arXiv subjects

Yang Zou

Publications and source records attributed to Yang Zou.

At least 55 records · Page 3Linked to original sources

arXiv2020

Joint Disentangling and Adaptation for Cross-Domain Person Re-Identification

Although a significant progress has been witnessed in supervised person re-identification (re-id), it remains challenging to generalize re-id models to new domains due to the huge domain gaps. Recently, there has been a growing interest in using unsupervised domain adaptation to address this scalability issue. Existing methods typically conduct adaptation on the representation space that contains both id-related and id-unrelated factors, thus inevitably undermining the adaptation efficacy of id-related features. In this paper, we seek to improve adaptation by purifying the representation space to be adapted. To this end, we propose a joint learning framework that disentangles id-related/unrelated features and enforces adaptation to work on the id-related feature space exclusively. Our model involves a disentangling module that encodes cross-domain images into a shared appearance space and two separate structure spaces, and an adaptation module that performs adversarial alignment and self-training on the shared appearance space. The two modules are co-designed to be mutually beneficial. Extensive experiments demonstrate that the proposed joint learning framework outperforms the state-of-the-art methods by clear margins.

cs.CV↗

arXiv2020

Confidence Regularized Self-Training

Recent advances in domain adaptation show that deep self-training presents a powerful means for unsupervised domain adaptation. These methods often involve an iterative process of predicting on target domain and then taking the confident predictions as pseudo-labels for retraining. However, since pseudo-labels can be noisy, self-training can put overconfident label belief on wrong classes, leading to deviated solutions with propagated errors. To address the problem, we propose a confidence regularized self-training (CRST) framework, formulated as regularized self-training. Our method treats pseudo-labels as continuous latent variables jointly optimized via alternating optimization. We propose two types of confidence regularization: label regularization (LR) and model regularization (MR). CRST-LR generates soft pseudo-labels while CRST-MR encourages the smoothness on network output. Extensive experiments on image classification and semantic segmentation show that CRSTs outperform their non-regularized counterpart with state-of-the-art performance. The code and models of this work are available at https://github.com/yzou2/CRST.

cs.CV↗

arXiv2019

Conservative Wasserstein Training for Pose Estimation

This paper targets the task with discrete and periodic class labels ($e.g.,$ pose/orientation estimation) in the context of deep learning. The commonly used cross-entropy or regression loss is not well matched to this problem as they ignore the periodic nature of the labels and the class similarity, or assume labels are continuous value. We propose to incorporate inter-class correlations in a Wasserstein training framework by pre-defining ($i.e.,$ using arc length of a circle) or adaptively learning the ground metric. We extend the ground metric as a linear, convex or concave increasing function $w.r.t.$ arc length from an optimization perspective. We also propose to construct the conservative target labels which model the inlier and outlier noises using a wrapped unimodal-uniform mixture distribution. Unlike the one-hot setting, the conservative label makes the computation of Wasserstein distance more challenging. We systematically conclude the practical closed-form solution of Wasserstein distance for pose data with either one-hot or conservative target label. We evaluate our method on head, body, vehicle and 3D object pose benchmarks with exhaustive ablation studies. The Wasserstein loss obtaining superior performance over the current methods, especially using convex mapping function for ground metric, conservative label, and closed-form solution.

cs.CV↗

arXiv2019

Deep Classification Network for Monocular Depth Estimation

Monocular Depth Estimation is usually treated as a supervised and regression problem when it actually is very similar to semantic segmentation task since they both are fundamentally pixel-level classification tasks. We applied depth increments that increases with depth in discretizing depth values and then applied Deeplab v2 and the result was higher accuracy. We were able to achieve a state-of-the-art result on the KITTI dataset and outperformed existing architecture by an 8% margin.

cs.CV↗

arXiv2019

Unsupervised Domain Adaptation via Calibrating Uncertainties

Unsupervised domain adaptation (UDA) aims at inferring class labels for unlabeled target domain given a related labeled source dataset. Intuitively, a model trained on source domain normally produces higher uncertainties for unseen data. In this work, we build on this assumption and propose to adapt from source to target domain via calibrating their predictive uncertainties. The uncertainty is quantified as the Renyi entropy, from which we propose a general Renyi entropy regularization (RER) framework. We further employ variational Bayes learning for reliable uncertainty estimation. In addition, calibrating the sample variance of network parameters serves as a plug-in regularizer for training. We discuss the theoretical properties of the proposed method and demonstrate its effectiveness on three domain-adaptation tasks.

cs.LG↗

arXiv2018

Simultaneous Edge Alignment and Learning

Edge detection is among the most fundamental vision problems for its role in perceptual grouping and its wide applications. Recent advances in representation learning have led to considerable improvements in this area. Many state of the art edge detection models are learned with fully convolutional networks (FCNs). However, FCN-based edge learning tends to be vulnerable to misaligned labels due to the delicate structure of edges. While such problem was considered in evaluation benchmarks, similar issue has not been explicitly addressed in general edge learning. In this paper, we show that label misalignment can cause considerably degraded edge learning quality, and address this issue by proposing a simultaneous edge alignment and learning framework. To this end, we formulate a probabilistic model where edge alignment is treated as latent variable optimization, and is learned end-to-end during network training. Experiments show several applications of this work, including improved edge detection with state of the art performance, and automatic refinement of noisy annotations.

cs.CV↗

arXiv2018

Domain Adaptation for Semantic Segmentation via Class-Balanced Self-Training

Recent deep networks achieved state of the art performance on a variety of semantic segmentation tasks. Despite such progress, these models often face challenges in real world `wild tasks' where large difference between labeled training/source data and unseen test/target data exists. In particular, such difference is often referred to as `domain gap', and could cause significantly decreased performance which cannot be easily remedied by further increasing the representation power. Unsupervised domain adaptation (UDA) seeks to overcome such problem without target domain labels. In this paper, we propose a novel UDA framework based on an iterative self-training procedure, where the problem is formulated as latent variable loss minimization, and can be solved by alternatively generating pseudo labels on target data and re-training the model with these labels. On top of self-training, we also propose a novel class-balanced self-training framework to avoid the gradual dominance of large classes on pseudo-label generation, and introduce spatial priors to refine generated labels. Comprehensive experiments show that the proposed methods achieve state of the art semantic segmentation performance under multiple major UDA settings.

cs.CV↗

arXiv2018

Scale Optimization for Full-Image-CNN Vehicle Detection

Many state-of-the-art general object detection methods make use of shared full-image convolutional features (as in Faster R-CNN). This achieves a reasonable test-phase computation time while enjoys the discriminative power provided by large Convolutional Neural Network (CNN) models. Such designs excel on benchmarks which contain natural images but which have very unnatural distributions, i.e. they have an unnaturally high-frequency of the target classes and a bias towards a "friendly" or "dominant" object scale. In this paper we present further study of the use and adaptation of the Faster R-CNN object detection method for datasets presenting natural scale distribution and unbiased real-world object frequency. In particular, we show that better alignment of the detector scale sensitivity to the extant distribution improves vehicle detection performance. We do this by modifying both the selection of Region Proposals, and through using more scale-appropriate full-image convolution features within the CNN model. By selecting better scales in the region proposal input and by combining feature maps through careful design of the convolutional neural network, we improve performance on smaller objects. We significantly increase detection AP for the KITTI dataset car class from 76.3% on our baseline Faster R-CNN detector to 83.6% in our improved detector.

cs.CV↗

arXiv2015

Experimental detection of polarization-frequency quantum correlations in a photonic quantum channel by local operations

The measurement of correlations between different degrees of freedom is an important, but in general extremely difficult task in many applications of quantum mechanics. Here, we report an all-optical experimental detection and quantification of quantum correlations between the polarization and the frequency degrees of freedom of single photons by means of local operations acting only on the polarization degree of freedom. These operations only require experimental control over an easily accessible two-dimensional subsystem, despite handling strongly mixed quantum states comprised of a continuum of orthogonal frequency states. Our experiment thus represents a photonic realization of a scheme for the local detection of quantum correlations in a truly infinite-dimensional continuous-variable system, which excludes an efficient finite-dimensional truncation.

quant-ph↗

arXiv2015

Sliced Wasserstein Kernels for Probability Distributions

Optimal transport distances, otherwise known as Wasserstein distances, have recently drawn ample attention in computer vision and machine learning as a powerful discrepancy measure for probability distributions. The recent developments on alternative formulations of the optimal transport have allowed for faster solutions to the problem and has revamped its practical applications in machine learning. In this paper, we exploit the widely used kernel methods and provide a family of provably positive definite kernels based on the Sliced Wasserstein distance and demonstrate the benefits of these kernels in a variety of learning tasks. Our work provides a new perspective on the application of optimal transport flavored distances through kernel methods in machine learning tasks.

cs.LG↗

arXiv2015

Storage of multiple single-photon pulses emitted from a quantum dot in a solid-state quantum memory

Quantum repeaters are critical components for distributing entanglement over long distances in presence of unavoidable optical losses during transmission. Stimulated by Duan-Lukin-Cirac-Zoller protocol, many improved quantum-repeater protocols based on quantum memories have been proposed, which commonly focus on the entanglement-distribution rate. Among these protocols, the elimination of multi-photons (multi-photon-pairs) and the use of multimode quantum memory are demonstrated to have the ability to greatly improve the entanglement-distribution rate. Here, we demonstrate the storage of deterministic single photons emitted from a quantum dot in a polarization-maintaining solid-state quantum memory; in addition, multi-temporal-mode memory with $1$, $20$ and $100$ narrow single-photon pulses is also demonstrated. Multi-photons are eliminated, and only one photon at most is contained in each pulse. Moreover, the solid-state properties of both sub-systems make this configuration more stable and easier to be scalable. Our work will be helpful in the construction of efficient quantum repeaters based on all-solid-state devices

quant-ph↗

arXiv2015

Unconventional superconductivity in the layered iron germanide YFe$_2$Ge$_2$

Since the discovery of superconductivity in LaFePO in 2006, numerous iron-based superconductors have been identified within diverse structure families, all of which combine iron with a group-V (pnictogen) or group-VI (chalco- gen) element. Unconventional superconductivity is extremely rare among transition metal compounds outside these layered iron systems and the cuprates, and it is almost universally associated with highly anisotropic electronic properties and nearly 2D Fermi surface geometries. The iron-based intermetallic YFe$_2$Ge$_2$ features a 3D Fermi surface and a strongly enhanced low temperature heat capacity, which signals strong electronic correlations. We present data from a new generation of high quality samples of YFe$_2$Ge$_2$, which show superconducting transition anomalies below 1.8 K in thermodynamic as well as transport measurements, establishing that superconductivity is intrinsic in this layered iron compound outside the known superconducting iron pnictide or chalcogenide families. The Fermi surface geometry of YFe$_2$Ge$_2$ resembles that of KFe$_2$As$_2$ in the high pressure collapsed tetragonal phase, in which superconductivity at temperatures as high as 10 K has recently been reported, suggesting an underlying connection between the two systems.

cond-mat.str-el↗

arXiv2013

Experimental test of the tradeoff relation in weak measurement

An upper bound between the information gain and state reversibility of weak measurement was first developed by Y. K. Cheong and S. W. Lee [Phys. Rev. Lett. 109, 150402 (2012)]. Their results are valid for arbitrary d-level quantum systems. In light of the commonly used qubit system in quantum information, a sharp tradeoff relation can be obtained. In this letter, this tradeoff relation is experimentally verified with polarization encoded single photons from a quantum dot. Furthermore, a complete traversal of weak measurement operators is realized, and the mapping to the least upper bound of this tradeoff relation is obtained. Our results complement the theoretical work and provide a universal ruler for the characterization of weak measurements.

quant-ph↗

arXiv2011

Non-classical correlation of cascaded photon pairs emitted from quantum dot

We studied the quantum correlation between the photon pairs generated by biexciton cascade decays of self-assembled quantum dots, and determined the temperature behavior associated with so-called sudden change of the quantum correlation. The relationship between the fine structure splitting and the sudden change temperature is also provided. Our study indicates that this correlation behavior sudden change temperature is independent on the back ground noise in the system and far lower than entanglement sudden death temperature, therefore it should be easier to observe the phenomenon of correlation sudden change in experiments than to observe entanglement sudden death.

quant-ph↗

arXiv2011

Violation of Leggett-Garg inequalities in single quantum dot

We investigate the violation of Leggett-Garg (LG) inequalities inquantum dots with the stationarity assumption. By comparing two types of LG inequalities, we find a better one which is easier to be tested in experiment. In addition, we show that the fine-structure splitting, background noise and temperature of quantum dots all influence the violation of LG inequalities.

quant-ph↗

arXiv2011

Evidence for Interlayer Electronic Coupling in Multilayer Epitaxial Graphene from Polarization Dependent Coherently Controlled Photocurrent Generation

Most experimental studies to date of multilayer epitaxial graphene on C-face SiC have indicated that the electronic states of different layers are decoupled as a consequence of rotational stacking. We have measured the third order nonlinear tensor in epitaxial graphene as a novel approach to probe interlayer electronic coupling, by studying THz emission from coherently controlled photocurrents as a function of the optical pump and THz beam polarizations. We find that the polarization dependence of the coherently controlled THz emission expected from perfectly uncoupled layers, i.e. a single graphene sheet, is not observed. We hypothesize that the observed angular dependence arises from weak coupling between the layers; a model calculation of the angular dependence treating the multilayer structure as a stack of independent bilayers with variable interlayer coupling qualitatively reproduces the polarization dependence, providing evidence for coupling.

cond-mat.mes-hall↗

arXiv2011

Experimental demonstration of nonlocal effects of the partial collapse measurement and reversal process

We experimentally demonstrate the nonlocal reversal of a partial-collapse quantum measurement on two-photon entangled state. Both the partial measurement and the reversal operation are implemented in linear optics with two displaced Sagnac interferometers, which are characterized by single qubit quantum process tomography. The recovered state is measured by quantum state tomography and its nonlocality is characterized by testing the Bell inequality. Our result will be helpful in quantum communication and quantum error correction.

quant-ph↗

arXiv2010

Non-Markovian Entanglement Sudden Death and Rebirth of a Two-Qubit System in the Presence of System-Bath Coherence

We present a detailed study of the entanglement dynamics of a two-qubit system coupled to independent non-Markovian environments, employing hierarchy equations. This recently developed theoretical treatment can conveniently solve non-Markovian problems and take into consideration the correlation between the system and bath in an initial state. We concentrate on calculating the death and rebirth time points of the entanglement to obtain a general view of the concurrence curve and explore the behavior of entanglement dynamics with respect to the coupling strength, the characteristic frequency of the noise bath and the environment temperature.

quant-ph↗