arXiv subjects

Song Liu

Publications and source records attributed to Song Liu.

At least 199 records · Page 11Linked to original sources

arXiv2020

Model Reuse with Reduced Kernel Mean Embedding Specification

Given a publicly available pool of machine learning models constructed for various tasks, when a user plans to build a model for her own machine learning application, is it possible to build upon models in the pool such that the previous efforts on these existing models can be reused rather than starting from scratch? Here, a grand challenge is how to find models that are helpful for the current application, without accessing the raw training data for the models in the pool. In this paper, we present a two-phase framework. In the upload phase, when a model is uploading into the pool, we construct a reduced kernel mean embedding (RKME) as a specification for the model. Then in the deployment phase, the relatedness of the current task and pre-trained models will be measured based on the value of the RKME specification. Theoretical results and extensive experiments validate the effectiveness of our approach.

cs.LG↗

arXiv2020

Personalized Dialogue Generation with Diversified Traits

Endowing a dialogue system with particular personality traits is essential to deliver more human-like conversations. However, due to the challenge of embodying personality via language expression and the lack of large-scale persona-labeled dialogue data, this research problem is still far from well-studied. In this paper, we investigate the problem of incorporating explicit personality traits in dialogue generation to deliver personalized dialogues. To this end, firstly, we construct PersonalDialog, a large-scale multi-turn dialogue dataset containing various traits from a large number of speakers. The dataset consists of 20.83M sessions and 56.25M utterances from 8.47M speakers. Each utterance is associated with a speaker who is marked with traits like Age, Gender, Location, Interest Tags, etc. Several anonymization schemes are designed to protect the privacy of each speaker. This large-scale dataset will facilitate not only the study of personalized dialogue generation, but also other researches on sociolinguistics or social science. Secondly, to study how personality traits can be captured and addressed in dialogue generation, we propose persona-aware dialogue generation models within the sequence to sequence learning framework. Explicit personality traits (structured by key-value pairs) are embedded using a trait fusion module. During the decoding process, two techniques, namely persona-aware attention and persona-aware bias, are devised to capture and address trait-related information. Experiments demonstrate that our model is able to address proper traits in different contexts. Case studies also show interesting results for this challenging research problem.

cs.CL↗

arXiv2019

Chiral spin-wave velocities induced by all-garnet interfacial Dzyaloshinskii-Moriya interaction in ultrathin yttrium iron garnet films

Spin waves can probe the Dzyaloshinskii-Moriya interaction (DMI) which gives rise to topological spin textures, such as skyrmions. However, the DMI has not yet been reported in yttrium iron garnet (YIG) with arguably the lowest damping for spin waves. In this work, we experimentally evidence the interfacial DMI in a 7~nm-thick YIG film by measuring the nonreciprocal spin wave propagation in terms of frequency, amplitude and most importantly group velocities using all electrical spin-wave spectroscopy. The velocities of propagating spin waves show chirality among three vectors, i.e. the film normal direction, applied field and spin-wave wavevector. By measuring the asymmetric group velocities, we extract a DMI constant of 16~$μ$J/m$^{2}$ which we independently confirm by Brillouin light scattering. Thickness-dependent measurements reveal that the DMI originates from the oxide interface between the YIG and garnet substrate. The interfacial DMI discovered in the ultrathin YIG films is of key importance for functional chiral magnonics as ultra-low spin-wave damping can be achieved.

cond-mat.mtrl-sci↗

arXiv2019

Fisher Efficient Inference of Intractable Models

Maximum Likelihood Estimators (MLE) has many good properties. For example, the asymptotic variance of MLE solution attains equality of the asymptotic Cram{é}r-Rao lower bound (efficiency bound), which is the minimum possible variance for an unbiased estimator. However, obtaining such MLE solution requires calculating the likelihood function which may not be tractable due to the normalization term of the density model. In this paper, we derive a Discriminative Likelihood Estimator (DLE) from the Kullback-Leibler divergence minimization criterion implemented via density ratio estimation and a Stein operator. We study the problem of model inference using DLE. We prove its consistency and show that the asymptotic variance of its solution can attain the equality of the efficiency bound under mild regularity conditions. We also propose a dual formulation of DLE which can be easily optimized. Numerical studies validate our asymptotic theorems and we give an example where DLE successfully estimates an intractable model constructed using a pre-trained deep neural network.

stat.ML↗

arXiv2019

WSe2/WS2 moiré superlattices: a new Hubbard model simulator

The Hubbard model, first formulated by physicist John Hubbard in the 1960s, is a simple theoretical model of interacting quantum particles in a lattice. The model is thought to capture the essential physics of high-temperature superconductors, magnetic insulators, and other complex emergent quantum many-body ground states. Although the Hubbard model is greatly simplified as a representation of most real materials, it has nevertheless proved difficult to solve accurately except in the one-dimensional case. Physical realization of the Hubbard model in two or three dimensions, which can act as quantum simulators, therefore have a vital role to play in solving the strong-correlation puzzle. Here we obtain a quantum phase diagram of the two-dimensional triangular lattice Hubbard model by studying angle-aligned WSe2/WS2 bilayers, which form moiré superlattices because of the difference in lattice constant between the two two-dimensional materials. We probe both charge and magnetic properties of the system by measuring the dependence of optical response on out-of-plane magnetic field, and on gate-tuned carrier density. At half filling of the first hole moiré superlattice band, we observe a Mott insulating state with antiferromagnetic Curie-Weiss behavior as expected for a Hubbard model in the strong interaction regime. Past half filling, our experiment suggests an antiferromagnetic to paramagnetic quantum phase transition near 0.6 filling. Our results establish a new solid-state platform based on moiré superlattices which can be used to simulate outstanding problems in strong correlation physics that are manifested by triangular lattice Hubbard models.

cond-mat.mtrl-sci↗

arXiv2019

Wafer-Scale and Deterministic Patterned Growth of Monolayer MoS2 via Vapor-Liquid-Solid Method

Vapor transportation is the core process in growing transition-metal dichalcogenides (TMDCs) by chemical vapor deposition (CVD). One inevitable problem is the spatial inhomogeneity of the vapors. The non-stoichiometric supply of transition-metal precursors and chalcogen leads to poor control in products' location, morphology, crystallinity, uniformity and batch to batch reproducibility. While vapor-liquid-solid (VLS) growth involves molten precursors at the growth temperatures higher than their melting points. The liquid sodium molybdate can precipitate solid MoS2 monolayers when saturated with sulfur vapor. Taking advantage of the VLS growth, we achieved three kinds of important achievements: (a) 4-inch-wafer-scale uniform growth of MoS2 flakes on SiO2/Si substrates, (b) 2-inch-wafer-scale growth of continuous MoS2 film with a grain size exceeding 100 um on sapphire substrates, and (c) pattern (site-controlled) growth of MoS2 flakes and film. We clarified that the VLS growth thus pave the new way for the high-efficient, scalable synthesis of two-dimensional TMDC monolayers.

cond-mat.mtrl-sci↗

arXiv2019

Refractive Index-Based Control of Hyperbolic Phonon-Polariton Propagation

Hyperbolic phonon polaritons (HPhPs) are generated when infrared photons couple to polar optic phonons in anisotropic media, confining long-wavelength light to nanoscale volumes. However, to realize the full potential of HPhPs for infrared optics, it is crucial to understand propagation and loss mechanisms on substrates suitable for applications from waveguiding to infrared sensing. In this paper, we employ scattering-type scanning near-field optical microscopy (s-SNOM) and nano-Fourier transform infrared (FTIR) spectroscopy, in concert with analytical and numerical calculations, to elucidate HPhP characteristics as a function of the complex substrate dielectric function. We consider propagation on suspended, dielectric and metallic substrates to demonstrate that the thickness-normalized wavevector can be reduced by a factor of 25 simply by changing the substrate from dielectric to metallic behavior. Moreover, by incorporating the imaginary contribution to the dielectric function in lossy materials, the wavevector can be dynamically controlled by small local variations in loss or carrier density. Such effects may therefore be used to spatially separate hyperbolic modes of different orders, and indicates that for index-based sensing schemes that HPhPs can be more sensitive than surface polaritons in the thin film limit. Our results advance our understanding of fundamental polariton excitations and their potential for on-chip photonics and planar metasurface optics.

physics.optics↗

arXiv2019

Infrared hyperbolic metasurface based on nanostructured van der Waals materials

Metasurfaces with strongly anisotropic optical properties can support deep subwavelength-scale confined electromagnetic waves (polaritons) that promise opportunities for controlling light in photonic and optoelectronic applications. We develop a mid-infrared hyperbolic metasurface by nanostructuring a thin layer of hexagonal boron nitride supporting deep subwavelength-scale phonon polaritons that propagate with in-plane hyperbolic dispersion. By applying an infrared nanoimaging technique, we visualize the concave (anomalous) wavefronts of a diverging polariton beam, which represent a landmark feature of hyperbolic polaritons. The results illustrate how near-field microscopy can be applied to reveal the exotic wavefronts of polaritons in anisotropic materials, and demonstrate that nanostructured van der Waals materials can form a highly variable and compact platform for hyperbolic infrared metasurface devices and circuits.

physics.optics↗

arXiv2019

Experimental Realization of Nonadiabatic Shortcut to Non-Abelian Geometric Gates

When a quantum system is driven adiabatically through a parametric cycle in a degenerate Hilbert space, the state would acquire a non-Abelian geometric phase, which is stable and forms the foundation for holonomic quantum computation (HQC). However, in the adiabatic limit, the environmental decoherence becomes a significant source of errors. Recently, various non-adiabatic HQC schemes have been proposed, but all at the price of increased sensitivity to control errors. Alternatively, there exist theoretical proposals for speeding up HQC by the technique of "shortcut to adiabaticity" (STA), but no experimental demonstration has been reported so far, as these proprosals involve a complicated control of four energy levels simultaneously. Here we propose and experimentally demonstrate that HQC via shortcut to adiabaticity can be constructed with only three energy levels, using a superconducting qubit in a scalable architecture. With this scheme, all holonomic single-qubit operations can be realized non-adiabatically through a single cycle of state evolution. As a result, we are able to experimentally benchmark the stability of STA+HQC against NHQC in the same platform. The flexibility and simplicity of our scheme makes it also implementable on other systems, such as nitrogen-vacancy center, quantum dots, and nuclear magnetic resonance.

quant-ph↗

arXiv2018

Probing hyperbolic polaritons using infrared attenuated total reflectance micro-spectroscopy

Hyperbolic polariton modes are highly appealing for a broad range of applications in nanophotonics, including surfaced enhanced sensing, sub-diffractional imaging and reconfigurable metasurfaces. Here we show that attenuated total reflectance micro-spectroscopy (ATR) using standard spectroscopic tools can launch hyperbolic polaritons in a Kretschmann-Raether configuration. We measure multiple hyperbolic and dielectric modes within the naturally hyperbolic material hexagonal boron nitride as a function of different isotopic enrichments and flake thickness. This overcomes the technical challenges of measurement approaches based on nanostructuring, or scattering scanning nearfield optical microscopy. Ultimately, our ATR approach allows us to compare the optical properties of small-scale materials prepared by different techniques systematically

physics.optics↗

arXiv2018

Reconfigurable Mid-Infrared Hyperbolic Metasurfaces using Phase-Change Materials

Metasurfaces offer the potential to control light propagation at the nanoscale for applications in both free-space and surface-confined geometries. Existing metasurfaces frequently utilize metallic polaritonic elements with high absorption losses, and/or fixed geometrical designs that serve a single function. Here we overcome these limitations by demonstrating a reconfigurable hyperbolic metasurface comprising of a heterostructure of isotopically enriched hexagonal boron nitride (hBN) in direct contact with the phase-change material (PCM) vanadium dioxide (VO2). Spatially localized metallic and dielectric domains in VO2 change the wavelength of the hyperbolic phonon polaritons (HPhPs) supported in hBN by a factor 1.6 at 1450cm-1. This induces in-plane launching, refraction and reflection of HPhPs in the hBN, proving reconfigurable control of in-plane HPhP propagation at the nanoscale15. These results exemplify a generalizable framework based on combining hyperbolic media and PCMs in order to design optical functionalities such as resonant cavities, beam steering, waveguiding and focusing with nanometric control.

physics.optics↗

arXiv2018

The electromagnetic waves propagation in unmagnetized plasma media using parallelized finite-difference time-domain method

The finite-difference time-domain (FDTD) method has been commonly utilized to simulate the electromagnetic (EM) waves propagation in the plasma media. However, the FDTD method may bring about extra run-time on concerning computationally large and complicated EM problems. Fortunately, the FDTD method is easy to parallelize. Besides, GPU has been widely used for parallel computing due to its unique SPMD (Single Program Multiple Data) architecture. In this paper, we represent the parallel Runge-Kutta exponential time differencing scheme FDTD (RKETD) method for the unmagnetized plasma implemented on GPU. The detailed flowchart of parallel RKETD-FDTD method is described. The accuracy and acceleration performance of the proposed parallel RKETD-FDTD method implemented on GPU are substantiated by calculating the reflection and transmission coefficients for one-dimensional unmagnetized plasma slab. The results indicate that the numerical precision of the parallel RKETD-FDTD scheme is consistent with that of the code implemented on CPU. The computation efficiency is greatly improved compared with merely CPU-based serial RKETD-FDTD method. Moreover, the comparisons of the performance of CUDA-based GPU parallel program, OpenMP (Open Multi-Processing)-based CPU parallel program, and single-CPU serial program on the same host computer are done. Compared with the serial program, both parallel programs get good results, while GPU-based parallel program gains better result.

physics.comp-ph↗

arXiv2018

GPU accelerated manifold correction method for spinning compact binaries

The conservative Post-Newtonian (PN) Hamiltonian formulation of spinning compact binaries has six integrals of motion including the total energy, the total angular momentum and the constant unit lengths of spins. The manifold correction method can effectively eliminate the integration errors accumulation in a long time. In this paper, the accelerated manifold correction method based on graphics processing unit (GPU) is designed to simulate the dynamic evolution of spinning compact binaries. The feasibility and the efficiency of parallel computation on GPU for spinning compact binaries have been confirmed by various numerical experiments. The numerical comparisons show that the accuracy on GPU execution of manifold corrections method has a good agreement with the execution of codes on merely central processing unit (CPU-based) method. The acceleration ability when the codes are implemented on GPU can increase enormously through the use of shared memory and register optimization techniques without additional hardware costs, implying that the speedup is nearly 13 times as compared with the codes executed on CPU for phase space scan (including orbits). In addition, GPU-accelerated manifold correction method is used to numerically study how dynamics are affected by the spin-induced quadrupole-monopole interaction for black hole binary system.

astro-ph.IM↗

arXiv2017

Trimmed Density Ratio Estimation

Density ratio estimation is a vital tool in both machine learning and statistical community. However, due to the unbounded nature of density ratio, the estimation procedure can be vulnerable to corrupted data points, which often pushes the estimated ratio toward infinity. In this paper, we present a robust estimator which automatically identifies and trims outliers. The proposed estimator has a convex formulation, and the global optimum can be obtained via subgradient descent. We analyze the parameter estimation error of this estimator under high-dimensional settings. Experiments are conducted to verify the effectiveness of the estimator.

stat.ML↗

arXiv2017

Ultra-low-loss Polaritons in Isotopically Pure Materials: A New Approach

Conventional optical components are limited to size-scales much larger than the wavelength of light, as changes in the amplitude, phase and polarization of the electromagnetic fields are accrued gradually along an optical path. However, advances in nanophotonics have produced ultra-thin, co-called "flat" optical components that beget abrupt changes in these properties over distances significantly shorter than the free space wavelength. While high optical losses still plague many approaches, phonon polariton (PhP) materials have demonstrated long lifetimes for sub-diffractional modes in comparison to plasmon-polariton-based nanophotonics. We experimentally observe a three-fold improvement in polariton lifetime through isotopic enrichment of hexagonal boron nitride (hBN). Commensurate increases in the polariton propagation length are demonstrated via direct imaging of polaritonic standing waves by means of infrared nano-optics. Our results provide the foundation for a materials-growth-directed approach towards realizing the loss control necessary for the development of PhP-based nanophotonic devices.

cond-mat.mtrl-sci↗

arXiv2017

Interpreting Outliers: Localized Logistic Regression for Density Ratio Estimation

We propose an inlier-based outlier detection method capable of both identifying the outliers and explaining why they are outliers, by identifying the outlier-specific features. Specifically, we employ an inlier-based outlier detection criterion, which uses the ratio of inlier and test probability densities as a measure of plausibility of being an outlier. For estimating the density ratio function, we propose a localized logistic regression algorithm. Thanks to the locality of the model, variable selection can be outlier-specific, and will help interpret why points are outliers in a high-dimensional space. Through synthetic experiments, we show that the proposed algorithm can successfully detect the important features for outliers. Moreover, we show that the proposed algorithm tends to outperform existing algorithms in benchmark datasets.

stat.ML↗

arXiv2017

Learning Sparse Structural Changes in High-dimensional Markov Networks: A Review on Methodologies and Theories

Recent years have seen an increasing popularity of learning the sparse \emph{changes} in Markov Networks. Changes in the structure of Markov Networks reflect alternations of interactions between random variables under different regimes and provide insights into the underlying system. While each individual network structure can be complicated and difficult to learn, the overall change from one network to another can be simple. This intuition gave birth to an approach that \emph{directly} learns the sparse changes without modelling and learning the individual (possibly dense) networks. In this paper, we review such a direct learning method with some latest developments along this line of research.

stat.ML↗

arXiv2017

Large-scale Isolated Gesture Recognition Using Convolutional Neural Networks

This paper proposes three simple, compact yet effective representations of depth sequences, referred to respectively as Dynamic Depth Images (DDI), Dynamic Depth Normal Images (DDNI) and Dynamic Depth Motion Normal Images (DDMNI). These dynamic images are constructed from a sequence of depth maps using bidirectional rank pooling to effectively capture the spatial-temporal information. Such image-based representations enable us to fine-tune the existing ConvNets models trained on image data for classification of depth sequences, without introducing large parameters to learn. Upon the proposed representations, a convolutional Neural networks (ConvNets) based method is developed for gesture recognition and evaluated on the Large-scale Isolated Gesture Recognition at the ChaLearn Looking at People (LAP) challenge 2016. The method achieved 55.57\% classification accuracy and ranked $2^{nd}$ place in this challenge but was very close to the best performance even though we only used depth data.

cs.CV↗