Searcharxiv⌕ Search

arXiv subjects

Yi Zhou

Publications and source records attributed to Yi Zhou.

At least 487 records · Page 27Linked to original sources
arXiv2019

Distributed SGD Generalizes Well Under Asynchrony

The performance of fully synchronized distributed systems has faced a bottleneck due to the big data trend, under which asynchronous distributed systems are becoming a major popularity due to their powerful scalability. In this paper, we study the generalization performance of stochastic gradient descent (SGD) on a distributed asynchronous system. The system consists of multiple worker machines that compute stochastic gradients which are further sent to and aggregated on a common parameter server to update the variables, and the communication in the system suffers from possible delays. Under the algorithm stability framework, we prove that distributed asynchronous SGD generalizes well given enough data samples in the training optimization. In particular, our results suggest to reduce the learning rate as we allow more asynchrony in the distributed system. Such adaptive learning rate strategy improves the stability of the distributed algorithm and reduces the corresponding generalization error. Then, we confirm our theoretical findings via numerical experiments.

stat.ML↗
arXiv2019

Extensive beam test study of prototype MRPCs for the T0 detector at the CSR external-target experiment

The CSR External-target Experiment (CEE) will be the first large-scale nuclear physics experiment device at the Cooling Storage Ring (CSR) of the Heavy-Ion Research Facility in Lanzhou (HIRFL) in China. A new T0 detector has been proposed to measure the multiplicity, angular distribution and timing information of charged particles produced in heavy-ion collisions at the target region. Multi-gap resistive plate chamber (MRPC) technology was chosen as part of the construction of the T0 detector, which provides precision event collision times (T0) and collision geometry information. The prototype was tested with hadron and heavy-ion beams to study its performance. By comparing the experimental results with a Monte Carlo simulation, the time resolution of the MRPCs are found to be $\sim$ 50 ps or better. The timing performance of the T0 detector, including both detector and readout electronics, we found to fulfil the requirements of the CEE.

physics.ins-det↗
arXiv2019

On some conjectures by Lu and Wenzel

In order to give a unified generalization of the BW inequality and the DDVV inequality, Lu and Wenzel proposed three Conjectures 1, 2, 3 and an open Question 1 in 2016. In this paper we discuss further these conjectures and put forward several new conjectures which will be shown equivalent to Conjecture 2. In particular, we prove Conjecture 2 and hence all conjectures in some special cases. For Conjecture 3, we obtain a bigger upper bound $2+\sqrt{10}/2$, and we also give a weaker answer for the more general Question 1. In addition, we obtain some new simple proofs of the complex BW inequality and the condition for equality.

math.DG↗
arXiv2019

A Hybrid Approach to Privacy-Preserving Federated Learning

Federated learning facilitates the collaborative training of models without the sharing of raw data. However, recent attacks demonstrate that simply maintaining data locality during training processes does not provide sufficient privacy guarantees. Rather, we need a federated learning system capable of preventing inference over both the messages exchanged during training and the final trained model while ensuring the resulting model also has acceptable predictive accuracy. Existing federated learning approaches either use secure multiparty computation (SMC) which is vulnerable to inference or differential privacy which can lead to low accuracy given a large number of parties with relatively small amounts of data each. In this paper, we present an alternative approach that utilizes both differential privacy and SMC to balance these trade-offs. Combining differential privacy with secure multiparty computation enables us to reduce the growth of noise injection as the number of parties increases without sacrificing privacy while maintaining a pre-defined rate of trust. Our system is therefore a scalable approach that protects against inference threats and produces models with high accuracy. Additionally, our system can be used to train a variety of machine learning models, which we validate with experimental results on 3 different machine learning algorithms. Our experiments demonstrate that our approach out-performs state of the art solutions.

cs.LG↗
arXiv2019

Hidden $SU(2)$ Symmetries, the Symmetry Hierarchy and the Emergent Eight-Fold Way in Spin-1 Quantum Magnets

The largest allowed symmetry in a spin-1 quantum system is an $SU(3)$ symmetry rather than the $SO(3)$ spin rotation. In this work, we reveal some $SU(2)$ symmetries as subgroups of $SU(3)$ that, to the best of our knowledge, have not previously been recognized. Then, we construct $SU(2)$ symmetric Hamiltonians and explore the ground-state phase diagram in accordance with the $SU(3)\supset SU(2)\times U(1)$ symmetry hierarchy. It is natural to treat the eight generators of the $SU(3)$ symmetry on an equal footing; this approach is called the eight-fold way. We find that the spin spectral functions and spin quadrupole spectral functions share the same structure, provided that the elementary excitations are flavor waves at low energies, which serves as a clue to the eight-fold way. An emergent $S=1/2$ quantum spin liquid is proposed to coexist with gapful spin nematic order in one of the ground states. In analogy to quantum chromodynamics, we find the gap relation for hydrodynamic modes in quantum spin-orbital liquid states, which is nothing but the Gell-Mann-Okubo formula.

cond-mat.str-el↗
arXiv2019

High Speed Mid-Infrared Interband Cascade Photodetector Based on InAs/GaSb Type-II Superlattice

High speed mid-wave infrared (MWIR) photodetectors have applications in the areas such as free space optical communication and frequency comb spectroscopy. However, most of the research on the MWIR photodetectors is focused on how to increase the quantum efficiency and reduce the dark current, in order to improve the detectivity (D*), and the 3dB bandwidth performance of the corresponding MWIR photodetectors is still not fully studied. In this work, we report and characterize a MWIR interband cascade photodetector based on InAs/GaSb type-II superlattice with a 50% cutoff wavelength at ~5.3 um at 300 K. The 3 dB cutoff frequency is 2.4 GHz at 300 K, for a 40 μm circular diameter device under -5 V applied bias. Limitations on the detector high speed performance are also discussed.High speed mid-wave infrared (MWIR) photodetectors have applications in the areas such as free space optical communication and frequency comb spectroscopy. However, most of the research on the MWIR photodetectors is focused on how to increase the quantum efficiency and reduce the dark current, in order to improve the detectivity (D*), and the 3dB bandwidth performance of the corresponding MWIR photodetectors is still not fully studied. In this work, we report and characterize a MWIR interband cascade photodetector based on InAs/GaSb type-II superlattice with a 50% cutoff wavelength at ~5.3 um at 300 K. The 3 dB cutoff frequency is 2.4 GHz at 300 K, for a 40 um circular diameter device under -5 V applied bias. Limitations on the detector high speed performance are also discussed.

physics.app-ph↗
arXiv2019

Cubic Regularization with Momentum for Nonconvex Optimization

Momentum is a popular technique to accelerate the convergence in practical training, and its impact on convergence guarantee has been well-studied for first-order algorithms. However, such a successful acceleration technique has not yet been proposed for second-order algorithms in nonconvex optimization.In this paper, we apply the momentum scheme to cubic regularized (CR) Newton's method and explore the potential for acceleration. Our numerical experiments on various nonconvex optimization problems demonstrate that the momentum scheme can substantially facilitate the convergence of cubic regularization, and perform even better than the Nesterov's acceleration scheme for CR. Theoretically, we prove that CR under momentum achieves the best possible convergence rate to a second-order stationary point for nonconvex optimization. Moreover, we study the proposed algorithm for solving problems satisfying an error bound condition and establish a local quadratic convergence rate. Then, particularly for finite-sum problems, we show that the proposed algorithm can allow computational inexactness that reduces the overall sample complexity without degrading the convergence rate.

math.OC↗
arXiv2019

Global smooth solutions of MHD equations with large data

In this paper, we establish the global existence of smooth solutions of the three-dimensional MHD system for a class of large initial data. Both the initial velocity and magnetic field can be arbitrarily large in the critical norm.

math.AP↗
arXiv2019

Electric-field noise from thermally-activated fluctuators in a surface ion trap

We probe electric-field noise near the metal surface of an ion trap chip in a previously unexplored high-temperature regime. We observe a non-trivial temperature dependence with the noise amplitude at 1-MHz frequency saturating around 500~K. Measurements of the noise spectrum reveal a $1/f^{α\approx1}$-dependence and a small decrease in $α$ between low and high temperatures. This behavior can be explained by considering noise from a distribution of thermally-activated two-level fluctuators with activation energies between 0.35~eV and 0.65~eV. Processes in this energy range may be relevant to understanding electric-field noise in ion traps; for example defect motion in the solid state and surface adsorbate binding energies. Studying these processes may aid in identifying the origin of excess electric-field noise in ion traps -- a major source of ion motional decoherence limiting the performance of surface traps as quantum devices.

quant-ph↗
arXiv2019

Elastic Neural Networks for Classification

In this work we propose a framework for improving the performance of any deep neural network that may suffer from vanishing gradients. To address the vanishing gradient issue, we study a framework, where we insert an intermediate output branch after each layer in the computational graph and use the corresponding prediction loss for feeding the gradient to the early layers. The framework - which we name Elastic network - is tested with several well-known networks on CIFAR10 and CIFAR100 datasets, and the experimental results show that the proposed framework improves the accuracy on both shallow networks (e.g., MobileNet) and deep convolutional neural networks (e.g., DenseNet). We also identify the types of networks where the framework does not improve the performance and discuss the reasons. Finally, as a side product, the computational complexity of the resulting networks can be adjusted in an elastic manner by selecting the output branch according to current computational budget.

cs.CV↗
arXiv2019

Topological Superconductivity Intertwined with Broken Symmetries

Recently the superconductor and topological semimetal PbTaSe$_2$ was experimentally found to exhibit surface-only lattice rotational symmetry breaking below $T_c$. We exploit the Ginzburg-Landau free energy and propose a microscopic two-channel model to study possible superconducting states on the surface of PbTaSe$_2$. We identify two types of topological superconducting states. One is time-reversal invariant and preserves the lattice hexagonal symmetry while the other breaks both symmetries. We find that such time-reversal symmetry breaking is unavoidable for a superconducting state in a two dimensional irreducible representation of crystal point group in a system where the spatial inversion symmetry is broken and the strong spin-orbit coupling is present. Our findings will guide the search for topological chiral superconductors.

cond-mat.supr-con↗
arXiv2019

Evidence for nematic superconductivity of topological surface states in PbTaSe2

Spontaneous symmetry breaking has been a paradigm to describe the phase transitions in condensed matter physics. In addition to the continuous electromagnetic gauge symmetry, an unconventional superconductor can break discrete symmetries simultaneously, such as time reversal and lattice rotational symmetry. In this work we report a characteristic in-plane 2-fold behaviour of the resistive upper critical field and point-contact spectra on the superconducting semimetal PbTaSe2 with topological nodal-rings, despite its hexagonal lattice symmetry (or D_3h in bulk while C_3v on surface, to be precise). However, we do not observe any lattice rotational symmetry breaking signal from field-angle-dependent specific heat. It is worth noting that such surface-only electronic nematicity is in sharp contrast to the observation in the topological superconductor candidate, CuxBi2Se3, where the nematicity occurs in various bulk measurements. In combination with theory, superconducting nematicity is likely to emerge from the topological surface states of PbTaSe2, rather than the proximity effect. The issue of time reversal symmetry breaking is also addressed. Thus, our results on PbTaSe2 shed new light on possible routes to realize nematic superconductivity with nontrivial topology.

cond-mat.supr-con↗
arXiv2019

Global regularity for Einstein-Klein-Gordon system with $U(1) \times \mathbb{R}$ isometry group, I

This is the first of the two papers devoted to the study of global regularity of the 3+1 dimensional Einstein-Klein-Gordon system with a $U(1)\times \mathbb{R}$ isometry group. In this first part, we reduce the Cauchy problem of the Einstein-Klein-Gordon system to a 2+1 dimensional system. Then, we will give energy estimates and construct the null coordinate system, under which we finally show that the first possible singularity can only occur at the axis.

math.AP↗
arXiv2019

Global regularity for Einstein-Klein-Gordon system with $U(1) \times \mathbb{R}$ isometry group, II

This paper is devoted to the study of the global existence of smooth solutions for the 3+1 dimensional Einstein-Klein-Gordon systems with a $U(1) \times \mathbb{R}$ isometry group for a class of regular Cauchy data. In our first paper \cite{chen}, we reduce the Einstein equations to a 2+1 dimensional Einstein-wave-Klein-Gordon system. And we show that the first possible singularity can only occur at the axis. In this paper, we give a proof for the global regularity for the 2+1 dimensional system. Firstly, we show the non-concentration of the energy near the first possible singularity. Then, we prove that the global regularity holds for initial data with small energy.

math.AP↗
arXiv2019

Momentum Schemes with Stochastic Variance Reduction for Nonconvex Composite Optimization

Two new stochastic variance-reduced algorithms named SARAH and SPIDER have been recently proposed, and SPIDER has been shown to achieve a near-optimal gradient oracle complexity for nonconvex optimization. However, the theoretical advantage of SPIDER does not lead to substantial improvement of practical performance over SVRG. To address this issue, momentum technique can be a good candidate to improve the performance of SPIDER. However, existing momentum schemes used in variance-reduced algorithms are designed specifically for convex optimization, and are not applicable to nonconvex scenarios. In this paper, we develop novel momentum schemes with flexible coefficient settings to accelerate SPIDER for nonconvex and nonsmooth composite optimization, and show that the resulting algorithms achieve the near-optimal gradient oracle complexity for achieving a generalized first-order stationary condition. Furthermore, we generalize our algorithm to online nonconvex and nonsmooth optimization, and establish an oracle complexity result that matches the state-of-the-art. Our extensive experiments demonstrate the superior performance of our proposed algorithm over other stochastic variance-reduced algorithms.

math.OC↗
arXiv2019

Possible 3D nematic odd-parity pairing in Sr$_2$RuO$_4$: experimental evidences and predictions

Due to the presence of a nontrivial three-dimensional spin-orbital entanglement, Sr$_2$RuO$_4$ may be a time-reversal invariant nematic $p$-wave superconductor with coexisting in-plane and out-of-plane pairings. Here we discuss various signatures of such a state if the out-of-plane pairing component is dominant. First, the enhancement of the superconducting $T_c$ under in-plane uniaxial strains is nearly a quadratic function of the strain, because the out-of-plane pairing lacks a linear-order coupling to the strain. Second, when the strain applies along a certain in-plane direction, the nematic $p$-wave pairing exhibits only a single phase transition as the temperature is lowered. These are consistent with several recent uniaxial strain measurements, which are otherwise hard to reconcile with chiral $p$-wave order. We further show that the nematic $p$-wave state can be distinguished from the chiral $p$-wave state through the velocity jumps of certain sound waves at the onset of superconductivity. Possible implications for $μ$SR experiment under strain and NMR Knight shift measurement are also discussed.

cond-mat.supr-con↗
arXiv2019

Non-Kitaev spin liquids in Kitaev materials

We point out that the Kitaev materials may not necessarily support Kitaev spin liquid. It is well-known that having a Kitaev term in the spin interaction is not the sufficient condition for the Kitaev spin liquid ground state. Many other spin liquids may be stabilized by the competing spin interactions of the systems. We thus explore the possibilities of non-Kitaev spin liquids in the honeycomb Kitaev materials. We carry out a systematic classification of gapped $\mathbb{Z}_2$ spin liquids using the Schwinger boson representation for the spin variables. The presence of strong spin-orbit coupling in the Kitaev materials brings new ingredients into the projective symmetry group classification of the non-Kitaev spin liquid. We predict the spectroscopic properties of these gapped non-Kitaev spin liquids. Moreover, among the gapped spin liquids that we discover, we identify the spin liquid whose spinon condensation leads to the magnetic Bragg peak structure of the zig-zag magnetic order that was observed in Na$_2$IrO$_3$ and $α$-RuCl$_3$. We further discuss the possibility of gapped $\mathbb{Z}_2$ spin liquid and the deconfined quantum criticality from the zig-zag magnetic order to spin dimerization in pressurized $α$-RuCl$_3$.

cond-mat.str-el↗