arXiv subjects

Cheng Yang

Publications and source records attributed to Cheng Yang.

257 records · Page 15Linked to original sources

arXiv2015

Comprehend DeepWalk as Matrix Factorization

Word2vec, as an efficient tool for learning vector representation of words has shown its effectiveness in many natural language processing tasks. Mikolov et al. issued Skip-Gram and Negative Sampling model for developing this toolbox. Perozzi et al. introduced the Skip-Gram model into the study of social network for the first time, and designed an algorithm named DeepWalk for learning node embedding on a graph. We prove that the DeepWalk algorithm is actually factoring a matrix M where each entry M_{ij} is logarithm of the average probability that node i randomly walks to node j in fix steps.

cs.LG↗

arXiv2014

Real-time Topic-aware Influence Maximization Using Preprocessing

Influence maximization is the task of finding a set of seed nodes in a social network such that the influence spread of these seed nodes based on certain influence diffusion model is maximized. Topic-aware influence diffusion models have been recently proposed to address the issue that influence between a pair of users are often topic-dependent and information, ideas, innovations etc. being propagated in networks (referred collectively as items in this paper) are typically mixtures of topics. In this paper, we focus on the topic-aware influence maximization task. In particular, we study preprocessing methods for these topics to avoid redoing influence maximization for each item from scratch. We explore two preprocessing algorithms with theoretical justifications. Our empirical results on data obtained in a couple of existing studies demonstrate that one of our algorithms stands out as a strong candidate providing microsecond online response time and competitive influence spread, with reasonable preprocessing effort.

cs.SI↗

arXiv2010

Strong and Tunable Nonlinear Optomechanical Coupling in a Low-Loss System

A major goal in optomechanics is to observe and control quantum behavior in a system consisting of a mechanical resonator coupled to an optical cavity. Work towards this goal has focused on increasing the strength of the coupling between the mechanical and optical degrees of freedom; however, the form of this coupling is crucial in determining which phenomena can be observed in such a system. Here we demonstrate that avoided crossings in the spectrum of an optical cavity containing a flexible dielectric membrane allow us to realize several different forms of the optomechanical coupling. These include cavity detunings that are (to lowest order) linear, quadratic, or quartic in the membrane's displacement, and a cavity finesse that is linear in (or independent of) the membrane's displacement. All these couplings are realized in a single device with extremely low optical loss and can be tuned over a wide range in situ; in particular, we find that the quadratic coupling can be increased three orders of magnitude beyond previous devices. As a result of these advances, the device presented here should be capable of demonstrating the quantization of the membrane's mechanical energy.

quant-ph↗

arXiv2008

Improved "Position Squared" Readout of a Mechanical Resonator in an Optical Cavity Using Degenerate Optical Modes

Optomechanical devices in which a flexible SiN membrane is placed inside an optical cavity allow for very high finesse and mechanical quality factor in a single device. They also provide fundamentally new functionality: the cavity detuning can be a quadratic function of membrane position. This enables a measurement of "position squared" ($x^2$) and in principle a QND phonon number readout of the membrane. However, the readout achieved using a single transverse cavity mode is not sensitive enough to observe quantum jumps between phonon Fock states. Here we demonstrate an $x^2$-sensitivity that is orders of magnitude stronger using two transverse cavity modes that are nearly degenerate. We derive a first-order perturbation theory to describe the interactions between nearly-degenerate cavity modes and achieve good agreement with our measurements using realistic parameters. We also demonstrate theoretically that the $x^2$-coupling should be easily tunable over a wide range.

quant-ph↗

arXiv2003

Dynamic Point-Formation in Dielectric Fluids

We use boundary-integral methods to compute the time-dependent deformation of a drop of dielectric fluid immersed in another dielectric fluid in a uniform electric field E. Steady state theory predicts, when the permittivity ratio, β, is large enough, a conical interface can exist at two cone angles, with θ_<(β) stable and θ_>(β) unstable. Our numerical evidence instead shows a dynamical process which produces a cone-formation and a transient finite-time singularity, when E and βare above their critical values. Based on a scaling analysis of the electric stress and the fluid motion, we are able to apply approximate boundary conditions to compute the evolution of the tip region. We find in our non-equilibrium case where the electric stress is substantially larger than the surface tension, the ratio of the electric stress to the surface tension in the newly-grown cone region can converge to a βdependent value, α_c(β)>1, while the cone angle converges to θ_<(β). This new dynamical solution is self-similar.

physics.flu-dyn↗