arXiv subjects

Kai Yang

Publications and source records attributed to Kai Yang.

At least 235 records · Page 13Linked to original sources

arXiv2020

Reversible Single Spin Control of Individual Magnetic Molecule by Hydrogen Atom Adsorption

The reversible control of a single spin of an atom or a molecule is of great interest in Kondo physics and a potential application in spin based electronics.Here we demonstrate that the Kondo resonance of manganese phthalocyanine molecules on an Au(111) substrate have been reversibly switched off and on via a robust route through attachment and detachment of single hydrogen atom to the magnetic core of the molecule. As further revealed by density functional theory calculations, even though the total number of electrons of the Mn ion remains almost the same in the process, gaining one single hydrogen atom leads to redistribution of charges within 3d orbitals with a reduction of the molecular spin state from S = 3/2 to S = 1 that directly contributes to the Kondo resonance disappearance. This process is reversed by a local voltage pulse or thermal annealing to desorb the hydrogen atom.

cond-mat.mes-hall↗

arXiv2020

Revealing the Atomic Site-Dependent g Factor within a Single Magnetic Molecule via the Extended Kondo Effect

The site-dependent g factor of a single magnetic molecule, with intramolecular resolution, is demonstrated for the first time by low-temperature, high-magnetic-field scanning tunneling microscopy of dehydrogenated Mn-phthalocyanine molecules on Au(111). This is achieved by exploring the magneticfield dependence of the extended Kondo effect at different atomic sites of the molecule. Importantly, an inhomogeneous distribution of the g factor inside a single molecule is revealed. Our results open up a new route to access local spin properties within a single molecule.

cond-mat.mes-hall↗

arXiv2020

Energy-Efficient Processing and Robust Wireless Cooperative Transmission for Edge Inference

Edge machine learning can deliver low-latency and private artificial intelligent (AI) services for mobile devices by leveraging computation and storage resources at the network edge. This paper presents an energy-efficient edge processing framework to execute deep learning inference tasks at the edge computing nodes whose wireless connections to mobile devices are prone to channel uncertainties. Aimed at minimizing the sum of computation and transmission power consumption with probabilistic quality-of-service (QoS) constraints, we formulate a joint inference tasking and downlink beamforming problem that is characterized by a group sparse objective function. We provide a statistical learning based robust optimization approach to approximate the highly intractable probabilistic-QoS constraints by nonconvex quadratic constraints, which are further reformulated as matrix inequalities with a rank-one constraint via matrix lifting. We design a reweighted power minimization approach by iteratively reweighted $\ell_1$ minimization with difference-of-convex-functions (DC) regularization and updating weights, where the reweighted approach is adopted for enhancing group sparsity whereas the DC regularization is designed for inducing rank-one solutions. Numerical results demonstrate that the proposed approach outperforms other state-of-the-art approaches.

cs.IT↗

arXiv2020

Communication-Efficient Edge AI: Algorithms and Systems

Artificial intelligence (AI) has achieved remarkable breakthroughs in a wide range of fields, ranging from speech processing, image classification to drug discovery. This is driven by the explosive growth of data, advances in machine learning (especially deep learning), and easy access to vastly powerful computing resources. Particularly, the wide scale deployment of edge devices (e.g., IoT devices) generates an unprecedented scale of data, which provides the opportunity to derive accurate models and develop various intelligent applications at the network edge. However, such enormous data cannot all be sent from end devices to the cloud for processing, due to the varying channel quality, traffic congestion and/or privacy concerns. By pushing inference and training processes of AI models to edge nodes, edge AI has emerged as a promising alternative. AI at the edge requires close cooperation among edge devices, such as smart phones and smart vehicles, and edge servers at the wireless access points and base stations, which however result in heavy communication overheads. In this paper, we present a comprehensive survey of the recent developments in various techniques for overcoming these communication challenges. Specifically, we first identify key communication challenges in edge AI systems. We then introduce communication-efficient techniques, from both algorithmic and system perspectives for training and inference tasks at the network edge. Potential future research directions are also highlighted.

cs.IT↗

arXiv2020

Stable blow-up dynamics in the $L^2$-critical and $L^2$-supercritical generalized Hartree equation

We study stable blow-up dynamics in the generalized Hartree equation with radial symmetry, a Schrödinger-type equation with a nonlocal, convolution-type nonlinearity: $iu_t+Δu +\left(|x|^{-(d-2)} \ast |u|^{p} \right) |u|^{p-2}u = 0, x \in \mathbb{R}^d$. First, we consider the $L^2$-critical case in dimensions d=3, 4, 5, 6, 7 and obtain that a generic blow-up has a self-similar structure and exhibits not only the square root blowup rate but also the log-log correction (via asymptotic analysis and functional fitting). In this setting we also study blow-up profiles and show that generic blow-up solutions converge to the rescaled $Q$, a ground state solution of the elliptic equation $-ΔQ+Q- \left(|x|^{-(d-2)} \ast |Q|^p \right) |Q|^{p-2} Q =0$. We also consider the $L^2$-supercritical case in dimensions d=3,4. We derive the profile equation for the self-similar blow-up and establish the existence and local uniqueness of its solutions. As in the NLS $L^2$-supercritical regime, the profile equation exhibits branches of non-oscillating, polynomially decaying (multi-bump) solutions. A numerical scheme of putting constraints into solving the corresponding ODE is applied during the process of finding the multi-bump solutions. Direct numerical simulation of solutions to the generalized Hartree equation by the dynamic rescaling method indicates that the $Q_{1,0}$ is the profile for the stable blow-up. In this supercritical case, we obtain the blow-up rate without any correction. This blow-up happens at the focusing level $10^{-5}$, and thus, numerically observable (unlike the $L^2$-critical case). In summary, we find that the results are similar to the behavior of stable blowup dynamics in the corresponding NLS settings. Consequently, one may expect that the form of the nonlinearity in the Schrödinger-type equations is not essential in stable blow-up.

math.AP↗

arXiv2020

Robust block preconditioners for poroelasticity

In this paper we study the linear systems arising from discretized poroelasticity problems. We formulate one block preconditioner for the two-filed Biot model and several preconditioners for the classical three-filed Biot model under the unified relationship framework between well-posedness and preconditioners. By the unified theory, we show all the considered preconditioners are uniformly optimal with respect to material and discretization parameters. Numerical tests demonstrate the robustness of these preconditioners.

math.NA↗

arXiv2019

Compressive Sensing Based Channel Estimation for Millimeter-Wave Full-Dimensional MIMO with Lens-Array

Channel estimation (CE) for millimeter-wave (mmWave) lens-array suffers from prohibitive training overhead, whereas the state-of-the-art solutions require an extra complicated radio frequency phase shift network. By contrast, lens-array using antenna switching network (ASN) simplifies the hardware, but the associated CE is a challenging task due to the constraint imposed by ASN. This paper proposes a compressive sensing (CS)-based CE solution for full-dimensional (FD) lens-array, where the mmWave channel sparsity is exploited. Specifically, we first propose an approach of pilot training under the more severe haraware constraint imposed by ASN, and formulate the associated CE of lens-array as a CS problem. Then, a redundant dictionary is tailored for FD lens-array to combat the power leakage caused by the continuous angles of multipath components. Further, we design the baseband pilot signals to minimize the total mutual coherence of the measurement matrix based on CS theory for more reliable CE performance. Our solution provides a framework for applying CS techniques to lens-array using simple and practical ASN. Simulation results demonstrate the effectiveness of the proposed scheme.

cs.IT↗

arXiv2019

Reconfigurable Intelligent Surface for Green Edge Inference

Reconfigurable intelligent surface (RIS) as an emerging cost-effective technology can enhance the spectrum- and energy-efficiency of wireless networks. In this paper, we consider an RIS-aided green edge inference system, where the inference tasks generated from resource-limited mobile devices (MDs) are uploaded to and cooperatively performed at multiple resource-enhanced base stations (BSs). Taking into account both the computation and uplink/downlink transmit power consumption, we formulate an overall network power consumption minimization problem, which calls for the joint design of the set of tasks performed by each BS, transmit and receive beamforming vectors of the BSs, transmit power of the MDs, and uplink/downlink phase-shift matrices at the RIS. Such a problem is a mixed combinatorial optimization problem with nonconvex constraints and is highly intractable. To address the challenge of the combinatorial objective, a group sparse reformulation is proposed by exploiting the group sparsity structure of the beamforming vectors, while a block-structured optimization (BSO) approach is proposed to decouple the optimization variables. Finally, we propose a BSO with mixed $\ell_{1,2}$-norm and difference-of-convex-functions (DC) based three-stage framework to solve the problem, where the mixed $\ell_{1,2}$-norm is adopted to induce the group sparsity of beamforming vectors and DC is adopted to effectively handle the nonconvex rank-one constraint after matrix lifting. Numerical results demonstrate the supreme gain of deploying an RIS and confirm the effectiveness of the proposed algorithm over the baseline algorithms.

eess.SP↗

arXiv2019

A Quasi-Newton Method Based Vertical Federated Learning Framework for Logistic Regression

Data privacy and security becomes a major concern in building machine learning models from different data providers. Federated learning shows promise by leaving data at providers locally and exchanging encrypted information. This paper studies the vertical federated learning structure for logistic regression where the data sets at two parties have the same sample IDs but own disjoint subsets of features. Existing frameworks adopt the first-order stochastic gradient descent algorithm, which requires large number of communication rounds. To address the communication challenge, we propose a quasi-Newton method based vertical federated learning framework for logistic regression under the additively homomorphic encryption scheme. Our approach can considerably reduce the number of communication rounds with a little additional communication cost per round. Numerical results demonstrate the advantages of our approach over the first-order method.

cs.LG↗

arXiv2019

Inexact Primal-Dual Gradient Projection Methods for Nonlinear Optimization on Convex Set

In this paper, we propose a novel primal-dual inexact gradient projection method for nonlinear optimization problems with convex-set constraint. This method only needs inexact computation of the projections onto the convex set for each iteration, consequently reducing the computational cost for projections per iteration. This feature is attractive especially for solving problems where the projections are computationally not easy to calculate. Global convergence guarantee and O(1/k) ergodic convergence rate of the optimality residual are provided under loose assumptions. We apply our proposed strategy to l1-ball constrained problems. Numerical results exhibit that our inexact gradient projection methods for solving l1-ball constrained problems are more efficient than the exact methods.

math.OC↗

arXiv2019

The effects of urbanization on the temperature over the Beijing-Tianjin-Hebei region under changing climate

Urbanization has a large impact on human society. Besides, global warming has become an increasingly popular topic since global warming will not only play a significant role in human beings daily life but also have a huge impact on the global climate. Under the background of global warming, the impacts of urbanization may have differences, which is the topic this project mainly discusses. In this study, the monthly air temperature data from 2 meteorological stations in Beijing under the different underlying surface: urban and suburban for a period of 1960-2014 were analyzed. Besides, two years, 1993 and 2010, selected with the different circulation conditions have been simulated by weather forecasting models. We conducted experiments using the Weather Research and Forecasting (WRF) model to investigate the impacts of urbanization under different circulation conditions on the summer temperature over the Beijing-Tianjin-Hebei Region (BTHR). The results indicate that the effect of urbanization under low-pressure system is greater than that under high-pressure system. That is because the difference of total cloud cover between the urban area and the suburban area is greater under low-pressure system in 1993. This research will not only increase our understanding of the effect of urbanization but also provide a scientific basis to enhance the ability to prevent disasters and reduce damages under the global warming background.

physics.ao-ph↗

arXiv2019

Coverage Analysis of Integrated Sub-6GHz-mmWave Cellular Networks with Hotspots

Deploying Sub-6GHz networks together with millimeter wave (mmWave) is a promising solution to achieve high data rates in traffic hotspots while guaranteeing sufficient coverage, where mmWave small cells are densely deployed to provide high quality of service. In this paper, we propose an analytical framework to investigate the integrated Sub-6GHz-mmWave cellular networks, in which the Sub-6GHz base stations (BSs) are modeled as a Poisson point process, and the mmWave BSs are clustered following a Poisson cluster process in traffic hotspots. We conduct stochastic geometry-based analysis and derive the performance metrics including the association probability, signal-to-interference-plus-noise ratio coverage probability and average achievable rate, which are validated to be accurate by Monte Carlo simulations. We analyze the impact of various deployment parameters on the network performance to give insights on the network design. In particular, it is shown that deploying mmWave small cells in traffic hotspots will outperform both traditional Sub-6GHz heterogeneous network and isolated mmWave system in terms of the coverage probability. It can also be shown that extremely high and extremely small association weight for mmWave BSs will deteriorate the performance for cell edge users and cell interior users, respectively. Moreover, there exists an optimal pre-decided dispersion parameter of mmWave BSs that contributes to the maximum coverage probability.

cs.IT↗

arXiv2019

Structural analysis of nuclear spin clusters via two-dimensional nanoscale nuclear magnetic resonance spectroscopy

Two-dimensional Nuclear Magnetic Resonance (NMR) is essential in molecular structure determination. The Nitrogen-Vacancy (NV) center in diamond has been proposed and developed as an outstanding quantum sensor to realize NMR in nanoscale. In this work, we develop a scheme for two-dimensional nanoscale NMR spectroscopy based on quantum controls on an NV center. We carry out a proof of principle experiment on a target of two coupled $^{13}$C nuclear spins in diamond. A COSY-like sequences is used to acquire the data on time domain, which is then converted to frequency domain with the fast Fourier transform (FFT). With the two-dimensional NMR spectrum, the structure and location of the set of nuclear spin are resolved. This work marks a fundamental step towards resolving the structure of a single molecule.

quant-ph↗

arXiv2019

A 6.7 GHz Methanol Maser Survey II. Low Galactic Latitudes

We report the results of our systematic survey for Galactic 6.7 GHz Class II CH$_3$OH maser emission toward a sample of young stellar objects. The survey was conducted with the Shanghai Tianma Radio Telescope (TMRT). The sample consists of 3348 sources selected from the all-sky \emph{Wide-Field Infrared Survey Explorer (WISE)} point source catalog. We have discussed the selection criteria in detail and the detection results of those at high Galactic latitudes (i.e. $|b|>$ 2$^\circ$) in a previous paper (paper I). Here, we present the results from the survey of those at low Galactic latitudes, i.e. $|b|<2^\circ$. Of 1875 selected \emph{WISE} point sources, 291 positions that were actually associated with 224 sources were detected with CH$_3$OH maser emission. Among them, 32 are newly detected. Majority of the newly detected sources are associated with bright WISE sources. The majority of the detected sources (209/224 = 93.3\%) are quite close to the Galactic Plane ($|b|<1^\circ$) and lie on the inner spiral arms with positive LSR velocities. Detection rate and the color-color distribution of our detection are all matched with our anticipation. Combining with detections from previous surveys, we compile a catalogue of 1085 sources with 6.7 GHz CH$_3$OH maser emission in our Galaxy.

astro-ph.GA↗

arXiv2019

Tuning the exchange bias on a single atom from 1 mT to 10 T

Shrinking spintronic devices to the nanoscale ultimately requires localized control of individual atomic magnetic moments. At these length scales, the exchange interaction plays important roles, such as in the stabilization of spin-quantization axes, the production of spin frustration, and creation of magnetic ordering. Here, we demonstrate the precise control of the exchange bias experienced by a single atom on a surface, covering an energy range of four orders of magnitude. The exchange interaction is continuously tunable from milli-eV to micro-eV by adjusting the separation between a spin-1/2 atom on a surface and the magnetic tip of a scanning tunneling microscope (STM). We seamlessly combine inelastic electron tunneling spectroscopy (IETS) and electron spin resonance (ESR) to map out the different energy scales. This control of exchange bias over a wide span of energies provides versatile control of spin states, with applications ranging from precise tuning of quantum state properties, to strong exchange bias for local spin doping. In addition we show that a time-varying exchange interaction generates a localized AC magnetic field that resonantly drives the surface spin. The static and dynamic control of the exchange interaction at the atomic-scale provides a new tool to tune the quantum states of coupled-spin systems.

cond-mat.mes-hall↗

arXiv2019

A 6.7 GHz Methanol Maser Survey at High Galactic Latitudes

We performed a systematic 6.7 GHz Class \uppercase\expandafter{\romannumeral2} methanol maser survey using the Shanghai Tianma Radio Telescope toward targets selected from the all-sky \emph{Wide-Field Infrared Survey Explorer (WISE)} point catalog. In this paper, we report the results from the survey of those at high Galactic latitudes, i.e. $|b|>$ 2$^\circ$. Of 1473 selected \emph{WISE} point sources at high latitude, 17 point positions that were actually associated with 12 sources were detected with maser emission, reflecting the rarity (1$-$2\%) of methanol masers in the region away from the Galactic plane. Out of the 12 sources, 3 are detected for the first time. The spectral energy distribution (SED) at infrared bands shows that these new detected masers occur in the massive star forming regions. Compared to previous detections, the methanol maser changes significantly in both spectral profiles and flux densities. The infrared \emph{WISE} images show that almost all of these masers are located in the positions of the bright \emph{WISE} point sources. Compared with the methanol masers at the Galactic plane, these high-latitude methanol masers provide good tracers for investigating the physics and kinematics around massive young stellar objects, because they are believed to be less affected by the surrounding cluster environment.

astro-ph.GA↗

arXiv2019

Hadoop Perfect File: A fast access container for small files with direct in disc metadata access

Storing and processing massive small files is one of the major challenges for the Hadoop Distributed File System (HDFS). In order to provide fast data access, the NameNode (NN) in HDFS maintains the metadata of all files in its main-memory. Hadoop performs well with a small number of large files that require relatively little metadata in the NN s memory. But for a large number of small files, Hadoop has problems such as NN memory overload caused by the huge metadata size of these small files. We present a new type of archive file, Hadoop Perfect File (HPF), to solve HDFS s small files problem by merging small files into a large file on HDFS. Existing archive files offer limited functionality and have poor performance when accessing a file in the merged file due to the fact that during metadata lookup it is necessary to read and process the entire index file(s). In contrast, HPF file can directly access the metadata of a particular file from its index file without having to process it entirely. The HPF index system uses two hash functions: file s metadata are distributed through index files by using a dynamic hash function and, for each index file, we build an order preserving perfect hash function that preserves the position of each file s metadata in the index file. The HPF design will only read the part of the index file that contains the metadata of the searched file during its access. HPF file also supports the file appending functionality after its creation. Our experiments show that HPF can be more than 40% faster file s access from the original HDFS. If we don t consider the caching effect, HPF s file access is around 179% faster than MapFile and 11294% faster than HAR file. If we consider caching effect, HPF is around 35% faster than MapFile and 105% faster than HAR file.

cs.DC↗

arXiv2019

Group-wise Correlation Stereo Network

Stereo matching estimates the disparity between a rectified image pair, which is of great importance to depth sensing, autonomous driving, and other related tasks. Previous works built cost volumes with cross-correlation or concatenation of left and right features across all disparity levels, and then a 2D or 3D convolutional neural network is utilized to regress the disparity maps. In this paper, we propose to construct the cost volume by group-wise correlation. The left features and the right features are divided into groups along the channel dimension, and correlation maps are computed among each group to obtain multiple matching cost proposals, which are then packed into a cost volume. Group-wise correlation provides efficient representations for measuring feature similarities and will not lose too much information like full correlation. It also preserves better performance when reducing parameters compared with previous methods. The 3D stacked hourglass network proposed in previous works is improved to boost the performance and decrease the inference computational cost. Experiment results show that our method outperforms previous methods on Scene Flow, KITTI 2012, and KITTI 2015 datasets. The code is available at https://github.com/xy-guo/GwcNet

cs.CV↗