Searcharxiv⌕ Search

arXiv subjects

Zheng Sun

Publications and source records attributed to Zheng Sun.

At least 127 records · Page 7Linked to original sources
arXiv2017

A discontinuous Galerkin method for nonlinear parabolic equations and gradient flow problems with interaction potentials

We consider a class of time dependent second order partial differential equations governed by a decaying entropy. The solution usually corresponds to a density distribution, hence positivity (non-negativity) is expected. This class of problems covers important cases such as Fokker-Planck type equations and aggregation models, which have been studied intensively in the past decades. In this paper, we design a high order discontinuous Galerkin method for such problems. If the interaction potential is not involved, or the interaction is defined by a smooth kernel, our semi-discrete scheme admits an entropy inequality on the discrete level. Furthermore, by applying the positivity-preserving limiter, our fully discretized scheme produces non-negative solutions for all cases under a time step constraint. Our method also applies to two dimensional problems on Cartesian meshes. Numerical examples are given to confirm the high order accuracy for smooth test cases and to demonstrate the effectiveness for preserving long time asymptotics.

math.NA↗
arXiv2016

Deep Recurrent Convolutional Neural Network: Improving Performance For Speech Recognition

A deep learning approach has been widely applied in sequence modeling problems. In terms of automatic speech recognition (ASR), its performance has significantly been improved by increasing large speech corpus and deeper neural network. Especially, recurrent neural network and deep convolutional neural network have been applied in ASR successfully. Given the arising problem of training speed, we build a novel deep recurrent convolutional network for acoustic modeling and then apply deep residual learning to it. Our experiments show that it has not only faster convergence speed but better recognition accuracy over traditional deep convolutional recurrent network. In the experiments, we compare the convergence speed of our novel deep recurrent convolutional networks and traditional deep convolutional recurrent networks. With faster convergence speed, our novel deep recurrent convolutional networks can reach the comparable performance. We further show that applying deep residual learning can boost the convergence speed of our novel deep recurret convolutional networks. Finally, we evaluate all our experimental networks by phoneme error rate (PER) with our proposed bidirectional statistical n-gram language model. Our evaluation results show that our newly proposed deep recurrent convolutional network applied with deep residual learning can reach the best PER of 17.33\% with the fastest convergence speed on TIMIT database. The outstanding performance of our novel deep recurrent convolutional neural network with deep residual learning indicates that it can be potentially adopted in other sequential problems.

cs.CL↗
arXiv2016

Composing Music with Grammar Argumented Neural Networks and Note-Level Encoding

Creating aesthetically pleasing pieces of art, including music, has been a long-term goal for artificial intelligence research. Despite recent successes of long-short term memory (LSTM) recurrent neural networks (RNNs) in sequential learning, LSTM neural networks have not, by themselves, been able to generate natural-sounding music conforming to music theory. To transcend this inadequacy, we put forward a novel method for music composition that combines the LSTM with Grammars motivated by music theory. The main tenets of music theory are encoded as grammar argumented (GA) filters on the training data, such that the machine can be trained to generate music inheriting the naturalness of human-composed pieces from the original dataset while adhering to the rules of music theory. Unlike previous approaches, pitches and durations are encoded as one semantic entity, which we refer to as note-level encoding. This allows easy implementation of music theory grammars, as well as closer emulation of the thinking pattern of a musician. Although the GA rules are applied to the training data and never directly to the LSTM music generation, our machine still composes music that possess high incidences of diatonic scale notes, small pitch intervals and chords, in deference to music theory.

cs.LG↗
arXiv2015

The Renormalizable Three-Term Polynomial Inflation with Large Tensor-to-Scalar Ratio

We systematically study the renormalizable three-term polynomial inflation in the supersymmetric and non-supersymmetric models. The supersymmetric inflaton potentials can be realized in supergravity theory, and only have two independent parameters. We show that the general renormalizable supergravity model is equivalent to one kind of our supersymmetric models. We find that the spectral index and tensor-to-scalar ratio can be consistent with the Planck and BICEP2 results, but the running of spectral index is always out of the $2σ$ range. If we do not consider the BICEP2 experiment, these inflationary models can be highly consistent with the Planck observations and saturate its upper bound on the tensor-to-scalar ratio ($r \le 0.11$). Thus, our models can be tested at the future Planck and QUBIC experiments.

hep-ph↗
arXiv2015

Weighted Elastic Net Penalized Mean-Variance Portfolio Design and Computation

It is well known that the out-of-sample performance of Markowitz's mean-variance portfolio criterion can be negatively affected by estimation errors in the mean and covariance. In this paper we address the problem by regularizing the mean-variance objective function with a weighted elastic net penalty. We show that the use of this penalty can be motivated by a robust reformulation of the mean-variance criterion that directly accounts for parameter uncertainty. With this interpretation of the weighted elastic net penalty we derive data driven techniques for calibrating the weighting parameters based on the level of uncertainty in the parameter estimates. We test our proposed technique on US stock return data and our results show that the calibrated weighted elastic net penalized portfolio outperforms both the unpenalized portfolio and uniformly weighted elastic net penalized portfolio. This paper also introduces a novel Adaptive Support Split-Bregman approach which leverages the sparse nature of $\ell_{1}$ penalized portfolios to efficiently compute a solution of our proposed portfolio criterion. Numerical results show that this modification to the Split-Bregman algorithm results in significant improvements in computational speed compared with other techniques.

q-fin.PM↗
arXiv2014

Preferred hierarchy scales from the product landscape

The product landscape method has been recently proposed to solve hierarchy problems such as the cosmological constant problem. We suggest that the parameter distribution on logarithmic scales should be used as a benchmark for hierarchy, and the preferred hierarchy scales can be obtained from the distribution peak. It is shown that generating hierarchy from purely product distribution is very inefficient. To achieve a reasonably acceptable efficiency, other effects such as accumulation of weak hierarchy in the effective theory should be incorporated.

hep-th↗
arXiv2014

Strong light-matter coupling in two-dimensional atomic crystals

Two dimensional (2D) atomic crystals of graphene, and transition metal dichalcogenides have emerged as a class of materials that show strong light-matter interaction. This interaction can be further controlled by embedding such materials into optical microcavities. When the interaction is engineered to be stronger than the dissipation of light and matter entities, one approaches the strong coupling regime resulting in the formation of half-light half-matter bosonic quasiparticles called microcavity polaritons. Here we report the evidence of strong light-matter coupling and formation of microcavity polaritons in a two dimensional atomic crystal of molybdenum disulphide (MoS2) embedded inside a dielectric microcavity at room temperature. A Rabi splitting of 46 meV and highly directional emission is observed from the MoS2 microcavity owing to the coupling between the 2D excitons and the cavity photons. Realizing strong coupling effects at room temperature in a disorder free potential landscape is central to the development of practical polaritonic circuits and switches.

physics.optics↗
arXiv2014

The Nelson-Seiberg theorem revised

The well-accepted Nelson-Seiberg theorem relates R-symmetries to supersymmetry (SUSY) breaking vacua, and provides a guideline for SUSY model building which is the most promising physics beyond the Standard Model. In the case of Wess-Zumino models with perturbative superpotentials, we revise the theorem to a combined necessary and sufficient condition for SUSY breaking which can be easily checked before solving the vacuum. The revised theorem provides a powerful tool to construct either SUSY breaking or SUSY vacua, and offers many practicable applications in low energy SUSY model building and string phenomenology.

hep-th↗
arXiv2012

Effect of Monolayer Thickness Fluctuations on Coherent Exciton Coupling in Single Quantum Wells

Monolayer fluctuations in the thickness of a semiconductor quantum well (QW) lead to three types of excitons, located in the narrower, average and thicker regions of the QW, which are clearly resolved in optical spectra. Whether or not these excitons are coherently coupled via Coulomb interactions is a long-standing debate. We demonstrate that different types of disorder in QWs distinctly affects the coherent coupling and that the coupling strength can be quantitatively measured using optical two-dimensional Fourier transform spectroscopy. We prove experimentally and theoretically that in narrow quantum wells the coherent coupling occurs predominantly between excitons residing in the disorder-free areas of the QWs and those residing in the plateau-type disorder. In contrast, excitons localized in the fault-type disorder potentials do not coherently couple to other excitons.

cond-mat.mes-hall↗
arXiv2012

Low energy supersymmetry from R-symmetries

In a generic setting of Wess-Zumino models, we prove that the existence of a supersymmetric vacuum with a vanishing superpotential can be a consequence of a continuous or discrete R-symmetry when invariant fields are not less than fields transforming in the same way as the superpotential under the R-symmetry. The realization in string theory is discussed. We show that a rich landscape of low energy supersymmetric vacua can be found in the Type IIB flux compactification setup ready for the KKLT construction of de Sitter vacua in string theory.

hep-th↗
arXiv2011

Vacuum statistics and parameter tuning for F-term supersymmetry breaking

We carry out a model-independent EFT method study on the vacuum statistics of general F-term SUSY breaking models. Assuming a smooth distribution of Lagrangian parameters, SUSY breaking vacua are rare in global SUSY models with a canonical Kähler potential, and have a peaked distribution near the cut-off of the SUSY breaking scale in both global SUSY and SUGRA models with a general Kähler potential. After including different mass scales in the Lagrangian, we compare the total number of SUSY and non-SUSY vacua and estimate quantitatively the rareness of SUSY breaking. The EFT method provides a general view to the amount of parameter tuning needed for a metastable SUSY breaking vacuum. The tuning also indicates the importance of R-symmetries in SUSY breaking even for metastable SUSY breaking.

hep-th↗
arXiv2011

Spin Selective Purcell Effect in a Quantum Dot Microcavity System

We demonstrate the selective coupling of a single quantum dot exciton spin state with the cavity mode in a quantum dot-micropillar cavity system. By tuning an external magnetic field, the Zeeman splitted exciton spin states coupled differently with the cavity due to field manipulated energy detuning. We found a 26 times increase in the emission intensity of spin-up exciton state with respect to spin-down exciton state at resonance due to Purcell effect, which gives rise to the selective enhancement of light emission with the circular polarization degree up to 93%. A four-level rate equation model is developed and quantitatively agrees well with our experimental data. Our results pave the way for the realization of future quantum light sources and the quantum information processing applications.

physics.optics↗
arXiv2011

Continuous degeneracy of non-supersymmetric vacua

In global supersymmetric Wess-Zumino models with minimal Kähler potentials, F-type supersymmetry breaking always yields instability or continuous degeneracy of non-supersymmetric vacua. As a generalization of the original O'Raifeartaigh's result, the existence of instability or degeneracy is true to any higher order corrections at tree level for models even with non-renormalizable superpotentials. The degeneracy generically coincides the R-axion direction under some assumptions of R-charge assignment, but generally requires neither R-symmetries nor any assumption of generic superpotentials. The result also confirms the well-known fact that tree level supersymmetry breaking is a very rare occurrence in global supersymmetric theories with minimal Kähler potentials. The implication for effective field theory method in the landscape is discussed and we point out that choosing models with minimal Kähler potentials may result in unexpected answers to the vacuum statistics. Supergravity theories or theories with non-minimal Kähler potentials in general do not suffer from the existence of instability or degeneracy. But very strong gauge dynamics or small compactification dimension reduces the Kähler potential from non-minimal to minimal, and gravity decoupling limit reduces supergravity to global supersymmetry. Instability or degeneracy may appear in these limits. Away from these limits, a large number of non-SUSY vacua may still be found in an intermediate region.

hep-th↗
arXiv2011

Tree level spontaneous R-symmetry breaking in O'Raifeartaigh models

We show that in O'Raifeartaigh models of spontaneous supersymmetry breaking, R-symmetries can be broken by non-zero values of fields at tree level, rather than by vacuum expectation values of pseudomoduli at loop level. As a complement of the recent result by Shih, we show that there must be a field in the theory with R-charge different from zero and two in order for R-symmetry breaking to occur, no matter whether the breaking happens at tree or loop level. We review the example by CDFM, and construct two types of tree level R-symmetry breaking models with a wide range of parameters and free of runaway problem. And the R-symmetry is broken everywhere on the pseudomoduli space in these models. This provides a rich set of candidates for SUSY model building and phenomenology.

hep-th↗
arXiv2008

Criteria on Utility Designing of Convex Optimization in FDMA Networks

In this paper, we investigate the network utility maximization problem in FDMA systems. We summarize with a suite of criteria on designing utility functions so as to achieve the global optimization convex. After proposing the general form of the utility functions, we present examples of commonly used utility function forms that are consistent with the criteria proposed in this paper, which include the well-known proportional fairness function and the sigmoidal-like functions. In the second part of this paper, we use numerical results to demonstrate a case study based on the criteria mentioned above, which deals with the subcarrier scheduling problem with dynamic rate allocation in FDMA system.

cs.NI↗
arXiv2008

Performance Comparison of Cooperative and Distributed Spectrum Sensing in Cognitive Radio

In this paper, we compare the performances of cooperative and distributed spectrum sensing in wireless sensor networks. After introducing the basic problem, we describe two strategies: 1) a cooperative sensing strategy, which takes advantage of cooperation diversity gain to increase probability of detection and 2) a distributed sensing strategy, which by passing the results in an inter-node manner increases energy efficiency and fairness among nodes. Then, we compare the performances of the strategies in terms of three criteria: agility, energy efficiency, and robustness against SNR changes, and summarize the comparison. It shows that: 1) the non-cooperative strategy has the best fairness of energy consumption, 2) the cooperative strategy leads to the best agility, and 3) the distributed strategy leads to the lowest energy consumption and the best robustness against SNR changes.

cs.NI↗
arXiv2008

A Heuristic Scheduling Scheme in Multiuser OFDMA Networks

Conventional heterogeneous-traffic scheduling schemes utilize zero-delay constraint for real-time services, which aims to minimize the average packet delay among real-time users. However, in light or moderate load networks this strategy is unnecessary and leads to low data throughput for non-real-time users. In this paper, we propose a heuristic scheduling scheme to solve this problem. The scheme measures and assigns scheduling priorities to both real-time and non-real-time users, and schedules the radio resources for the two user classes simultaneously. Simulation results show that the proposed scheme efficiently handles the heterogeneous-traffic scheduling with diverse QoS requirements and alleviates the unfairness between real-time and non-real-time services under various traffic loads.

cs.NI↗