arXiv subjects

Yi Yang

Publications and source records attributed to Yi Yang.

At least 973 records · Page 54Linked to original sources

arXiv2019

Syntax-Infused Variational Autoencoder for Text Generation

We present a syntax-infused variational autoencoder (SIVAE), that integrates sentences with their syntactic trees to improve the grammar of generated sentences. Distinct from existing VAE-based text generative models, SIVAE contains two separate latent spaces, for sentences and syntactic trees. The evidence lower bound objective is redesigned correspondingly, by optimizing a joint distribution that accommodates two encoders and two decoders. SIVAE works with long short-term memory architectures to simultaneously generate sentences and syntactic trees. Two versions of SIVAE are proposed: one captures the dependencies between the latent variables through a conditional prior network, and the other treats the latent variables independently such that syntactically-controlled sentence generation can be performed. Experimental results demonstrate the generative superiority of SIVAE on both reconstruction and targeted syntactic evaluations. Finally, we show that the proposed models can be used for unsupervised paraphrasing given different syntactic tree templates.

stat.ML↗

arXiv2019

Spin polarization independence of hard polarized fermion string scattering amplitudes

We calculate a class of polarized fermion string scattering amplitudes (PFSSA) at arbitrary mass levels. We discover that, in the hard scattering limit, the functional forms of the non-vanishing PFSSA at each fixed mass level are independent of the choices of spin polarizations. This result justifies and extends Gross conjecture on high energy string scattering amplitudes to the fermionic sector. In addition, this peculiar property of hard PFSSA is to be compared with the usual spin polarization dependence of the hard polarized fermion field theory scatterings.

hep-th↗

arXiv2019

Cubic LSTMs for Video Prediction

Predicting future frames in videos has become a promising direction of research for both computer vision and robot learning communities. The core of this problem involves moving object capture and future motion prediction. While object capture specifies which objects are moving in videos, motion prediction describes their future dynamics. Motivated by this analysis, we propose a Cubic Long Short-Term Memory (CubicLSTM) unit for video prediction. CubicLSTM consists of three branches, i.e., a spatial branch for capturing moving objects, a temporal branch for processing motions, and an output branch for combining the first two branches to generate predicted frames. Stacking multiple CubicLSTM units along the spatial branch and output branch, and then evolving along the temporal branch can form a cubic recurrent neural network (CubicRNN). Experiment shows that CubicRNN produces more accurate video predictions than prior methods on both synthetic and real-world datasets.

cs.CV↗

arXiv2019

Contrastive Adaptation Network for Unsupervised Domain Adaptation

Unsupervised Domain Adaptation (UDA) makes predictions for the target domain data while manual annotations are only available in the source domain. Previous methods minimize the domain discrepancy neglecting the class information, which may lead to misalignment and poor generalization performance. To address this issue, this paper proposes Contrastive Adaptation Network (CAN) optimizing a new metric which explicitly models the intra-class domain discrepancy and the inter-class domain discrepancy. We design an alternating update strategy for training CAN in an end-to-end manner. Experiments on two real-world benchmarks Office-31 and VisDA-2017 demonstrate that CAN performs favorably against the state-of-the-art methods and produces more discriminative features.

cs.CV↗

arXiv2019

Sim-Real Joint Reinforcement Transfer for 3D Indoor Navigation

There has been an increasing interest in 3D indoor navigation, where a robot in an environment moves to a target according to an instruction. To deploy a robot for navigation in the physical world, lots of training data is required to learn an effective policy. It is quite labour intensive to obtain sufficient real environment data for training robots while synthetic data is much easier to construct by rendering. Though it is promising to utilize the synthetic environments to facilitate navigation training in the real world, real environment are heterogeneous from synthetic environment in two aspects. First, the visual representation of the two environments have significant variances. Second, the houseplans of these two environments are quite different. Therefore two types of information, i.e. visual representation and policy behavior, need to be adapted in the reinforcement model. The learning procedure of visual representation and that of policy behavior are presumably reciprocal. We propose to jointly adapt visual representation and policy behavior to leverage the mutual impacts of environment and policy. Specifically, our method employs an adversarial feature adaptation model for visual representation transfer and a policy mimic strategy for policy behavior imitation. Experiment shows that our method outperforms the baseline by 19.47% without any additional human annotations.

cs.CV↗

arXiv2019

Invariance Matters: Exemplar Memory for Domain Adaptive Person Re-identification

This paper considers the domain adaptive person re-identification (re-ID) problem: learning a re-ID model from a labeled source domain and an unlabeled target domain. Conventional methods are mainly to reduce feature distribution gap between the source and target domains. However, these studies largely neglect the intra-domain variations in the target domain, which contain critical factors influencing the testing performance on the target domain. In this work, we comprehensively investigate into the intra-domain variations of the target domain and propose to generalize the re-ID model w.r.t three types of the underlying invariance, i.e., exemplar-invariance, camera-invariance and neighborhood-invariance. To achieve this goal, an exemplar memory is introduced to store features of the target domain and accommodate the three invariance properties. The memory allows us to enforce the invariance constraints over global training batch without significantly increasing computation cost. Experiment demonstrates that the three invariance properties and the proposed memory are indispensable towards an effective domain adaptation system. Results on three re-ID domains show that our domain adaptation accuracy outperforms the state of the art by a large margin. Code is available at: https://github.com/zhunzhong07/ECN

cs.CV↗

arXiv2019

DM-GAN: Dynamic Memory Generative Adversarial Networks for Text-to-Image Synthesis

In this paper, we focus on generating realistic images from text descriptions. Current methods first generate an initial image with rough shape and color, and then refine the initial image to a high-resolution one. Most existing text-to-image synthesis methods have two main problems. (1) These methods depend heavily on the quality of the initial images. If the initial image is not well initialized, the following processes can hardly refine the image to a satisfactory quality. (2) Each word contributes a different level of importance when depicting different image contents, however, unchanged text representation is used in existing image refinement processes. In this paper, we propose the Dynamic Memory Generative Adversarial Network (DM-GAN) to generate high-quality images. The proposed method introduces a dynamic memory module to refine fuzzy image contents, when the initial images are not well generated. A memory writing gate is designed to select the important text information based on the initial image content, which enables our method to accurately generate images from the text description. We also utilize a response gate to adaptively fuse the information read from the memories and the image features. We evaluate the DM-GAN model on the Caltech-UCSD Birds 200 dataset and the Microsoft Common Objects in Context dataset. Experimental results demonstrate that our DM-GAN model performs favorably against the state-of-the-art approaches.

cs.CV↗

arXiv2019

Operation-aware Neural Networks for User Response Prediction

User response prediction makes a crucial contribution to the rapid development of online advertising system and recommendation system. The importance of learning feature interactions has been emphasized by many works. Many deep models are proposed to automatically learn high-order feature interactions. Since most features in advertising system and recommendation system are high-dimensional sparse features, deep models usually learn a low-dimensional distributed representation for each feature in the bottom layer. Besides traditional fully-connected architectures, some new operations, such as convolutional operations and product operations, are proposed to learn feature interactions better. In these models, the representation is shared among different operations. However, the best representation for different operations may be different. In this paper, we propose a new neural model named Operation-aware Neural Networks (ONN) which learns different representations for different operations. Our experimental results on two large-scale real-world ad click/conversion datasets demonstrate that ONN consistently outperforms the state-of-the-art models in both offline-training environment and online-training environment.

cs.IR↗

arXiv2019

SN 2016hil-- a Type II supernova in the remote outskirts of an elliptical host and its origin

Type II supernovae (SNe) stem from the core collapse of massive ($>8\ M_{\odot}$) stars. Owing to their short lifespan, we expect a very low rate of such events in elliptical host galaxies, where the star-formation rate is low, and which mostly consist of an old stellar population. SN 2016hil (iPTF16hil) is a Type II supernova located in the extreme outskirts of an elliptical galaxy at redshift $z=0.0608$ (projected distance $27.2$ kpc). It was detected near peak brightness ($M_{r} \approx -17$ mag) 9 days after the last nondetection. SN 2016hil has some potentially peculiar properties: while presenting a characteristic spectrum, the event was unusually short lived and declined by $\sim 1.5$ mag in $< 40$ days, following an apparently double-peaked light curve. Its spectra suggest a low metallicity ($Z<0.4\ Z_{\odot}$). We place a tentative upper limit on the mass of a potential faint host at $\log(M/M_{\odot}) =7.27^{+0.43}_{-0.24}$ using deep Keck optical imaging. In light of this, we discuss the possibility of the progenitor forming locally, and other more exotic formation scenarios such as a merger or common-envelope evolution causing a time-delayed explosion. Further observations of the explosion site in the ultraviolet are needed in order to distinguish between the cases. Regardless of the origin of the transient, observing a population of such seemingly hostless Type II SNe could have many uses, including an estimate the number of faint galaxies in a given volume, and tests of the prediction of a time-delayed population of core-collapse SNe in locations otherwise unfavorable for the detection of such events.

astro-ph.HE↗

arXiv2019

Taking A Closer Look at Domain Shift: Category-level Adversaries for Semantics Consistent Domain Adaptation

We consider the problem of unsupervised domain adaptation in semantic segmentation. The key in this campaign consists in reducing the domain shift, i.e., enforcing the data distributions of the two domains to be similar. A popular strategy is to align the marginal distribution in the feature space through adversarial learning. However, this global alignment strategy does not consider the local category-level feature distribution. A possible consequence of the global movement is that some categories which are originally well aligned between the source and target may be incorrectly mapped. To address this problem, this paper introduces a category-level adversarial network, aiming to enforce local semantic consistency during the trend of global alignment. Our idea is to take a close look at the category-level data distribution and align each class with an adaptive adversarial loss. Specifically, we reduce the weight of the adversarial loss for category-level aligned features while increasing the adversarial force for those poorly aligned. In this process, we decide how well a feature is category-level aligned between source and target by a co-training approach. In two domain adaptation tasks, i.e., GTA5 -> Cityscapes and SYNTHIA -> Cityscapes, we validate that the proposed method matches the state of the art in segmentation accuracy.

cs.CV↗

arXiv2019

Significance-aware Information Bottleneck for Domain Adaptive Semantic Segmentation

For unsupervised domain adaptation problems, the strategy of aligning the two domains in latent feature space through adversarial learning has achieved much progress in image classification, but usually fails in semantic segmentation tasks in which the latent representations are overcomplex. In this work, we equip the adversarial network with a "significance-aware information bottleneck (SIB)", to address the above problem. The new network structure, called SIBAN, enables a significance-aware feature purification before the adversarial adaptation, which eases the feature alignment and stabilizes the adversarial training course. In two domain adaptation tasks, i.e., GTA5 -> Cityscapes and SYNTHIA -> Cityscapes, we validate that the proposed method can yield leading results compared with other feature-space alternatives. Moreover, SIBAN can even match the state-of-the-art output-space methods in segmentation accuracy, while the latter are often considered to be better choices for domain adaptive segmentation task.

cs.CV↗

arXiv2019

The Young and Nearby Normal Type Ia Supernova 2018gv: UV-Optical Observations and the Earliest Spectropolarimetry

The non-detection of companion stars in Type Ia supernova (SN) progenitor systems lends support to the notion of double-degenerate (DD) systems and explosions triggered by the merging of two white dwarfs. This very asymmetric process should lead to a conspicuous polarimetric signature. By contrast, observations consistently find very low continuum polarization as the signatures from the explosion process largely dominate over the pre-explosion configuration within several days. Critical information about the interaction of the ejecta with a companion and any circumstellar matter is encoded in the early polarization spectra. In this study, we obtain spectropolarimetry of SN\,2018gv with the ESO Very Large Telescope at $-$13.6 days relative to the $B-$band maximum light, or $\sim$5 days after the estimated explosion --- the earliest spectropolarimetric observations to date of any Type Ia SN. These early observations still show a low continuum polarization ($\lesssim$0.2\%) and moderate line polarization (0.30$\pm$0.04\% for the prominent \ion{Si}{2} $λ$6355 feature and 0.85$\pm$0.04\% for the high-velocity Ca component). The high degree of spherical symmetry implied by the low line and continuum polarization at this early epoch is consistent with explosion models of delayed detonations and is inconsistent with the merger-induced explosion scenario. The dense UV and optical photometry and optical spectroscopy within the first $\sim$100 days after the maximum light indicate that SN\,2018gv is a normal Type Ia SN with similar spectrophotometric behavior to SN\,2011fe.

astro-ph.HE↗

arXiv2019

Discovery of Highly Blueshifted Broad Balmer and Metastable Helium Absorption Lines in a Tidal Disruption Event

We report the discovery of non-stellar hydrogen Balmer and metastable helium absorption lines accompanying a transient, high-velocity (0.05$c$) broad absorption line (BAL) system in the optical spectra of the tidal disruption event (TDE) AT2018zr ($z=0.071$). In the HST UV spectra, absorption of high- and low-ionization lines are also present at this velocity, making AT2018zr resemble a low-ionization broad absorption line (LoBAL) QSO. We conclude that these transient absorption features are more likely to arise in fast outflows produced by the TDE than absorbed by the unbound debris. In accordance with the outflow picture, we are able to reproduce the flat-topped H$α$ emission in a spherically expanding medium, without invoking the typical prescription of an elliptical disk. We also report the appearance of narrow ($\sim$1000~km~s$^{-1}$) NIII$λ$4640, HeII$\lambda4686$, H$α$, and H$β$, emission in the late-time optical spectra of AT2018zr, which may be a result of UV continuum hardening at late time as observed by Swift. Including AT2018zr, we find a high association rate (3 out of 4) of BALs in the UV spectra of TDEs. This suggests that outflows may be ubiquitous among TDEs and may be less sensitive to viewing angle effects compared to QSO outflows.

astro-ph.HE↗

arXiv2019

Holographic Superconductors: An Analytic Method Revisit

We study a non-minimal holographic superconductors model in both non-backreaction and fullbackreaction cases using an analytic method. We calculate the condensate of the dilaton and the critical temperature of the phase transition. We also study the properties of the electric conductivity in various parameters.

hep-th↗

arXiv2019

The SL(K+3,C) Symmetry of the Bosonic String Scattering Amplitudes

We discover that the exact string scattering amplitudes (SSA) of three tachyons and one arbitrary string state, or the Lauricella SSA (LSSA), in the 26D open bosonic string theory can be expressed in terms of the basis functions in the infinite dimensional representation space of the SL(K+3,C) group. In addition, we find that the K+2 recurrence relations among the LSSA discovered by the present authors previously can be used to reproduce the Cartan subalgebra and simple root system of the SL(K+3,C) group with rank K+2. As a result, the SL(K+3,C) group can be used to solve all the LSSA and express them in terms of one amplitude. As an application in the hard scattering limit, the SL(K+3,C) group can be used to directly prove Gross conjecture [1-3], which was previously corrected and proved by the method of decoupling of zero norm states [4-10].

hep-th↗

arXiv2019

Learning to Propagate Labels: Transductive Propagation Network for Few-shot Learning

The goal of few-shot learning is to learn a classifier that generalizes well even when trained with a limited number of training instances per class. The recently introduced meta-learning approaches tackle this problem by learning a generic classifier across a large number of multiclass classification tasks and generalizing the model to a new task. Yet, even with such meta-learning, the low-data problem in the novel classification task still remains. In this paper, we propose Transductive Propagation Network (TPN), a novel meta-learning framework for transductive inference that classifies the entire test set at once to alleviate the low-data problem. Specifically, we propose to learn to propagate labels from labeled instances to unlabeled test instances, by learning a graph construction module that exploits the manifold structure in the data. TPN jointly learns both the parameters of feature embedding and the graph construction in an end-to-end manner. We validate TPN on multiple benchmark datasets, on which it largely outperforms existing few-shot learning approaches and achieves the state-of-the-art results.

cs.LG↗

arXiv2019

Maximal Spontaneous Photon Emission and Energy Loss from Free Electrons

Free electron radiation such as Cerenkov, Smith--Purcell, and transition radiation can be greatly affected by structured optical environments, as has been demonstrated in a variety of polaritonic, photonic-crystal, and metamaterial systems. However, the amount of radiation that can ultimately be extracted from free electrons near an arbitrary material structure has remained elusive. Here we derive a fundamental upper limit to the spontaneous photon emission and energy loss of free electrons, regardless of geometry, which illuminates the effects of material properties and electron velocities. We obtain experimental evidence for our theory with quantitative measurements of Smith--Purcell radiation. Our framework allows us to make two predictions. One is a new regime of radiation operation---at subwavelength separations, slower (nonrelativistic) electrons can achieve stronger radiation than fast (relativistic) electrons. The second is a divergence of the emission probability in the limit of lossless materials. We further reveal that such divergences can be approached by coupling free electrons to photonic bound states in the continuum (BICs). Our findings suggest that compact and efficient free-electron radiation sources from microwaves to the soft X-ray regime may be achievable without requiring ultrahigh accelerating voltages.

physics.optics↗

arXiv2019

A General Theoretical and Experimental Framework for Nanoscale Electromagnetism

Local, bulk response functions, e.g permittivity, and the macroscopic Maxwell equations completely specify the classical electromagnetic problem, which features only wavelength $λ$ and geometric scales. The above neglect of intrinsic electronic length scales $L_{\text{e}}$ leads to an eventual breakdown in the nanoscopic limit. Here, we present a general theoretical and experimental framework for treating nanoscale electromagnetic phenomena. The framework features surface-response functions---known as the Feibelman $d$-parameters---which reintroduce the missing electronic length scales. As a part of our framework, we establish an experimental procedure to measure these complex, dispersive surface response functions, enabled by quasi-normal-mode perturbation theory and observations of pronounced nonclassical effects---spectral shifts in excess of 30% and the breakdown of Kreibig-like broadening---in a quintessential multiscale architecture: film-coupled nanoresonators, with feature-sizes comparable to both $L_{\text{e}}$ and $λ$.

physics.optics↗