arXiv subjects

Ping Li

Publications and source records attributed to Ping Li.

At least 397 records · Page 22Linked to original sources

arXiv2016

2-Bit Random Projections, NonLinear Estimators, and Approximate Near Neighbor Search

The method of random projections has become a standard tool for machine learning, data mining, and search with massive data at Web scale. The effective use of random projections requires efficient coding schemes for quantizing (real-valued) projected data into integers. In this paper, we focus on a simple 2-bit coding scheme. In particular, we develop accurate nonlinear estimators of data similarity based on the 2-bit strategy. This work will have important practical applications. For example, in the task of near neighbor search, a crucial step (often called re-ranking) is to compute or estimate data similarities once a set of candidate data points have been identified by hash table techniques. This re-ranking step can take advantage of the proposed coding scheme and estimator. As a related task, in this paper, we also study a simple uniform quantization scheme for the purpose of building hash tables with projected data. Our analysis shows that typically only a small number of bits are needed. For example, when the target similarity level is high, 2 or 3 bits might be sufficient. When the target similarity level is not so high, it is preferable to use only 1 or 2 bits. Therefore, a 2-bit scheme appears to be overall a good choice for the task of sublinear time approximate near neighbor search via hash tables. Combining these results, we conclude that 2-bit random projections should be recommended for approximate near neighbor search and similarity estimation. Extensive experimental results are provided.

stat.ML↗

arXiv2016

Robust Room-Temperature Quantum Spin Hall Effect in Methyl-functionalized InBi honeycomb film

Two-dimensional (2D) group-III-V honeycomb films have attracted significant interest for their potential application in fields of quantum computing and nanoeletronics. Searching for 2D III-V films with high structural stability and large-gap are crucial for the realizations of dissipationless transport edge states using quantum spin Hall (QSH) effect. Based on first-principles calculations, we predict that the methyl-functionalized InBi monolayer (InBiCH3) has no dynamic instability, and host a QSH state with a band gap as large as 0.29 eV, exhibiting an interesting electronic behavior viable for room-temperature applications. The topological characteristic is confirmed by s-pxy bands inversion, topological invariant Z2 number, and the time-reversal symmetry protected helical edge states. Noticeably, the QSH states are tunable and robust against the mechanical strain, electric field and different levels of methyl coverages. We also find that InBiCH3 supported on h-BN substrate maintains a nontrivial QSH state, which harbors the edge states lying within the band gap of substrate. These findings demonstrate that the methyl-functionalized III-V films may be a good QSH platform for device design and fabrication in spintronics.

cond-mat.mtrl-sci↗

arXiv2016

Large-gap quantum spin Hall state in functionalized dumbbell stanene

Two-dimensional dumbbell (DB) stanene has been proposed as a promising candidate material for realizing quantum spin Hall effect (QSHE) by Tang et al [P. Tang, P. Chen, W. Cao, H. Huang, S. Cahangirov, L. Xian, Y. Xu, S. C. Zhang, W. Duan, A. Rubio. Phys. Rev. B, 90, 121408 (2014)]. However, the small bulk-gap limits its possible applications at room temperature. Based on first-principles calculations, we predict that its band gap can be enhanced to 148 meV under methyl-functionalization, which can be further tuned by applying lattice strain. The QSHE is confirmed by s-px,y band inversion, topological invariant Z2 = 1, and helical gapless edge within bulk band gap. Notably, the characteristic properties of edge states, such as the large Fermi velocity and Dirac cone, can be modulated by edge modification. The effects of substrates on topological properties are explored when it is grown on various substrates, like SiC, h-BN, and Bi2Te3 sheets. These findings provide significant guidance for future fabrication and realistic applications of QSHE based on stanene in spintronics.

cond-mat.mtrl-sci↗

arXiv2016

Binary and Multi-Bit Coding for Stable Random Projections

We develop efficient binary (i.e., 1-bit) and multi-bit coding schemes for estimating the scale parameter of $α$-stable distributions. The work is motivated by the recent work on one scan 1-bit compressed sensing (sparse signal recovery) using $α$-stable random projections, which requires estimating of the scale parameter at bits-level. Our technique can be naturally applied to data stream computations for estimating the $α$-th frequency moment. In fact, the method applies to the general scale family of distributions, not limited to $α$-stable distributions. Due to the heavy-tailed nature of $α$-stable distributions, using traditional estimators will potentially need many bits to store each measurement in order to ensure sufficient accuracy. Interestingly, our paper demonstrates that, using a simple closed-form estimator with merely 1-bit information does not result in a significant loss of accuracy if the parameter is chosen appropriately. For example, when $α=0+$, 1, and 2, the coefficients of the optimal estimation variances using full (i.e., infinite-bit) information are 1, 2, and 2, respectively. With the 1-bit scheme and appropriately chosen parameters, the corresponding variance coefficients are 1.544, $π^2/4$, and 3.066, respectively. Theoretical tail bounds are also provided. Using 2 or more bits per measurements reduces the estimation variance and importantly, stabilizes the estimate so that the variance is not sensitive to parameters. With look-up tables, the computational cost is minimal.

stat.ME↗

arXiv2016

Some results of linear codes over the ring $\mathbb{Z}_4+u\mathbb{Z}_4+v\mathbb{Z}_4+uv\mathbb{Z}_4$

In this paper, we mainly study the theory of linear codes over the ring $R =\mathbb{Z}_4+u\mathbb{Z}_4+v\mathbb{Z}_4+uv\mathbb{Z}_4$. By the Chinese Remainder Theorem, we have $R$ is isomorphic to the direct sum of four rings $\mathbb{Z}_4$. We define a Gray map $Φ$ from $R^{n}$ to $\mathbb{Z}_4^{4n}$, which is a distance preserving map. The Gray image of a cyclic code over $R^{n}$ is a linear code over $\mathbb{Z}_4$. Furthermore, we study the MacWilliams identities of linear codes over $R$ and give the the generator polynomials of cyclic codes over $R$. Finally, we discuss some properties of MDS codes over $R$.

cs.IT↗

arXiv2015

Recovery of Sparse Signals Using Multiple Orthogonal Least Squares

We study the problem of recovering sparse signals from compressed linear measurements. This problem, often referred to as sparse recovery or sparse reconstruction, has generated a great deal of interest in recent years. To recover the sparse signals, we propose a new method called multiple orthogonal least squares (MOLS), which extends the well-known orthogonal least squares (OLS) algorithm by allowing multiple $L$ indices to be chosen per iteration. Owing to inclusion of multiple support indices in each selection, the MOLS algorithm converges in much fewer iterations and improves the computational efficiency over the conventional OLS algorithm. Theoretical analysis shows that MOLS ($L > 1$) performs exact recovery of all $K$-sparse signals within $K$ iterations if the measurement matrix satisfies the restricted isometry property (RIP) with isometry constant $δ_{LK} < \frac{\sqrt{L}}{\sqrt{K} + 2 \sqrt{L}}.$ The recovery performance of MOLS in the noisy scenario is also studied. It is shown that stable recovery of sparse signals can be achieved with the MOLS algorithm when the signal-to-noise ratio (SNR) scales linearly with the sparsity level of input signals.

stat.ME↗

arXiv2015

A new structure of two-dimensional allotropes of group V elements

The elemental two-dimensional (2D) materials such as graphene, silicene, germanene, and black phosphorus have attracted considerable attention due to their fascinating physical properties. Structurally they possess the honeycomb or distorted honeycomb lattices, which are composed of six-atom rings. Here we find a new structure of 2D allotropes of group V elements composed of eight-atom rings, which we name as the octagonal tiling (OT) structure. First-principles calculations indicate that these allotropes are dynamically stable and are also thermally stable at temperatures up to 600 K. These allotropes are semiconductors with band gaps ranging from 0.3 to 2.0 eV, thus they are potentially useful in near- and mid-infrared optoelectronic devices. OT-Bi is also a 2D topological insulator (TI) with a band gap of 0.33 eV, which is the largest among the reported elemental 2D TIs, and this gap can be increased further by applying compressive strains.

cond-mat.mtrl-sci↗

arXiv2015

New approach for fabrication germanene with Dirac electrons preserved: A first principle study

How to obtain germanene with Dirac electrons preserved is still an open challenge. Here we report a sandwich-dehydrogenation approach, i.e., to fabricate germanene through dehydrogenating germanane in a sandwiched structure. The dehydrogenation can spontaneously occur for the sandwiched structure, which overcomes the problem of amorphization in the heating dehydrogenation approach. The obtained germanene preserve the Dirac electronic properties very well. Moreover, the Fermi velocity of germanene can be efficiently manipulated through controlling the interlayer spacing between germanane and the sandwiching surfaces. Our results indicate a guideline for fabrication of prefect two-dimensional materials.

cond-mat.mes-hall↗

arXiv2015

One Scan 1-Bit Compressed Sensing

Based on $α$-stable random projections with small $α$, we develop a simple algorithm for compressed sensing (sparse signal recovery) by utilizing only the signs (i.e., 1-bit) of the measurements. Using only 1-bit information of the measurements results in substantial cost reduction in collection, storage, communication, and decoding for compressed sensing. The proposed algorithm is efficient in that the decoding procedure requires only one scan of the coordinates. Our analysis can precisely show that, for a $K$-sparse signal of length $N$, $12.3K\log N/δ$ measurements (where $δ$ is the confidence) would be sufficient for recovering the support and the signs of the signal. While the method is very robust against typical measurement noises, we also provide the analysis of the scheme under random flipping of the signs of the measurements. \noindent Compared to the well-known work on 1-bit marginal regression (which can also be viewed as a one-scan method), the proposed algorithm requires orders of magnitude fewer measurements. Compared to 1-bit Iterative Hard Thresholding (IHT) (which is not a one-scan algorithm), our method is still significantly more accurate. Furthermore, the proposed method is reasonably robust against random sign flipping while IHT is known to be very sensitive to this type of noise.

stat.ME↗

arXiv2015

Tests for High-Dimensional Covariance Matrices Using Random Matrix Projection

The classic likelihood ratio test for testing the equality of two covariance matrices breakdowns due to the singularity of the sample covariance matrices when the data dimension $p$ is larger than the sample size $n$. In this paper, we present a conceptually simple method using random projection to project the data onto the one-dimensional random subspace so that the conventional methods can be applied. Both one-sample and two-sample tests for high-dimensional covariance matrices are studied. Asymptotic results are established and numerical results are given to compare our method with state-of-the-art methods in the literature.

stat.ME↗

arXiv2015

Functionalized Thallium Antimony Films as Excellent Candidates for Large-Gap Quantum Spin Hall Insulator

Group III-V films are of great importance for their potential application in spintronics and quantum computing. Search for two-dimensional III-V films with a nontrivial large-gap are quite crucial for the realization of dissipationless transport edge channels using quantum spin Hall (QSH) effects. Here we use first-principles calculations to predict a class of large-gap QSH insulators in functionalized TlSb monolayers (TlSbX2; (X = H, F, Cl, Br, I)), with sizable bulk gaps as large as 0.22~0.40 eV. The QSH state is identified by Z2 topological invariant together with helical edge states induced by spin-orbit coupling (SOC). Noticeably, the inverted band gap in the nontrivial states can be effectively tuned by the electric field and strain. Additionally, these films on BN substrate also maintain a nontrivial QSH state, which harbors a Dirac cone lying within the band gap. These findings may shed new light in future design and fabrication of QSH insulators based on two-dimensional honeycomb lattices in spintronics.

cond-mat.mtrl-sci↗

arXiv2015

Tunable Quantum Spin Hall Effect via Strain in two-Dimensional Arsenene Monolayer

The search for new quantum spin Hall (QSH) phase and effective manipulations of their edge states are very important for both fundamental sciences and practical applications. Here, we use first-principles calculations to study the strain-driven topological phase transition of two-dimensional (2D) arsenene monolayer. We find that the band gap of arsenene decreases with increasing strain and changes from indirect to direct, and then the s-p band inversion takes place at Γ point as the tensile strain is larger than 11.14%, which lead to a nontrivially topological state. A single pair of topologically protected helical edge states is established for the edge of arsenene, and their QSH states are confirmed with nontrivial topological invariant Z2 = 1. We also propose high-dielectric BN as an ideal substrate for the experimental synthesis of arsenene, maintaining its nontrivial topology. These findings provide a promising candidate platform for topological phenomena and new quantum devices operating at nanoelectronics.

cond-mat.mtrl-sci↗

arXiv2015

Reexamining generalized Chaplygin gas with the sign-changeable interaction

In this paper, we reexamine the generalized Chaplygin gas (GCG) cosmology with the sign-changeable interaction. The dynamical analysis show that there exists de-Sitter attractors in this model, which means that the late-time behaviors of the model is insensitive to the initial condition and thus alleviates the coincidence problem. Furthermore, we find that this interaction oscillates and tends to zero finally by numerical calculations of the nonlinear equations. In some specific cases of initial conditions, the interaction is positive (the energy transition from dark energy to dark matter) at high redshift while it is negative (the energy transition from dark matter to dark energy) at low redshift for suitable ranges of the parameters.

gr-qc↗

arXiv2015

Some remarks on the uniqueness of the complex projective spaces

We first notice in this article that if a compact Kähler manifold has the same integral cohomology ring and Pontrjagin classes as the complex projective space $\mathbb{C}P^n$, then it is biholomorphic to $\mathbb{C}P^n$ provided $n$ is odd. The same holds for even $n$ if we further assume that $M$ is simply-connected. This technically refines a classical result of Hirzebruch-Kodaira and Yau. This observation, together with a result of Dessai and Wilking, enables us to characterize all $\mathbb{C}P^n$ in terms of homotopy type under mild symmetry. When $n=4$, we can drop the requirement on Pontrjagin classes by showing that a simply-connected compact Kähler manifold having the same integral cohomology ring as $\mathbb{C}P^4$ is biholomorphic to $\mathbb{C}P^4$, which improves on results of Fujita and Libgober-Wood.

math.DG↗

arXiv2015

Sign Stable Random Projections for Large-Scale Learning

We study the use of "sign $α$-stable random projections" (where $0<α\leq 2$) for building basic data processing tools in the context of large-scale machine learning applications (e.g., classification, regression, clustering, and near-neighbor search). After the processing by sign stable random projections, the inner products of the processed data approximate various types of nonlinear kernels depending on the value of $α$. Thus, this approach provides an effective strategy for approximating nonlinear learning algorithms essentially at the cost of linear learning. When $α=2$, it is known that the corresponding nonlinear kernel is the arc-cosine kernel. When $α=1$, the procedure approximates the arc-cos-$χ^2$ kernel (under certain condition). When $α\rightarrow0+$, it corresponds to the resemblance kernel. From practitioners' perspective, the method of sign $α$-stable random projections is ready to be tested for large-scale learning applications, where $α$ can be simply viewed as a tuning parameter. What is missing in the literature is an extensive empirical study to show the effectiveness of sign stable random projections, especially for $α\neq 2$ or 1. The paper supplies such a study on a wide variety of classification datasets. In particular, we compare shoulder-by-shoulder sign stable random projections with the recently proposed "0-bit consistent weighted sampling (CWS)" (Li 2015).

stat.ML↗

arXiv2015

Regularization-free estimation in trace regression with symmetric positive semidefinite matrices

Over the past few years, trace regression models have received considerable attention in the context of matrix completion, quantum state tomography, and compressed sensing. Estimation of the underlying matrix from regularization-based approaches promoting low-rankedness, notably nuclear norm regularization, have enjoyed great popularity. In the present paper, we argue that such regularization may no longer be necessary if the underlying matrix is symmetric positive semidefinite (\textsf{spd}) and the design satisfies certain conditions. In this situation, simple least squares estimation subject to an \textsf{spd} constraint may perform as well as regularization-based approaches with a proper choice of the regularization parameter, which entails knowledge of the noise level and/or tuning. By contrast, constrained least squares estimation comes without any tuning parameter and may hence be preferred due to its simplicity.

stat.ML↗

arXiv2015

A characterization of the $\hat{A}$-genus as a linear combination of Pontrjagin numbers

We show in this short note that if a rational linear combination of Pontrjagin numbers vanishes on all simply-connected $4k$-dimensional closed connected and oriented spin manifolds admitting a Riemannian metric whose Ricci curvature is nonnegative and nonzero at any point, then this linear combination must be a multiple of the $\hat{A}$-genus, which improves on a result of Gromov and Lawson. Our proof combines an idea of Atiyah and Hirzebruch and the celebrated Calabi-Yau theorem.

math.DG↗

arXiv2015

Recovery of Sparse Signals via Generalized Orthogonal Matching Pursuit: A New Analysis

As an extension of orthogonal matching pursuit (OMP) improving the recovery performance of sparse signals, generalized OMP (gOMP) has recently been studied in the literature. In this paper, we present a new analysis of the gOMP algorithm using restricted isometry property (RIP). We show that if the measurement matrix $\mathbfΦ \in \mathcal{R}^{m \times n}$ satisfies the RIP with $$δ_{\max \left\{9, S + 1 \right\}K} \leq \frac{1}{8},$$ then gOMP performs stable reconstruction of all $K$-sparse signals $\mathbf{x} \in \mathcal{R}^n$ from the noisy measurements $\mathbf{y} = \mathbf{Φx} + \mathbf{v}$ within $\max \left\{K, \left\lfloor \frac{8K}{S} \right\rfloor \right\}$ iterations where $\mathbf{v}$ is the noise vector and $S$ is the number of indices chosen in each iteration of the gOMP algorithm. For Gaussian random measurements, our results indicate that the number of required measurements is essentially $m = \mathcal{O}(K \log \frac{n}{K})$, which is a significant improvement over the existing result $m = \mathcal{O}(K^2 \log \frac{n}{K})$, especially for large $K$.

cs.IT↗