arXiv subjects

Yong Xu

Publications and source records attributed to Yong Xu.

At least 433 records · Page 24Linked to original sources

arXiv2018

Intrinsic magnetic topological insulators in van der Waals layered MnBi$_2$Te$_4$-family materials

The interplay of magnetism and topology is a key research subject in condensed matter physics and material science, which offers great opportunities to explore emerging new physics, like the quantum anomalous Hall (QAH) effect, axion electrodynamics and Majorana fermions. However, these exotic physical effects have rarely been realized in experiment, due to the lacking of suitable working materials. Here we predict that van der Waals layered MnBi$_2$Te$_4$-family materials show two-dimensional (2D) ferromagnetism in the single layer and three-dimensional (3D) $A$-type antiferromagnetism in the bulk, which could serve as a next-generation material platform for the state-of-art research. Remarkably, we predict extremely rich topological quantum effects with outstanding features in an experimentally available material MnBi$_2$Te$_4$, including a 3D antiferromagnetic topological insulator with the long-sought topological axion states, the type-II magnetic Weyl semimetal (WSM) with simply one pair of Weyl points, and the high-temperature intrinsic QAH effect. These striking predictions, if proved experimentally, could profoundly transform future research and technology of topological quantum physics.

cond-mat.mtrl-sci↗

arXiv2018

Surface symmetry breaking and disorder effects on superconductivity in perovskite BaBi3 epitaxial films

The structural or electronic symmetry breaking of the host lattice is a recurrent phenomenon in many quantum materials, including superconductors. Yet, how these broken symmetry states affect the electronic pair wave function of superconductivity have been rarely elucidated. Here, using low-temperature scanning tunneling microscopy and first-principles calculations, we identify the broken rotational symmetry via stripe ordering on the (001) surface of perovskite BaBi3 films grown by molecular beam epitaxy, and show that it consequently leads to anisotropic superconductivity with twofold symmetry. In contrast, the structural disorder smears out the anisotropy of electron pairing and fills superconducting subgap density of states as the film thickness is reduced. A quasi-long range model of superconducting fluctuations is revealed to describe the tunneling conductance spectra of thin BaBi3 films well, and to exemplify how disorders contribute to the low-energy quasiparticle excitations in superconductors. Our findings help understand the effects of symmetry breaking states and disorders on superconductivity, particularly the existing tunneling conductance spectra there.

cond-mat.supr-con↗

arXiv2018

Scheme to Equilibrate the Quantized Hall Response of Topological Systems from Coherent Dynamics

Two-dimensional topologically distinct insulators are separated by topological gapless points, which exist as Weyl points in three-dimensional momentum space. Slowly varying parameters in the two-dimensional Hamiltonian across two distinct phases therefore necessarily experiences the gap closing process, which prevents the intrinsic physical observable, the Hall response, from equilibrating. To equilibrate the Hall response, engineered laser noises were introduced at the price of destroying the quantum coherence. Here we demonstrate a new scheme to equilibrate the quantized Hall response from pure coherent dynamics as the Hamiltonian is slowly tuned from the topologically trivial to nontrivial regimes. We show the elements that affect the process of equilibration including the sequence when the electric field is switched on, its strength and the band dispersion of the final Hamiltonian. We further apply our method to Weyl semimetals in three dimensions and find the equilibrated Hall response despite the underlying gapless band structure. Our finding not only lays the theoretical foundation for observing the two-dimensional topological phase transition but also for observing and controlling Weyl semimetals in ultracold atomic gases.

cond-mat.quant-gas↗

arXiv2018

Time-reversal-invariant spin-orbit-coupled bilayer Bose-Einstein Condensates

Time-reversal invariance plays a crucial role for many exotic quantum phases, particularly for topologically nontrivial states, in spin-orbit coupled electronic systems. Recently realized spin-orbit coupled cold-atom systems, however, lack the time-reversal symmetry due to the inevitable presence of an effective transverse Zeeman field. We address this issue by analyzing a realistic scheme to preserve time-reversal symmetry in spin-orbit coupled ultracold atoms, with the use of Hermite-Gaussian-laser induced Raman transitions that preserve spin-layer time-reversal symmetry. We find that the system's quantum states form Kramers pairs, resulting in symmetry-protected gap closing of the lowest two bands at arbitrarily large Raman coupling. We also show that Bose gases in this setup exhibit interaction-induced layer-stripe and uniform phases as well as intriguing spin-layer symmetry and spin-layer correlation.

cond-mat.quant-gas↗

arXiv2018

Capsule Routing for Sound Event Detection

The detection of acoustic scenes is a challenging problem in which environmental sound events must be detected from a given audio signal. This includes classifying the events as well as estimating their onset and offset times. We approach this problem with a neural network architecture that uses the recently-proposed capsule routing mechanism. A capsule is a group of activation units representing a set of properties for an entity of interest, and the purpose of routing is to identify part-whole relationships between capsules. That is, a capsule in one layer is assumed to belong to a capsule in the layer above in terms of the entity being represented. Using capsule routing, we wish to train a network that can learn global coherence implicitly, thereby improving generalization performance. Our proposed method is evaluated on Task 4 of the DCASE 2017 challenge. Results show that classification performance is state-of-the-art, achieving an F-score of 58.6%. In addition, overfitting is reduced considerably compared to other architectures.

cs.SD↗

arXiv2018

Bidirectional Attentive Fusion with Context Gating for Dense Video Captioning

Dense video captioning is a newly emerging task that aims at both localizing and describing all events in a video. We identify and tackle two challenges on this task, namely, (1) how to utilize both past and future contexts for accurate event proposal predictions, and (2) how to construct informative input to the decoder for generating natural event descriptions. First, previous works predominantly generate temporal event proposals in the forward direction, which neglects future video context. We propose a bidirectional proposal method that effectively exploits both past and future contexts to make proposal predictions. Second, different events ending at (nearly) the same time are indistinguishable in the previous works, resulting in the same captions. We solve this problem by representing each event with an attentive fusion of hidden states from the proposal module and video contents (e.g., C3D features). We further propose a novel context gating mechanism to balance the contributions from the current event and its surrounding contexts dynamically. We empirically show that our attentively fused event representation is superior to the proposal hidden states or video contents alone. By coupling proposal and captioning modules into one unified framework, our model outperforms the state-of-the-arts on the ActivityNet Captions dataset with a relative gain of over 100% (Meteor score increases from 4.82 to 9.65).

cs.CV↗

arXiv2018

Dynamical Weyl Points and 4D Nodal Rings in Cold Atomic Gases

Controllability of ultracold atomic gases has reached an unprecedented level, allowing for experimental realization of the long-sought-after Thouless pump, which can be interpreted as a dynamical quantum Hall effect. On the other hand, Weyl semimetals and Weyl nodal line semimetals with touching points and rings in band structures have sparked tremendous interest in various fields in the past few years. Here, we show that dynamical Weyl points and dynamical 4D Weyl nodal rings, which are protected by the first Chern number on a parameter surface formed by quasi-momentum and time, emerge in a two-dimensional and three-dimensional system, respectively. We find that the topological pump occurs in these systems but the amount of pumped particles is not quantized and can be continuously tuned by controlling experimental parameters over a wide range. We also propose an experimental scheme to realize the dynamical Weyl points and 4D Weyl nodal rings and to observe their corresponding topological pump in cold atomic gases.

cond-mat.quant-gas↗

arXiv2018

Audio Set classification with attention model: A probabilistic perspective

This paper investigates the classification of the Audio Set dataset. Audio Set is a large scale weakly labelled dataset of sound clips. Previous work used multiple instance learning (MIL) to classify weakly labelled data. In MIL, a bag consists of several instances, and a bag is labelled positive if at least one instances in the audio clip is positive. A bag is labelled negative if all the instances in the bag are negative. We propose an attention model to tackle the MIL problem and explain this attention model from a novel probabilistic perspective. We define a probability space on each bag, where each instance in the bag has a trainable probability measure for each class. Then the classification of a bag is the expectation of the classification output of the instances in the bag with respect to the learned probability measure. Experimental results show that our proposed attention model modeled by fully connected deep neural network obtains mAP of 0.327 on Audio Set dataset, outperforming the Google's baseline of 0.314 and recurrent neural network of 0.325.

cs.SD↗

arXiv2018

A joint separation-classification model for sound event detection of weakly labelled data

Source separation (SS) aims to separate individual sources from an audio recording. Sound event detection (SED) aims to detect sound events from an audio recording. We propose a joint separation-classification (JSC) model trained only on weakly labelled audio data, that is, only the tags of an audio recording are known but the time of the events are unknown. First, we propose a separation mapping from the time-frequency (T-F) representation of an audio to the T-F segmentation masks of the audio events. Second, a classification mapping is built from each T-F segmentation mask to the presence probability of each audio event. In the source separation stage, sources of audio events and time of sound events can be obtained from the T-F segmentation masks. The proposed method achieves an equal error rate (EER) of 0.14 in SED, outperforming deep neural network baseline of 0.29. Source separation SDR of 8.08 dB is obtained by using global weighted rank pooling (GWRP) as probability mapping, outperforming the global max pooling (GMP) based probability mapping giving SDR at 0.03 dB. Source code of our work is published.

cs.SD↗

arXiv2017

Electrically modulated SQUID with single Josephson junction coupled by a time-reversal breaking Weyl semimetal thin film

Usually, the superconducting quantum interference device (SQUID) consists of two Josephson junctions and the interference therein is modulated by a magnetic flux. In this work, we propose an electrically modulated SQUID consisting of single Josephson junction coupled by a time-reversal breaking Weyl semimetal thin film. For a low Fermi energy, the Josephson current is only mediated by Fermi arc surface states, and has an arbitrary ground-state phase difference \phi0 which is directly proportional to the product of the transverse electric field and the cross section area of the junction.For a suitable Fermi energy, the bulk states make comparable contributions to the Josephson current with the current-phase relation of a 0-junction. The interference between the surface channel and the bulk channel results in an electrically modulated SQUID with single Josephson junction, which provides an experimental proposal to identify magnetic Weyl semimetals and may have potential applications in superconducting quantum computation.

cond-mat.supr-con↗

arXiv2017

Pseudospins and topological effects of phonons in a Kekule lattice

The search for exotic topological effects of phonons has attracted enormous interest for both fundamental science and practical applications. By studying phonons in a Kekulé lattice, we find a new type of pseudospins characterized by quantized Berry phases and pseudoangular momenta, which introduces various novel topological effects, including topologically protected pseudospin-polarized interface states and a phonon pseudospin Hall effect. We further demonstrate a pseudospin-contrasting optical selection rule and a pseudospin Zeeman effect, giving a complete generation-manipulation-detection paradigm of the phonon pseudospin. The pseudospin and topology-related physics revealed for phonons is general and applicable for electrons, photons and other particles.

cond-mat.mes-hall↗

arXiv2017

Surrey-cvssp system for DCASE2017 challenge task4

In this technique report, we present a bunch of methods for the task 4 of Detection and Classification of Acoustic Scenes and Events 2017 (DCASE2017) challenge. This task evaluates systems for the large-scale detection of sound events using weakly labeled training data. The data are YouTube video excerpts focusing on transportation and warnings due to their industry applications. There are two tasks, audio tagging and sound event detection from weakly labeled data. Convolutional neural network (CNN) and gated recurrent unit (GRU) based recurrent neural network (RNN) are adopted as our basic framework. We proposed a learnable gating activation function for selecting informative local features. Attention-based scheme is used for localizing the specific events in a weakly-supervised mode. A new batch-level balancing strategy is also proposed to tackle the data unbalancing problem. Fusion of posteriors from different systems are found effective to improve the performance. In a summary, we get 61% F-value for the audio tagging subtask and 0.73 error rate (ER) for the sound event detection subtask on the development set. While the official multilayer perceptron (MLP) based baseline just obtained 13.1% F-value for the audio tagging and 1.02 for the sound event detection.

cs.SD↗

arXiv2017

Size Dependence and Ballistic Limits of Thermal Transport in Anisotropic Layered Two-Dimensional Materials

Layered materials have uncommonly anisotropic thermal properties due to their strong in-plane covalent bonds and weak out-of-plane van der Waals interactions. Here we examine heat flow in graphene (graphite), h-BN, MoS2, and WS2 monolayers and bulk films, from diffusive to ballistic limits. We determine the ballistic thermal conductance limit (Gball) both in-plane and out-of-plane, based on full phonon dispersions from first-principles calculations. An overall phonon mean free path (λ) is expressed in terms of Gball and the diffusive thermal conductivity, consistent with kinetic theory if proper averaging of phonon group velocity is used. We obtain a size-dependent thermal conductivity k(L) in agreement with available experiments, and find that k(L) only converges to >90% of the diffusive thermal conductivity for sample sizes L > 16λ, which ranges from ~140 nm for MoS2 cross-plane to ~10 um for suspended graphene in-plane. These results provide a deeper understanding of microscopic thermal transport, revealing that device scales below which thermal size effects should be taken into account are generally larger than previously thought.

cond-mat.mes-hall↗

arXiv2017

Large-scale weakly supervised audio classification using gated convolutional neural network

In this paper, we present a gated convolutional neural network and a temporal attention-based localization method for audio classification, which won the 1st place in the large-scale weakly supervised sound event detection task of Detection and Classification of Acoustic Scenes and Events (DCASE) 2017 challenge. The audio clips in this task, which are extracted from YouTube videos, are manually labeled with one or a few audio tags but without timestamps of the audio events, which is called as weakly labeled data. Two sub-tasks are defined in this challenge including audio tagging and sound event detection using this weakly labeled data. A convolutional recurrent neural network (CRNN) with learnable gated linear units (GLUs) non-linearity applied on the log Mel spectrogram is proposed. In addition, a temporal attention method is proposed along the frames to predicate the locations of each audio event in a chunk from the weakly labeled data. We ranked the 1st and the 2nd as a team in these two sub-tasks of DCASE 2017 challenge with F value 55.6\% and Equal error 0.73, respectively.

cs.SD↗

arXiv2017

Learning Inverse Mapping by Autoencoder based Generative Adversarial Nets

The inverse mapping of GANs'(Generative Adversarial Nets) generator has a great potential value.Hence, some works have been developed to construct the inverse function of generator by directly learning or adversarial learning.While the results are encouraging, the problem is highly challenging and the existing ways of training inverse models of GANs have many disadvantages, such as hard to train or poor performance.Due to these reasons, we propose a new approach based on using inverse generator ($IG$) model as encoder and pre-trained generator ($G$) as decoder of an AutoEncoder network to train the $IG$ model. In the proposed model, the difference between the input and output, which are both the generated image of pre-trained GAN's generator, of AutoEncoder is directly minimized. The optimizing method can overcome the difficulty in training and inverse model of an non one-to-one function.We also applied the inverse model of GANs' generators to image searching and translation.The experimental results prove that the proposed approach works better than the traditional approaches in image searching.

cs.LG↗

arXiv2017

Unconventional quantum Hall effects in two-dimensional massive spin-1 fermion systems

Unconventional fermions with high degeneracies in three dimensions beyond Weyl and Dirac fermions have sparked tremendous interest in condensed matter physics. Here, we study quantum Hall effects (QHEs) in a two-dimensional (2D) unconventional fermion system with a pair of gapped spin-1 fermions. We find that the original unlimited number of zero energy Landau levels (LLs) in the gapless case develop into a series of bands, leading to a novel QHE phenomenon that the Hall conductance first decreases (or increases) to zero and then revives as an infinite ladder of fine staircase when the Fermi surface is moved toward zero energy, and it suddenly reverses with its sign being flipped due to a Van Hove singularity when the Fermi surface is moved across zero. We further investigate the peculiar QHEs in a dice model with a pair of spin-1 fermions, which agree well with the results of the continuous model.

cond-mat.str-el↗

arXiv2017

Model for Topological Phononics and Phonon Diode

The quantum anomalous Hall effect, an exotic topological state first theoretically predicted by Haldane and recently experimentally observed, has attracted enormous interest for low-power-consumption electronics. In this work, we derived a Schr{ö}dinger-like equation of phonons, where topology-related quantities, time reversal symmetry and its breaking can be naturally introduced similar as for electrons. Furthermore, we proposed a phononic analog of the Haldane model, which gives the novel quantum (anomalous) Hall-like phonon states characterized by one-way gapless edge modes immune to scattering. The topologically nontrivial phonon states are useful not only for conducting phonons without dissipation but also for designing highly efficient phononic devices, like an ideal phonon diode, which could find important applications in future phononics.

cond-mat.mes-hall↗

arXiv2017

Berry Phase and Topological Effects of Phonons

Phonons as collective excitations of lattice vibrations are the main heat carriers in solids. Tremendous effort has been devoted to investigate phonons and related properties, giving rise to an intriguing field of phononics, which is of great importance to many practical applications, including heat dissipation, thermal barrier coating, thermoelectrics and thermal control devices. Meanwhile, the research of topology-related physics, awarded the 2016 Nobel Prize in Physics, has led to discoveries of various exotic quantum states of matter, including the quantum (anomalous/spin) Hall [Q(A/S)H] effects, topological insulators/semimetals and topological superconductors. An emerging research field is to bring topological concepts for a new paradigm phononics---"topological phononics". In this Perspective, we will briefly introduce this emerging field and discuss the use of novel quantum degrees of freedom like the Berry phase and topology for manipulating phonons in unprecedentedly new ways.

cond-mat.mtrl-sci↗