing-based Photo Crowdsourcing Framework for Real-time 3D Reconstruction

Image-based three-dimensional (3D) reconstruction utilizes a set of photos to build 3D model and can be widely used in many emerging applications such as augmented reality (AR) and disaster recovery. Most of existing 3D reconstruction methods require a mobile user to walk around the target area and reconstruct objectives with a hand-held camera, which is inefficient and time-consuming. To meet the requirements of delay intensive and resource hungry applications in 5G, we propose an edge computing-based photo crowdsourcing (EC-PCS) framework in this paper. The main objective is to collect a set of representative photos from ubiquitous mobile and Internet of Things (IoT) devices at the network edge for real-time 3D model reconstruction, with network resource and monetary cost considerations. Specifically, we first propose a photo pricing mechanism by jointly considering their freshness, resolution and data size. Then, we design a novel photo selection scheme to dynamically select a set of photos with the required target coverage and the minimum monetary cost. We prove the NP-hardness of such problem, and develop an efficient greedy-based approximation algorithm to obtain a near-optimal solution. Moreover, an optimal network resource allocation scheme is presented, in order to minimize the maximum uploading delay of the selected photos to the edge server. Finally, a 3D reconstruction algorithm and a 3D model caching scheme are performed by the edge server in real time. Extensive experimental results based on real-world datasets demonstrate the superior performance of our EC-PCS system over the existing mechanisms.

cs.NI↗
arXiv2020

EasyQuant: Post-training Quantization via Scale Optimization

The 8 bits quantization has been widely applied to accelerate network inference in various deep learning applications. There are two kinds of quantization methods, training-based quantization and post-training quantization. Training-based approach suffers from a cumbersome training process, while post-training quantization may lead to unacceptable accuracy drop. In this paper, we present an efficient and simple post-training method via scale optimization, named EasyQuant (EQ),that could obtain comparable accuracy with the training-based method.Specifically, we first alternately optimize scales of weights and activations for all layers target at convolutional outputs to further obtain the high quantization precision. Then, we lower down bit width to INT7 both for weights and activations, and adopt INT16 intermediate storage and integer Winograd convolution implementation to accelerate inference.Experimental results on various computer vision tasks show that EQ outperforms the TensorRT method and can achieve near INT8 accuracy in 7 bits width post-training.

cs.CV↗
arXiv2020

Study of Charge Radii with Neural Networks

A feed-forward neural network model is trained to calculate the nuclear charge radii. The model trained with input data set of proton and neutron number $Z,N$, the electric quadrupole transition strength $B(E2)$ from the first excited 2$^+$ state to the ground state, together with the symmetry energy. The model reproduces well not only the isotope dependence of charge radii, but also the kinks of charge radii at the neutron magic numbers $N=82$ for Sn and Sm isotopes, and also $N=126$ for Pb isotopes. The important role of $B(E2)$ value is pointed out to reproduce the kink of the isotope dependence of charge radii in these nuclei. Moreover, with the inclusion of the symmetry energy term in the inputs, the charge radii of Ca isotopes are well reproduced. This result suggests a new correlation between the symmetry energy and charge radii of Ca isotopes. The Skyrme HFB calculation is performed to confirm the existence of this correlation in a microscopic model.

nucl-th↗
arXiv2020

Exploring open cluster properties with Gaia and LAMOST

In Gaia DR2, the unprecedented high-precision level reached in sub-mas for astrometry and mmag for photometry. Using cluster members identified with these astrometry and photometry in Gaia DR2, we can obtain a reliable determination of cluster properties. However, because of the shortcoming of Gaia spectroscopic observation in dealing with densely crowded cluster region, the number of radial velocity and metallicity for cluster member stars from Gaia DR2 is still lacking. In this study, we aim to improve the cluster properties by combining the LAMOST spectra. In particular, we provide the list of cluster members with spectroscopic parameters as an add-value catalog in LAMOST DR5, which can be used to perform detailed study for a better understanding on the stellar properties, by using their spectra and fundamental properties from the host cluster. We cross-matched the spectroscopic catalog in LAMOST DR5 with the identified cluster members in Cantat-Gaudin et al.2018 and then used members with spectroscopic parameters to derive statistical properties of open clusters. We obtained a list of 8811 members with spectroscopic parameters and a catalog of 295 cluster properties. In addition, we study the radial and vertical metallicity gradient and age-metallicity relation with the compiled open clusters as tracers, finding slopes of -0.053$\pm$0.004 dex kpc$^{-1}$, -0.252$\pm$0.039 dex kpc$^{-1}$ and 0.022$\pm$0.008 dex Gyr$^{-1}$, respectively. Both slopes of metallicity distribution relation for young clusters (0.1 Gyr < Age < 2 Gyr) and the age-metallicity relation for clusters within 6 Gyr are consistent with literature results. In order to fully study the chemical evolution history in the disk, more spectroscopic observations for old and distant open clusters are needed for further investigation.

astro-ph.GA↗
arXiv2020

Scaling invariant Serrin criterion via one velocity component for the Navier-Stokes equations

In this paper, we prove that the Leray weak solution $u : \mathbb{R}^3\times (0, T)\rightarrow\mathbb{R}^3 $ of the Navier-Stokes equations is regular in $\mathbb{R}^3\times (0,T)$ under the scaling invariant Serrin condition imposed on one component of the velocity $u_3\in L^{q,1}(0, T;L^p(\mathbb{R}^3))$ with \[ \frac{2}{q}+\frac{3}{p}\leq 1,\quad 3<p<+\infty. \] This result is an immediate consequence of a new local regularity criterion in terms of one velocity component for suitable weak solutions.

math.AP↗
arXiv2020

Ferromagnetic MnSn monolayer epitaxially grown on silicon substrate

Two-dimensional (2D) ferromagnetic materials have been exhibiting promising potential in applications, such as spintronics devices. To grow epitaxial magnetic films on silicon substrate, in the single-layer limit, is practically important but challenging. In this study, we realized the epitaxial growth of MnSn monolayer on Si(111) substrate, with an atomically thin Sn/Si(111)-$2\sqrt{3}\times2\sqrt{3}$- buffer layer, and controlled the MnSn thickness with atomic-layer precision. We discovered the ferromagnetism in MnSn monolayer with the Curie temperature (Tc) of ~54 K. As the MnSn film is grown to 4 monolayers, Tc increases accordingly to ~235 K. The lattice of the epitaxial MnSn monolayer as well as the Sn/Si(111)-$2\sqrt{3}\times2\sqrt{3}$ is perfectly compatible with silicon, and thus an sharp interface is formed between MnSn, Sn and Si. This system provides a new platform for exploring the 2D ferromagnetism, integrating magnetic monolayers into silicon-based technology, and engineering the spintronics heterostructures.

cond-mat.mtrl-sci↗
arXiv2020

Energy Model for UAV Communications: Experimental Validation and Model Generalization

Wireless communication involving unmanned aerial vehicles (UAVs) is expected to play an important role in future wireless networks. However, different from conventional terrestrial communication systems, UAVs typically have rather limited onboard energy on one hand, and require additional flying energy consumption on the other hand, which renders energy-efficient UAV communication with smart energy expenditure of paramount importance. In this paper, via extensive flight experiments, we aim to firstly validate the recently derived theoretical energy model for rotary-wing UAVs, and then develop a general model for those complicated flight scenarios where rigorous theoretical model derivation is quite challenging, if not impossible. Specifically, we first investigate how UAV power consumption varies with its flying speed for the simplest straight-and-level flight. With about 12,000 valid power-speed data points collected, we first apply the model-based curve fitting to obtain the modelling parameters based on the theoretical closed-form energy model in the existing literature. In addition, in order to exclude the potential bias caused by the theoretical energy model, the obtained measurement data is also trained using a model-free deep neural network. It is found that the obtained curve from both methods can match quite well with the theoretical energy model. Next, we further extend the study to arbitrary 2-dimensional (2-D) flight, where, to our best knowledge, no rigorous theoretical derivation is available for the closed-form energy model as a function of its flying speed, direction, and acceleration. To fill the gap, we first propose a heuristic energy model for these more complicated cases, and then provide experimental validation based on the measurement results for circular level flight.

eess.SP↗
arXiv2020

Chiral symmetry breaking for deterministic switching of perpendicular magnetization by spin-orbit torque

Symmetry breaking is a characteristic to determine which branch of a bifurcation system follows upon crossing a critical point. Specifically, in spin-orbit torque (SOT) devices, a fundamental question arises: how to break the symmetry of the perpendicular magnetic moment by the in-plane spin polarization? Here, we show that the chiral symmetry breaking by the DMI can induce the deterministic SOT switching of the perpendicular magnetization. By introducing a gradient of saturation magnetization or magnetic anisotropy, non-collinear spin textures are formed by the gradient of effective SOT strength, and thus the chiral symmetry of the SOT-induced spin textures is broken by the DMI, resulting in the deterministic magnetization switching. We introduce a strategy to induce an out-of-plane (z) gradient of magnetic properties, as a practical solution for the wafer-scale manufacture of SOT devices.

cond-mat.mes-hall↗
arXiv2020

Neural Mesh Refiner for 6-DoF Pose Estimation

How can we effectively utilise the 2D monocular image information for recovering the 6D pose (6-DoF) of the visual objects? Deep learning has shown to be effective for robust and real-time monocular pose estimation. Oftentimes, the network learns to regress the 6-DoF pose using a naive loss function. However, due to a lack of geometrical scene understanding from the directly regressed pose estimation, there are misalignments between the rendered mesh from the 3D object and the 2D instance segmentation result, e.g., bounding boxes and masks prediction. This paper bridges the gap between 2D mask generation and 3D location prediction via a differentiable neural mesh renderer. We utilise the overlay between the accurate mask prediction and less accurate mesh prediction to iteratively optimise the direct regressed 6D pose information with a focus on translation estimation. By leveraging geometry, we demonstrate that our technique significantly improves direct regression performance on the difficult task of translation estimation and achieve the state of the art results on Peking University/Baidu - Autonomous Driving dataset and the ApolloScape 3D Car Instance dataset. The code can be found at \url{https://bit.ly/2IRihfU}.

cs.CV↗
arXiv2020

Flow by Gauss curvature to Dual Orlicz-Minkowski problems

In this paper we study a normalised anisotropic Gauss curvature flow of strictly convex, closed hypersurfaces in the Euclidean space R^n+1. We prove that the flow exists for all time and converges smoothly to the unique, strictly convex solution of a Monge-Amp`ere type equation. Our argument provides a parabolic proof in the smooth category for the existence of solutions to the Dual Orlicz-Minkowski problem introduced by Zhu, Xing and Ye.

math.AP↗