Searcharxiv⌕ Search

arXiv subjects

Peng Gao

Publications and source records attributed to Peng Gao.

At least 379 records · Page 21Linked to original sources
arXiv2020

Controllable generations of several nonlinear waves in optical fibers with third-order dispersion

We propose a method to controllably generate six kinds of nonlinear waves on continuous waves, including the one- and multi-peak solitons, the Akhmediev, Kuznetsov-Ma, and Taijiri-Watanabe breathers, and stable periodic waves. In the nonlinear fiber system with third-order dispersion, we illustrate their generation conditions by the modified linear stability analysis, and numerically generate them from initial perturbations on continuous waves. We implement the quantitative control over their dynamical features, including the wave type, velocity, periodicity, and localization. Our results may provide an effective scheme for generating optical solitons on continuous waves, and it can also be applied for wave generations in other various nonlinear systems.

nlin.PS↗
arXiv2020

Broad-Spectral-Range Sustainability and Controllable Excitation of Hyperbolic Phonon Polaritons in $α$-MoO3

Hyperbolic phonon polaritons (HPhPs) in orthorhombic-phase molybdenum trioxide ($α$-MoO3) show in-plane hyperbolicity, great wavelength compression and ultra-long lifetime, therefore holding great potential in nanophotonic applications. However, its polaritonic response in the far-infrared (FIR) range has long remained unexplored due to challenges in experimental characterization. Here, using monochromated electron energy loss spectroscopy (EELS) in a scanning transmission electron microscope (STEM), we probe HPhPs in $α$-MoO3 in both mid-infrared (MIR) and FIR frequencies and correlate their behaviors with microstructures and orientations. We find that low-structural symmetry leads to various phonon modes and multiple Reststrahlen bands (RBs) over a broad spectral range (over 70 meV) and in different directions (55-63 meV and 119-125 meV along b axis, 68-106 meV along c axis, 101-121 meV along a axis). These HPhPs can be selectively excited by controlling the direction of swift electrons. These findings provide new opportunities in nanophotonic and optoelectronic applications such as directed light propagation, hyperlenses and heat transfer.

cond-mat.mtrl-sci↗
arXiv2020

Multi-Channel Automatic Speech Recognition Using Deep Complex Unet

The front-end module in multi-channel automatic speech recognition (ASR) systems mainly use microphone array techniques to produce enhanced signals in noisy conditions with reverberation and echos. Recently, neural network (NN) based front-end has shown promising improvement over the conventional signal processing methods. In this paper, we propose to adopt the architecture of deep complex Unet (DCUnet) - a powerful complex-valued Unet-structured speech enhancement model - as the front-end of the multi-channel acoustic model, and integrate them in a multi-task learning (MTL) framework along with cascaded framework for comparison. Meanwhile, we investigate the proposed methods with several training strategies to improve the recognition accuracy on the 1000-hours real-world XiaoMi smart speaker data with echos. Experiments show that our proposed DCUnet-MTL method brings about 12.2% relative character error rate (CER) reduction compared with the traditional approach with array processing plus single-channel acoustic model. It also achieves superior performance than the recently proposed neural beamforming method.

cs.SD↗
arXiv2020

A Predictive Autoscaler for Elastic Batch Jobs

Large batch jobs such as Deep Learning, HPC and Spark require far more computational resources and higher cost than conventional online service. Like the processing of other time series data, these jobs possess a variety of characteristics such as trend, burst, and seasonality. Cloud providers offer short-term instances to achieve scalability, stability, and cost-efficiency. Given the time lag caused by joining into the cluster and initialization, crowded workloads may lead to a violation in the scheduling system. Based on the assumption that there are infinite resources and ideal placements available for users to require in the cloud environment, we propose a predictive autoscaler to provide an elastic interface for the customers and overprovision instances based on the trained regression model. We contribute to a method to embed heterogeneous resource requirements in continuous space into discrete resource buckets and an autoscaler to do predictive expand plans on the time series of resource bucket counts. Our experimental evaluation of the production resources usage data validates the solution and the results show that the predictive autoscaler relieves the burden of making scaling plans, avoids long launching time at lower cost and outperforms other prediction methods with fine-tuned settings.

cs.LG↗
arXiv2020

Multi-Pass Transformer for Machine Translation

In contrast with previous approaches where information flows only towards deeper layers of a stack, we consider a multi-pass transformer (MPT) architecture in which earlier layers are allowed to process information in light of the output of later layers. To maintain a directed acyclic graph structure, the encoder stack of a transformer is repeated along a new multi-pass dimension, keeping the parameters tied, and information is allowed to proceed unidirectionally both towards deeper layers within an encoder stack and towards any layer of subsequent stacks. We consider both soft (i.e., continuous) and hard (i.e., discrete) connections between parallel encoder stacks, relying on a neural architecture search to find the best connection pattern in the hard case. We perform an extensive ablation study of the proposed MPT architecture and compare it with other state-of-the-art transformer architectures. Surprisingly, Base Transformer equipped with MPT can surpass the performance of Large Transformer on the challenging machine translation En-De and En-Fr datasets. In the hard connection case, the optimal connection pattern found for En-De also leads to improved performance for En-Fr.

cs.CL↗
arXiv2020

Zone-folded longnitude acoustic phonons driving self-trapped state emission in colloidal CdSe nanoplate superlattice

Colloidal cadmium chalcogenide nanoplates are two-dimensional semiconductors that have shown great application prospect for light-emitting technologies. Self-trapped state (STS), a special localized state originated from strong electron-phonon coupling (EPC), has great potential in one-step white light luminance owing to its broadband emission linewidth. However, achieving STS in cadmium chalcogenide nanocrystals is extremely challenging due to their intrinic weak EPC nature. By building hybrid superlattice (SL) structures via self-assembly of colloidal CdSe nanoplates (NPLs), we demonstrated an emergence of zone-folded longnitude acoustic phonons (ZFLAP) differ from monodispersed NPLs, and observed a broadband STS emission in spectra range of 450-600 nm. Through femtosecond transient absorption and impulsive vibrational spectroscopy, we revealed that STS is generated in time scale of ~500 fs and is driven by strong coupling of excitons and ZFLAPs with Huang-Rhys parameter as large as ~22.7. Our findings provide a new avenue for generating and manipulating STS emission by artificially designing and building hybrid periodic structures superior to single material optimization.

cond-mat.mtrl-sci↗
arXiv2020

Four-dimensional Vibrational Spectroscopy for Nanoscale Mapping of Phonon Dispersion in BN Nanotubes

Direct measurement of local phonon dispersion in individual nanostructures can greatly advance our understanding of their electrical, thermal, and mechanical properties. However, such experimental measurements require extremely high detection sensitivity and combined spatial, energy and momentum resolutions, thus has been elusive. Here, we develop a four-dimensional electron energy loss spectroscopy (4D-EELS) technique based a monochromated scanning transmission electron microscope (STEM), and present the position-dependent phonon dispersion measurement in individual boron nitride nanotubes (BNNTs). Our measurement shows that the unfolded phonon dispersion of multi-walled BNNTs is close to hexagonal-boron nitride (h-BN) crystals, suggesting that interlayer coupling and curved geometry have no substantial impacts on phonon dispersion. We also find that the acoustic phonons are extremely sensitive to momentum-dependent defect scattering, while optical phonons are much less susceptible. This work not only provides useful insights into vibrational properties of BNNTs, but also demonstrates huge prospects of the developed 4D-EELS technique in nanoscale phonon dispersion measurements.

cond-mat.mes-hall↗
arXiv2020

Low-lying zeros of a family of quadratic Hecke $L$-functions via ratios conjecture

In this paper, we apply the ratio conjecture of $L$-functions to derive the lower order terms of the $1$-level density of the low-lying zeros of a family quadratic Hecke $L$-functions in the Gaussian field. Up to the first lower order term, we show that our result is consistent with that obtained from previous work under the generalized Riemann hypothesis, when the Fourier transforms of the test functions are supported in $(-2, 2)$.

math.NT↗
arXiv2020

Creating topological polar structure in a nonpolar matter

Nontrivial topological structures offer rich playground in condensed matter physics including fluid dynamics, superconductivity, and ferromagnetism, and they promise alternative device configurations for post-Moore spintronics and electronics. Indeed, magnetic skyrmions are actively pursued for high-density data storage, while polar vortices with exotic negative capacitance may enable ultralow power consumption in microelectronics. Following extensive investigations on a variety of magnetic textures including vortices, domain walls and skyrmions in the past decades, studies on polar topologies have taken off in recent years, resulting in discoveries of closure domains, vortices, and skyrmions in ferroelectric materials. Nevertheless, the atomic-scale creation of topological polar structures is largely confined in a single ferroelectric system, PbTiO3 (PTO) with large polarization, casting doubt on the generality of polar topologies and limiting their potential applications. In this work, we successfully create previously unrealized atomic-scale polar antivortices in the nominally nonpolar SrTiO3 (STO), expanding the reaches of topological structures and completing an important missing link in polar topologies. The work shed considerable new insight into the formation of topological polar structures, and offers guidance in searching for new polar textures.

cond-mat.mtrl-sci↗
arXiv2020

Reconstruction Regularized Deep Metric Learning for Multi-label Image Classification

In this paper, we present a novel deep metric learning method to tackle the multi-label image classification problem. In order to better learn the correlations among images features, as well as labels, we attempt to explore a latent space, where images and labels are embedded via two unique deep neural networks, respectively. To capture the relationships between image features and labels, we aim to learn a \emph{two-way} deep distance metric over the embedding space from two different views, i.e., the distance between one image and its labels is not only smaller than those distances between the image and its labels' nearest neighbors, but also smaller than the distances between the labels and other images corresponding to the labels' nearest neighbors. Moreover, a reconstruction module for recovering correct labels is incorporated into the whole framework as a regularization term, such that the label embedding space is more representative. Our model can be trained in an end-to-end manner. Experimental results on publicly available image datasets corroborate the efficacy of our method compared with the state-of-the-arts.

cs.CV↗
arXiv2020

Contrastive Visual-Linguistic Pretraining

Several multi-modality representation learning approaches such as LXMERT and ViLBERT have been proposed recently. Such approaches can achieve superior performance due to the high-level semantic information captured during large-scale multimodal pretraining. However, as ViLBERT and LXMERT adopt visual region regression and classification loss, they often suffer from domain gap and noisy label problems, based on the visual features having been pretrained on the Visual Genome dataset. To overcome these issues, we propose unbiased Contrastive Visual-Linguistic Pretraining (CVLP), which constructs a visual self-supervised loss built upon contrastive learning. We evaluate CVLP on several down-stream tasks, including VQA, GQA and NLVR2 to validate the superiority of contrastive learning on multi-modality representation learning. Our code is available at: https://github.com/ArcherYunDong/CVLP-.

cs.CV↗
arXiv2020

Gradient Regularized Contrastive Learning for Continual Domain Adaptation

Human beings can quickly adapt to environmental changes by leveraging learning experience. However, the poor ability of adapting to dynamic environments remains a major challenge for AI models. To better understand this issue, we study the problem of continual domain adaptation, where the model is presented with a labeled source domain and a sequence of unlabeled target domains. There are two major obstacles in this problem: domain shifts and catastrophic forgetting. In this work, we propose Gradient Regularized Contrastive Learning to solve the above obstacles. At the core of our method, gradient regularization plays two key roles: (1) enforces the gradient of contrastive loss not to increase the supervised training loss on the source domain, which maintains the discriminative power of learned features; (2) regularizes the gradient update on the new domain not to increase the classification loss on the old target domains, which enables the model to adapt to an in-coming target domain while preserving the performance of previously observed domains. Hence our method can jointly learn both semantically discriminative and domain-invariant features with labeled source domain and unlabeled target domains. The experiments on Digits, DomainNet and Office-Caltech benchmarks demonstrate the strong performance of our approach when compared to the state-of-the-art.

cs.CV↗
arXiv2020

Learning Where to Focus for Efficient Video Object Detection

Transferring existing image-based detectors to the video is non-trivial since the quality of frames is always deteriorated by part occlusion, rare pose, and motion blur. Previous approaches exploit to propagate and aggregate features across video frames by using optical flow-warping. However, directly applying image-level optical flow onto the high-level features might not establish accurate spatial correspondences. Therefore, a novel module called Learnable Spatio-Temporal Sampling (LSTS) has been proposed to learn semantic-level correspondences among adjacent frame features accurately. The sampled locations are first randomly initialized, then updated iteratively to find better spatial correspondences guided by detection supervision progressively. Besides, Sparsely Recursive Feature Updating (SRFU) module and Dense Feature Aggregation (DFA) module are also introduced to model temporal relations and enhance per-frame features, respectively. Without bells and whistles, the proposed method achieves state-of-the-art performance on the ImageNet VID dataset with less computational complexity and real-time speed. Code will be made available at https://github.com/jiangzhengkai/LSTS.

cs.CV↗
arXiv2020

Eightfold Fermionic Excitation in a Charge Density Wave Compound

Unconventional quasiparticle excitations in condensed matter systems have become one of the most important research frontiers. Beyond two- and fourfold degenerate Weyl and Dirac fermions, three-, six- and eightfold symmetry protected degeneracies have been predicted however remain challenging to realize in solid state materials. Here, charge density wave compound TaTe4 is proposed to hold eightfold fermionic excitation and Dirac point in energy bands. High quality TaTe4 single crystals are prepared, where the charge density wave is revealed by directly imaging the atomic structure and a pseudogap of about 45 meV on the surface. Shubnikov de-Haas oscillations of TaTe4 are consistent with band structure calculation. Scanning tunneling microscopy reveals atomic step edge states on the surface of TaTe4. This work uncovers that charge density wave is able to induce new topological phases and sheds new light on the novel excitations in condensed matter materials.

cond-mat.mtrl-sci↗
arXiv2020

Extreme Low-Light Imaging with Multi-granulation Cooperative Networks

Low-light imaging is challenging since images may appear to be dark and noised due to low signal-to-noise ratio, complex image content, and the variety in shooting scenes in extreme low-light condition. Many methods have been proposed to enhance the imaging quality under extreme low-light conditions, but it remains difficult to obtain satisfactory results, especially when they attempt to retain high dynamic range (HDR). In this paper, we propose a novel method of multi-granulation cooperative networks (MCN) with bidirectional information flow to enhance extreme low-light images, and design an illumination map estimation function (IMEF) to preserve high dynamic range (HDR). To facilitate this research, we also contribute to create a new benchmark dataset of real-world Dark High Dynamic Range (DHDR) images to evaluate the performance of high dynamic preservation in low light environment. Experimental results show that the proposed method outperforms the state-of-the-art approaches in terms of both visual effects and quantitative analysis.

eess.IV↗
arXiv2020

Character Matters: Video Story Understanding with Character-Aware Relations

Different from short videos and GIFs, video stories contain clear plots and lists of principal characters. Without identifying the connection between appearing people and character names, a model is not able to obtain a genuine understanding of the plots. Video Story Question Answering (VSQA) offers an effective way to benchmark higher-level comprehension abilities of a model. However, current VSQA methods merely extract generic visual features from a scene. With such an approach, they remain prone to learning just superficial correlations. In order to attain a genuine understanding of who did what to whom, we propose a novel model that continuously refines character-aware relations. This model specifically considers the characters in a video story, as well as the relations connecting different characters and objects. Based on these signals, our framework enables weakly-supervised face naming through multi-instance co-occurrence matching and supports high-level reasoning utilizing Transformer structures. We train and test our model on the six diverse TV shows in the TVQA dataset, which is by far the largest and only publicly available dataset for VSQA. We validate our proposed approach over TVQA dataset through extensive ablation study.

cs.CV↗