arXiv subjects

Fei Song

Publications and source records attributed to Fei Song.

51 records · Page 3Linked to original sources

arXiv2021

Neural-based Modeling for Performance Tuning of Spark Data Analytics

Cloud data analytics has become an integral part of enterprise business operations for data-driven insight discovery. Performance modeling of cloud data analytics is crucial for performance tuning and other critical operations in the cloud. Traditional modeling techniques fail to adapt to the high degree of diversity in workloads and system behaviors in this domain. In this paper, we bring recent Deep Learning techniques to bear on the process of automated performance modeling of cloud data analytics, with a focus on Spark data analytics as representative workloads. At the core of our work is the notion of learning workload embeddings (with a set of desired properties) to represent fundamental computational characteristics of different jobs, which enable performance prediction when used together with job configurations that control resource allocation and other system knobs. Our work provides an in-depth study of different modeling choices that suit our requirements. Results of extensive experiments reveal the strengths and limitations of different modeling methods, as well as superior performance of our best performing method over a state-of-the-art modeling tool for cloud analytics.

cs.DC↗

arXiv2021

A stabilized nonconforming Nitsche's extended finite element method for Stokes interface problems

In this paper, a stabilized extended finite element method is proposed for Stokes interface problems on unfitted triangulation elements which do not require the interface align with the triangulation. The velocity solution and pressure solution on each side of the interface are separately expanded in the standard nonconforming piecewise linear polynomials and the piecewise constant polynomials, respectively. Harmonic weighted fluxes and arithmetic fluxes are used across the interface and cut edges (segment of the edges cut by the interface), respectively. Extra stabilization terms involving velocity and pressure are added to ensure the stable inf-sup condition. We show a priori error estimates under additional regularity hypothesis. Moreover, the errors {in energy and $L^2$ norms for velocity and the error in $L^2$ norm for pressure} are robust with respect to the viscosity {and independent of the location of the interface}. Results of numerical experiments are presented to {support} the theoretical analysis.

math.NA↗

arXiv2020

Boosting Cloud Data Analytics using Multi-Objective Optimization

Data analytics in the cloud has become an integral part of enterprise businesses. Big data analytics systems, however, still lack the ability to take user performance goals and budgetary constraints for a task, collectively referred to as task objectives, and automatically configure an analytic job to achieve these objectives. This paper presents a data analytics optimizer that can automatically determine a cluster configuration with a suitable number of cores as well as other system parameters that best meet the task objectives. At a core of our work is a principled multi-objective optimization (MOO) approach that computes a Pareto optimal set of job configurations to reveal tradeoffs between different user objectives, recommends a new job configuration that best explores such tradeoffs, and employs novel optimizations to enable such recommendations within a few seconds. We present efficient incremental algorithms based on the notion of a Progressive Frontier for realizing our MOO approach and implement them into a Spark-based prototype. Detailed experiments using benchmark workloads show that our MOO techniques provide a 2-50x speedup over existing MOO methods, while offering good coverage of the Pareto frontier. When compared to Ottertune, a state-of-the-art performance tuning system, our approach recommends configurations that yield 26\%-49\% reduction of running time of the TPCx-BB benchmark while adapting to different application preferences on multiple objectives.

cs.DB↗

arXiv2020

Probabilistic Caching for Small-Cell Networks with Terrestrial and Aerial Users

The support for aerial users has become the focus of recent 3GPP standardizations of 5G, due to their high maneuverability and flexibility for on-demand deployment. In this paper, probabilistic caching is studied for ultra-dense small-cell networks with terrestrial and aerial users, where a dynamic on-off architecture is adopted under a sophisticated path loss model incorporating both line-of-sight and non-line-of-sight transmissions. Generally, this paper focuses on the successful download probability (SDP) of user equipments (UEs) from small-cell base stations (SBSs) that cache the requested files under various caching strategies. To be more specific, the SDP is first analyzed using stochastic geometry theory, by considering the distribution of such two-tier UEs and SBSs as Homogeneous Poisson Point Processes. Second, an optimized caching strategy (OCS) is proposed to maximize the average SDP. Third, the performance limits of the average SDP are developed for the popular caching strategy (PCS) and the uniform caching strategy (UCS). Finally, the impacts of the key parameters, such as the SBS density, the cache size, the exponent of Zipf distribution and the height of aerial user, are investigated on the average SDP. The analytical results indicate that the UCS outperforms the PCS if the SBSs are sufficiently dense, while the PCS is better than the UCS if the exponent of Zipf distribution is large enough. Furthermore, the proposed OCS is superior to both the UCS and PCS.

eess.SP↗

arXiv2020

Dynamic Virtual Resource Allocation for 5G and Beyond Network Slicing

The fifth generation and beyond wireless communication will support vastly heterogeneous services and use demands such as massive connection, low latency and high transmission rate. Network slicing has been envisaged as an efficient technology to meet these diverse demands. In this paper, we propose a dynamic virtual resources allocation scheme based on the radio access network (RAN) slicing for uplink communications to ensure the quality-of-service (QoS). To maximum the weighted-sum transmission rate performance under delay constraint, formulate a joint optimization problem of subchannel allocation and power control as an infinite-horizon average-reward constrained Markov decision process (CMDP) problem. Based on the equivalent Bellman equation, the optimal control policy is first derived by the value iteration algorithm. However, the optimal policy suffers from the widely known curse-of-dimensionality problem. To address this problem, the linear value function approximation (approximate dynamic programming) is adopted. Then, the subchannel allocation Q-factor is decomposed into the per-slice Q-factor. Furthermore, the Q-factor and Lagrangian multipliers are updated by the use of an online stochastic learning algorithm. Finally, simulation results reveal that the proposed algorithm can meet the delay requirements and improve the user transmission rate compared with baseline schemes.

eess.SP↗

arXiv2019

Non-Hermitian Topological Invariants in Real Space

The topology of non-Hermitian systems is drastically shaped by the non-Hermitian skin effect, which leads to the generalized bulk-boundary correspondence and non-Bloch band theory. The essential part in formulations of bulk-boundary correspondence is a general and computable definition of topological invariants. In this paper, we introduce a construction of non-Hermitian topological invariants based directly on real-space wavefunctions, which provides a general and straightforward approach for determining non-Hermitian topology. As an illustration, we apply this formulation to several representative models of non-Hermitian systems, efficiently obtaining their topological invariants in the presence of non-Hermitian skin effect. Our formulation also provides a dual picture of the non-Bloch band theory based on the generalized Brillouin zone, offering a unique perspective of bulk-boundary correspondence.

cond-mat.mes-hall↗

arXiv2019

Non-Hermitian skin effect and chiral damping in open quantum systems

One of the unique features of non-Hermitian Hamiltonians is the non-Hermitian skin effect, namely that the eigenstates are exponentially localized at the boundary of the system. For open quantum systems, a short-time evolution can often be well described by the effective non-Hermitian Hamiltonians, while long-time dynamics calls for the Lindblad master equations, in which the Liouvillian superoperators generate time evolution. In this Letter, we find that Liouvillian superoperators can exhibit the non-Hermitian skin effect, and uncover its unexpected physical consequences. It is shown that the non-Hermitian skin effect dramatically shapes the long-time dynamics, such that the damping in a class of open quantum systems is algebraic under periodic boundary condition but exponential under open boundary condition. Moreover, the non-Hermitian skin effect and non-Bloch bands cause a chiral damping with a sharp wavefront. These phenomena are beyond the effective non-Hermitian Hamiltonians; instead, they belong to the non-Hermitian physics of full-fledged open quantum dynamics.

cond-mat.quant-gas↗

arXiv2018

Non-Hermitian Chern bands

The relation between chiral edge modes and bulk Chern numbers of quantum Hall insulators is a paradigmatic example of bulk-boundary correspondence. We show that the chiral edge modes are not strictly tied to the Chern numbers defined by a non-Hermitian Bloch Hamiltonian. This breakdown of conventional bulk-boundary correspondence stems from the non-Bloch-wave behavior of eigenstates (non-Hermitian skin effect), which generates pronounced deviations of phase diagrams from the Bloch theory. We introduce non-Bloch Chern numbers that faithfully predict the numbers of chiral edge modes. The theory is backed up by the open-boundary energy spectra, dynamics, and phase diagram of representative lattice models. Our results highlight a unique feature of non-Hermitian bands and suggest a non-Bloch framework to characterize their topology.

cond-mat.mes-hall↗

arXiv2018

Majorana corner modes in a high-temperature platform

We introduce two-dimensional topological insulators in proximity to high-temperature cuprate or iron-based superconductors as high-temperature platforms of Majorana Kramers pairs of zero modes. The proximity-induced pairing at the helical edge state of the topological insulator serves as a Dirac mass, whose sign changes at the sample corner because of the pairing symmetry of high-$T_c$ superconductors. This sign changing naturally creates at each corner a pair of Majorana zero modes protected by time-reversal symmetry. Conceptually, this is a topologically trivial superconductor-based approach for Majorana zero modes. We provide quantitative criteria and suggest candidate materials for this proposal.

cond-mat.mes-hall↗

arXiv2015

Cellular-Base-Station Assisted Device-to-Device Communications in TV White Space

This paper presents a systematic approach to exploit TV white space (TVWS) for device-to-device (D2D) communications with the aid of the existing cellular infrastructure. The goal is to build a location-specific TVWS database, which provides a look-up table service for any D2D link to determine its maximum permitted emission power (MPEP) in an unlicensed digital TV (DTV) band. To achieve this goal, the idea of mobile crowd sensing is firstly introduced to collect active spectrum measurements from massive personal mobile devices. Considering the incompleteness of crowd measurements, we formulate the problem of unknown measurements recovery as a matrix completion problem and apply a powerful fixed point continuation algorithm to reconstruct the unknown elements from the known elements. By joint exploitation of the big spectrum data in its vicinity, each cellular base station further implements a nonlinear support vector machine algorithm to perform irregular coverage boundary detection of a licensed DTV transmitter. With the knowledge of the detected coverage boundary, an opportunistic spatial reuse algorithm is developed for each D2D link to determine its MPEP. Simulation results show that the proposed approach can successfully enable D2D communications in TVWS while satisfying the interference constraint from the licensed DTV services. In addition, to our best knowledge, this is the first try to explore and exploit TVWS inside the DTV protection region resulted from the shadowing effect. Potential application scenarios include communications between internet of vehicles in the underground parking, D2D communications in hotspots such as subway, game stadiums, and airports, etc.

cs.NI↗

arXiv2014

Fidelity susceptibility and quantum Fisher information for density operators with arbitrary ranks

Taking into account the density matrices with non-full ranks, we show that the fidelity susceptibility is determined by the support of the density matrix. Combining with the corresponding expression of the quantum Fisher information, we rigorously prove that the fidelity susceptibility is proportional to the quantum Fisher information. As this proof can be naturally extended to the full rank case, this proportional relation is generally established for density matrices with arbitrary ranks. Furthermore, we give an analytical expression of the quantum Fisher information matrix, and show that the quantum Fisher information matrix can also be represented in the density matrix's support.

quant-ph↗

arXiv2013

Modeling of Multipath Transport

In this paper, we propose a model for evaluating the transmission performance of multipath transport. Previous researches focused exclusively on single pair users in simple scenarios. The distinct perspective in this paper is to build models for analyzing the performance when multipath transport is used in the entire network scope. We illustrate the influences on the transmission performance caused by the variation of network topologies, the services' arrival rate, the services' size and other parameters. We demonstrate through simulation that multipath transport could conditionally increase the throughput than single-path transport. And it has the capability to support higher services' arrival rate in various network topologies. And higher multi-parent probability will be beneficial for multipath transport to take its advantages.

cs.NI↗

arXiv2013

Probabilistic Topic and Syntax Modeling with Part-of-Speech LDA

This article presents a probabilistic generative model for text based on semantic topics and syntactic classes called Part-of-Speech LDA (POSLDA). POSLDA simultaneously uncovers short-range syntactic patterns (syntax) and long-range semantic patterns (topics) that exist in document collections. This results in word distributions that are specific to both topics (sports, education, ...) and parts-of-speech (nouns, verbs, ...). For example, multinomial distributions over words are uncovered that can be understood as "nouns about weather" or "verbs about law". We describe the model and an approximate inference algorithm and then demonstrate the quality of the learned topics both qualitatively and quantitatively. Then, we discuss an NLP application where the output of POSLDA can lead to strong improvements in quality: unsupervised part-of-speech tagging. We describe algorithms for this task that make use of POSLDA-learned distributions that result in improved performance beyond the state of the art.

cs.CL↗

arXiv2012

A Practical Regularity Partitioning Algorithm and its Applications in Clustering

In this paper we introduce a new clustering technique called Regularity Clustering. This new technique is based on the practical variants of the two constructive versions of the Regularity Lemma, a very useful tool in graph theory. The lemma claims that every graph can be partitioned into pseudo-random graphs. While the Regularity Lemma has become very important in proving theoretical results, it has no direct practical applications so far. An important reason for this lack of practical applications is that the graph under consideration has to be astronomically large. This requirement makes its application restrictive in practice where graphs typically are much smaller. In this paper we propose modifications of the constructive versions of the Regularity Lemma that work for smaller graphs as well. We call this the Practical Regularity partitioning algorithm. The partition obtained by this is used to build the reduced graph which can be viewed as a compressed representation of the original graph. Then we apply a pairwise clustering method such as spectral clustering on this reduced graph to get a clustering of the original graph that we call Regularity Clustering. We present results of using Regularity Clustering on a number of benchmark datasets and compare them with standard clustering techniques, such as $k$-means and spectral clustering. These empirical results are very encouraging. Thus in this paper we report an attempt to harness the power of the Regularity Lemma for real-world applications.

math.CO↗

arXiv2012

Robust Surface Doping of Bi$_2$Se$_3$ by Rb Intercalation

Rubidium adsorption on the surface of the topological insulator Bi$_2$Se$_3$ is found to induce a strong downward band bending, leading to the appearance of a quantum-confined two dimensional electron gas states (2DEGs) in the conduction band. The 2DEGs shows a strong Rashba-type spin-orbit splitting, and it has previously been pointed out that this has relevance to nano-scale spintronics devices. The adsorption of Rb atoms, on the other hand, renders the surface very reactive and exposure to oxygen leads to a rapid degrading of the 2DEGs. We show that intercalating the Rb atoms, presumably into the van der Waals gaps in the quintuple layer structure of Bi$_2$Se$_3$, drastically reduces the surface reactivity while not affecting the promising electronic structure. The intercalation process is observed above room temperature and accelerated with increasing initial Rb coverage, an effect that is ascribed to the Coulomb interaction between the charged Rb ions. Coulomb repulsion is also thought to be responsible for a uniform distribution of Rb on the surface.

cond-mat.mtrl-sci↗