arXiv subjects

Jian Sun

Publications and source records attributed to Jian Sun.

At least 397 records · Page 22Linked to original sources

arXiv2019

Learnable Tree Filter for Structure-preserving Feature Transform

Learning discriminative global features plays a vital role in semantic segmentation. And most of the existing methods adopt stacks of local convolutions or non-local blocks to capture long-range context. However, due to the absence of spatial structure preservation, these operators ignore the object details when enlarging receptive fields. In this paper, we propose the learnable tree filter to form a generic tree filtering module that leverages the structural property of minimal spanning tree to model long-range dependencies while preserving the details. Furthermore, we propose a highly efficient linear-time algorithm to reduce resource consumption. Thus, the designed modules can be plugged into existing deep neural networks conveniently. To this end, tree filtering modules are embedded to formulate a unified framework for semantic segmentation. We conduct extensive ablation studies to elaborate on the effectiveness and efficiency of the proposed method. Specifically, it attains better performance with much less overhead compared with the classic PSP block and Non-local operation under the same backbone. Our approach is proved to achieve consistent improvements on several benchmarks without bells-and-whistles. Code and models are available at https://github.com/StevenGrove/TreeFilter-Torch.

cs.CV↗

arXiv2019

3/2 Fractional Quantum Hall Plateau in Confined Two-Dimensional Electron Gas

Even-denominator fractional quantum Hall (FQH) states, such as 5/2 and 7/2, have been well known in a two-dimensional electron gas (2DEG) for decades and are still investigated as candidates of non-Abelian statistics. In this paper, we present the observation of a 3/2 FQH plateau in a single-layer 2DEG with lateral confinement at a bulk filling factor of 5/3. The 3/2 FQH plateau is quantized at (h/e^2 )/(3/2) within 0.02%, and can survive up to 300 mK. This even-denominator FQH plateau may imply intriguing edge structure and excitation in FQH system with lateral confinement. The observations in this work demonstrate that understanding the effect of the lateral confinement on the many-body system is critical in the pursuit of important theoretical proposals involving edge physics, such as the demonstration of non-Abelian statistics and the realization of braiding for fault-tolerant quantum computation.

cond-mat.mes-hall↗

arXiv2019

Disentangled Image Matting

Most previous image matting methods require a roughly-specificed trimap as input, and estimate fractional alpha values for all pixels that are in the unknown region of the trimap. In this paper, we argue that directly estimating the alpha matte from a coarse trimap is a major limitation of previous methods, as this practice tries to address two difficult and inherently different problems at the same time: identifying true blending pixels inside the trimap region, and estimate accurate alpha values for them. We propose AdaMatting, a new end-to-end matting framework that disentangles this problem into two sub-tasks: trimap adaptation and alpha estimation. Trimap adaptation is a pixel-wise classification problem that infers the global structure of the input image by identifying definite foreground, background, and semi-transparent image regions. Alpha estimation is a regression problem that calculates the opacity value of each blended pixel. Our method separately handles these two sub-tasks within a single deep convolutional neural network (CNN). Extensive experiments show that AdaMatting has additional structure awareness and trimap fault-tolerance. Our method achieves the state-of-the-art performance on Adobe Composition-1k dataset both qualitatively and quantitatively. It is also the current best-performing method on the alphamatting.com online evaluation for all commonly-used metrics.

cs.CV↗

arXiv2019

A dynamical mean-field study of orbital-selective Mott phase enhanced by next-nearest neighbor hopping

The dynamical mean-field theory is employed to study the orbital-selective Mott transition (OSMT) of the two-orbital Hubbard model with nearest neighbor hopping and next-nearest neighbor (NNN) hopping. The NNN hopping breaks the particle-hole symmetry at half filling and gives rise to an asymmetric density of states (DOS). Our calculations show that the broken symmetry of DOS benefits the OSMT, where the region of the orbital-selective Mott phase significantly extends with the increasing NNN hopping integral. We also find that Hund's rule coupling promotes OSMT by blocking the orbital fluctuations, but the influence of NNN hopping is more remarkable.

cond-mat.str-el↗

arXiv2019

Doublon-holon excitations split by Hund's rule coupling within the orbital-selective Mott phase

Multiorbital interactions have the capacity to produce an interesting kind of doublon-holon bound state that consists of a single-hole state in one band and a doubly-occupied state in another band. Interband doublon-holon pair excitations in the two-orbital Hubbard model are studied by using dynamical mean-field theory with the Lanczos method as the impurity solver. We find that the interband bound states may provide several in-gap quasiparticle peaks in the density of states of the narrow band in the orbital-selective Mott phase with a small Hund's rule coupling ($J$). There exists a corresponding energy relation between the in-gap states of the narrow band and the peaks in the excitation spectrum of the doublon for the wide band. We also find that the spin flip and pair-hopping Hund interactions can divide one quasiparticle peak into two peaks, where the splitting energy increases linearly with increasing $J$. Strong Hund's rule coupling can move the interband doublon-holon pair excitations outside the Mott gap and restrict the bound states by suppressing the orbital selectivity of the doubly-occupied and single-hole states.

cond-mat.str-el↗

arXiv2019

HRGE-Net: Hierarchical Relational Graph Embedding Network for Multi-view 3D Shape Recognition

View-based approach that recognizes 3D shape through its projected 2D images achieved state-of-the-art performance for 3D shape recognition. One essential challenge for view-based approach is how to aggregate the multi-view features extracted from 2D images to be a global 3D shape descriptor. In this work, we propose a novel feature aggregation network by fully investigating the relations among views. We construct a relational graph with multi-view images as nodes, and design relational graph embedding by modeling pairwise and neighboring relations among views. By gradually coarsening the graph, we build a hierarchical relational graph embedding network (HRGE-Net) to aggregate the multi-view features to be a global shape descriptor. Extensive experiments show that HRGE-Net achieves stateof-the-art performance for 3D shape classification and retrieval on benchmark datasets.

cs.CV↗

arXiv2019

MetaPruning: Meta Learning for Automatic Neural Network Channel Pruning

In this paper, we propose a novel meta learning approach for automatic channel pruning of very deep neural networks. We first train a PruningNet, a kind of meta network, which is able to generate weight parameters for any pruned structure given the target network. We use a simple stochastic structure sampling method for training the PruningNet. Then, we apply an evolutionary procedure to search for good-performing pruned networks. The search is highly efficient because the weights are directly generated by the trained PruningNet and we do not need any finetuning at search time. With a single PruningNet trained for the target network, we can search for various Pruned Networks under different constraints with little human participation. Compared to the state-of-the-art pruning methods, we have demonstrated superior performances on MobileNet V1/V2 and ResNet. Codes are available on https://github.com/liuzechun/MetaPruning.

cs.CV↗

arXiv2019

Superconducting single-layer T-graphene and novel synthesis routes

Single-layer superconductors are ideal materials for fabricating superconducting nano devices. However, up to date, very few single-layer elemental superconductors have been predicted and especially no one has been successfully synthesized yet. Here, using crystal structure search techniques and ab initio calculations, we predict that a single-layer planar carbon sheet with 4- and 8-membered rings called T-graphene is a new intrinsic elemental superconductor with superconducting critical temperature (Tc) up to around 20.8 K. More importantly, we propose a synthesis route to obtain such a single-layer T-graphene, that is, a T-graphene potassium intercalation compound (C4K with P4/mmm symmetry) is firstly synthesized at high pressure (>11.5GPa) and then quenched to ambient condition; and finally, the single-layer T-graphene can be either exfoliated using the electrochemical method from the bulk C4K, or peeled off from bulk T-graphite C4, where C4 can be obtained from C4K by evaporating the K atoms. Interestingly, we find that the calculated Tc of C4K is about 30.4K at 0GPa, which sets a new record for layered carbon-based superconductors. The present findings add a new class of carbon based superconductors. In particular, once the single-layer T-graphene is synthesized, it can pave the way for fabricating superconducting devices together with other 2D materials using the layer-by-layer growth techniques.

cond-mat.mtrl-sci↗

arXiv2019

Rethinking on Multi-Stage Networks for Human Pose Estimation

Existing pose estimation approaches fall into two categories: single-stage and multi-stage methods. While multi-stage methods are seemingly more suited for the task, their performance in current practice is not as good as single-stage methods. This work studies this issue. We argue that the current multi-stage methods' unsatisfactory performance comes from the insufficiency in various design choices. We propose several improvements, including the single-stage module design, cross stage feature aggregation, and coarse-to-fine supervision. The resulting method establishes the new state-of-the-art on both MS COCO and MPII Human Pose dataset, justifying the effectiveness of a multi-stage architecture. The source code is publicly available for further research.

cs.CV↗

arXiv2019

Distributed Solver for Discrete-Time Lyapunov Equations Over Dynamic Networks with Linear Convergence Rate

This paper investigates the problem of solving discrete-time Lyapunov equations (DTLE) over a multi-agent system, where every agent has access to its local information and communicates with its neighbors. To obtain a solution to DTLE, a distributed algorithm with uncoordinated constant step sizes is proposed over time-varying topologies. The convergence properties and the range of constant step sizes of the proposed algorithm are analyzed. Moreover, a linear convergence rate is proved and the convergence performances over dynamic networks are verified by numerical simulations.

math.OC↗

arXiv2019

Meta-SR: A Magnification-Arbitrary Network for Super-Resolution

Recent research on super-resolution has achieved great success due to the development of deep convolutional neural networks (DCNNs). However, super-resolution of arbitrary scale factor has been ignored for a long time. Most previous researchers regard super-resolution of different scale factors as independent tasks. They train a specific model for each scale factor which is inefficient in computing, and prior work only take the super-resolution of several integer scale factors into consideration. In this work, we propose a novel method called Meta-SR to firstly solve super-resolution of arbitrary scale factor (including non-integer scale factors) with a single model. In our Meta-SR, the Meta-Upscale Module is proposed to replace the traditional upscale module. For arbitrary scale factor, the Meta-Upscale Module dynamically predicts the weights of the upscale filters by taking the scale factor as input and use these weights to generate the HR image of arbitrary size. For any low-resolution image, our Meta-SR can continuously zoom in it with arbitrary scale factor by only using a single model. We evaluated the proposed method through extensive experiments on widely used benchmark datasets on single image super-resolution. The experimental results show the superiority of our Meta-Upscale.

cs.CV↗

arXiv2019

DFANet: Deep Feature Aggregation for Real-Time Semantic Segmentation

This paper introduces an extremely efficient CNN architecture named DFANet for semantic segmentation under resource constraints. Our proposed network starts from a single lightweight backbone and aggregates discriminative features through sub-network and sub-stage cascade respectively. Based on the multi-scale feature propagation, DFANet substantially reduces the number of parameters, but still obtains sufficient receptive field and enhances the model learning ability, which strikes a balance between the speed and segmentation performance. Experiments on Cityscapes and CamVid datasets demonstrate the superior performance of DFANet with 8$\times$ less FLOPs and 2$\times$ faster than the existing state-of-the-art real-time semantic segmentation methods while providing comparable accuracy. Specifically, it achieves 70.3\% Mean IOU on the Cityscapes test dataset with only 1.7 GFLOPs and a speed of 160 FPS on one NVIDIA Titan X card, and 71.3\% Mean IOU with 3.4 GFLOPs while inferring on a higher resolution image.

cs.CV↗

arXiv2019

Perceive Where to Focus: Learning Visibility-aware Part-level Features for Partial Person Re-identification

This paper considers a realistic problem in person re-identification (re-ID) task, i.e., partial re-ID. Under partial re-ID scenario, the images may contain a partial observation of a pedestrian. If we directly compare a partial pedestrian image with a holistic one, the extreme spatial misalignment significantly compromises the discriminative ability of the learned representation. We propose a Visibility-aware Part Model (VPM), which learns to perceive the visibility of regions through self-supervision. The visibility awareness allows VPM to extract region-level features and compare two images with focus on their shared regions (which are visible on both images). VPM gains two-fold benefit toward higher accuracy for partial re-ID. On the one hand, compared with learning a global feature, VPM learns region-level features and benefits from fine-grained information. On the other hand, with visibility awareness, VPM is capable to estimate the shared regions between two images and thus suppresses the spatial misalignment. Experimental results confirm that our method significantly improves the learned representation and the achieved accuracy is on par with the state of the art.

cs.CV↗

arXiv2019

Improving Cross-Domain Chinese Word Segmentation with Word Embeddings

Cross-domain Chinese Word Segmentation (CWS) remains a challenge despite recent progress in neural-based CWS. The limited amount of annotated data in the target domain has been the key obstacle to a satisfactory performance. In this paper, we propose a semi-supervised word-based approach to improving cross-domain CWS given a baseline segmenter. Particularly, our model only deploys word embeddings trained on raw text in the target domain, discarding complex hand-crafted features and domain-specific dictionaries. Innovative subsampling and negative sampling methods are proposed to derive word embeddings optimized for CWS. We conduct experiments on five datasets in special domains, covering domains in novels, medicine, and patent. Results show that our model can obviously improve cross-domain CWS, especially in the segmentation of domain-specific noun entities. The word F-measure increases by over 3.0% on four datasets, outperforming state-of-the-art semi-supervised and unsupervised cross-domain CWS approaches with a large margin. We make our code and data available on Github.

cs.CL↗

arXiv2019

Quantum correlations near the exceptional point

Recent advances in non-Hermitian physical systems have led to numerous novel optical phenomena and applications. However, most realizations are limited to classical systems and quantum fluctuations of light is unexplored. For the first time, we report the observation of quantum correlations between light channels in an anti-symmetric optical system made of flying atoms. Two distant optical channels coupled dissipatively, display gain, phase sensitivity and quantum correlations with each other, even under linear atom-light interaction within each channel. We found that quantum correlations emerge in the phase unbroken regime and disappears after crossing the exceptional point. Our microscopic model considering quantum noise evolution produces results in good qualitative agreement with experimental observations. This work opens up a new direction of experimental quantum nonlinear optics using non-Hermitian systems, and demonstrates the viability of nonlinear coupling with linear systems by using atomic motion as feedback.

quant-ph↗

arXiv2019

Growth and Thermo-driven Crystalline Phase Transition of Metastable Monolayer 1T'-WSe2 Thin Film

Two-dimensional (2D) transition metal dichalcogenides MX2 (M = Mo, W, X = S, Se, Te) attracts enormous research interests in recent years. Its 2H phase possesses an indirect to direct bandgap transition in 2D limit, and thus shows great application potentials in optoelectronic devices [1]. The 1T' crystalline phase transition can drive the monolayer MX2 to be a 2D topological insulator. Here we realized the molecular beam epitaxial (MBE) growth of both the 1T' and 2H phase monolayer WSe2 on bilayer graphene (BLG) substrate. The crystalline structures of these two phases were characterized using scanning tunneling microscopy. The monolayer 1T'-WSe2 was found to be metastable, and can transform into 2H phase under post-annealing procedure. The phase transition temperature of 1T'-WSe2 grown on BLG is lower than that of 1T' phase grown on 2H-WSe2 layers. This thermo-driven crystalline phase transition makes the monolayer WSe2 to be an ideal platform for the controlling of topological phase transitions in 2D materials family.

cond-mat.mtrl-sci↗

arXiv2019

Piezo-driven sample rotation system with ultra-low electron temperature

Piezo-driven rotator is convenient for tilted magnetic field experiments due to its precise angle control. However, the rotator itself and the sample mounted on it are difficult to be cooled down because of extra heat leaks and presumably bad thermal contacts from the piezo. Here, we report a piezo-driven sample rotation system designed for ultra-low temperature environment. The sample, as well as the rotating sample holder, can be cooled to as low as 25 mK by customized thermal links and thermal contacts. More importantly, the electron temperature in the electrical transport measurements can also be cooled down to 25 mK with the help of home-made filters. To demonstrate the application of our rotation system at ultra-low electron temperature, a measurement revealing tilt-induced localization and delocalization in the second Landau level of two-dimensional electron gas is provided.

cond-mat.mes-hall↗

arXiv2019

Quantum coherence across bosonic superconductor-anomalous metal-insulator transitions

After decades of explorations, suffering from low critical temperature and subtle nature, whether a metallic ground state exists in a two-dimensional system beyond Anderson localization is still a mystery. Supremely, phase coherence could be the key that unlocks its intriguing nature. This work reveals how quantum phase coherence evolves across bosonic superconductor-metal-insulator transitions via magneto-conductance quantum oscillations in high-Tc superconducting films. A robust intervening anomalous metallic state characterized by both resistance and oscillation amplitude saturations in the low temperature regime is detected. By contrast, with decreasing temperature the oscillation amplitude monotonically grows on the superconducting side, but decreases at low temperatures on the insulating side. It suggests that the saturation of phase coherence plays a prominent role in the formation of this anomalous metallic state.

cond-mat.supr-con↗