arXiv subjects

Yan Wang

Publications and source records attributed to Yan Wang.

At least 991 records · Page 55Linked to original sources

arXiv2017

Deep Regression Forests for Age Estimation

Age estimation from facial images is typically cast as a nonlinear regression problem. The main challenge of this problem is the facial feature space w.r.t. ages is heterogeneous, due to the large variation in facial appearance across different persons of the same age and the non-stationary property of aging patterns. In this paper, we propose Deep Regression Forests (DRFs), an end-to-end model, for age estimation. DRFs connect the split nodes to a fully connected layer of a convolutional neural network (CNN) and deal with heterogeneous data by jointly learning input-dependant data partitions at the split nodes and data abstractions at the leaf nodes. This joint learning follows an alternating strategy: First, by fixing the leaf nodes, the split nodes as well as the CNN parameters are optimized by Back-propagation; Then, by fixing the split nodes, the leaf nodes are optimized by iterating a step-size free and fast-converging update rule derived from Variational Bounding. We verify the proposed DRFs on three standard age estimation benchmarks and achieve state-of-the-art results on all of them.

cs.CV↗

arXiv2017

Minimum co-degree condition for perfect matchings in k-partite k-graphs

Let $H$ be a $k$-partite $k$-graph with $n$ vertices in each partition class, and let $δ_{k-1}(H)$ denote the minimum co-degree of $H$. We characterize those $H$ with $δ_{k-1}(H) \geq n/2$ and with no perfect matching. As a consequence we give an affirmative answer to the following question of Rödl and Ruciński: If $k$ is even or $n \not\equiv 2 \pmod 4$, does $δ_{k-1}(H) \geq n/2$ imply that $H$ has a perfect matching? We also give an example indicating that it is not sufficient to impose this degree bound on only two types of $(k-1)$-sets.

math.CO↗

arXiv2017

A Quantum Walk Enhanced Grover Search Algorithm for Global Optimization

One of the significant breakthroughs in quantum computation is Grover's algorithm for unsorted database search. Recently, the applications of Grover's algorithm to solve global optimization problems have been demonstrated, where unknown optimum solutions are found by iteratively improving the threshold value for the selective phase shift operator in Grover rotation. In this paper, a hybrid approach that combines continuous-time quantum walks with Grover search is proposed so that the search is accelerated with improved threshold values. By taking advantage of the quantum tunneling effect, better threshold values can be found at the early stage of the search process so that the sharpness of probability improves. The results between the new algorithm, existing Grover search, and classical heuristic algorithms are compared.

quant-ph↗

arXiv2017

Prospects for gravitational wave astronomy with next generation large-scale pulsar timing arrays

Next generation radio telescopes, namely the Five-hundred-meter Aperture Spherical Telescope (FAST) and the Square Kilometer Array (SKA), will revolutionize the pulsar timing arrays (PTAs) based gravitational wave (GW) searches. We review some of the characteristics of FAST and SKA, and the resulting PTAs, that are pertinent to the detection of gravitational wave signals from individual supermassive black hole binaries.

astro-ph.IM↗

arXiv2017

SORT: Second-Order Response Transform for Visual Recognition

In this paper, we reveal the importance and benefits of introducing second-order operations into deep neural networks. We propose a novel approach named Second-Order Response Transform (SORT), which appends element-wise product transform to the linear sum of a two-branch network module. A direct advantage of SORT is to facilitate cross-branch response propagation, so that each branch can update its weights based on the current status of the other branch. Moreover, SORT augments the family of transform operations and increases the nonlinearity of the network, making it possible to learn flexible functions to fit the complicated distribution of feature space. SORT can be applied to a wide range of network architectures, including a branched variant of a chain-styled network and a residual network, with very light-weighted modifications. We observe consistent accuracy gain on both small (CIFAR10, CIFAR100 and SVHN) and big (ILSVRC2012) datasets. In addition, SORT is very efficient, as the extra computation overhead is less than 5%.

cs.CV↗

arXiv2017

Multi-stage Multi-recursive-input Fully Convolutional Networks for Neuronal Boundary Detection

In the field of connectomics, neuroscientists seek to identify cortical connectivity comprehensively. Neuronal boundary detection from the Electron Microscopy (EM) images is often done to assist the automatic reconstruction of neuronal circuit. But the segmentation of EM images is a challenging problem, as it requires the detector to be able to detect both filament-like thin and blob-like thick membrane, while suppressing the ambiguous intracellular structure. In this paper, we propose multi-stage multi-recursive-input fully convolutional networks to address this problem. The multiple recursive inputs for one stage, i.e., the multiple side outputs with different receptive field sizes learned from the lower stage, provide multi-scale contextual boundary information for the consecutive learning. This design is biologically-plausible, as it likes a human visual system to compare different possible segmentation solutions to address the ambiguous boundary issue. Our multi-stage networks are trained end-to-end. It achieves promising results on two public available EM segmentation datasets, the mouse piriform cortex dataset and the ISBI 2012 EM dataset.

cs.CV↗

arXiv2017

Deep View-Sensitive Pedestrian Attribute Inference in an end-to-end Model

Pedestrian attribute inference is a demanding problem in visual surveillance that can facilitate person retrieval, search and indexing. To exploit semantic relations between attributes, recent research treats it as a multi-label image classification task. The visual cues hinting at attributes can be strongly localized and inference of person attributes such as hair, backpack, shorts, etc., are highly dependent on the acquired view of the pedestrian. In this paper we assert this dependence in an end-to-end learning framework and show that a view-sensitive attribute inference is able to learn better attribute predictions. Our proposed model jointly predicts the coarse pose (view) of the pedestrian and learns specialized view-specific multi-label attribute predictions. We show in an extensive evaluation on three challenging datasets (PETA, RAP and WIDER) that our proposed end-to-end view-aware attribute prediction model provides competitive performance and improves on the published state-of-the-art on these datasets.

cs.CV↗

arXiv2017

DeepSkeleton: Learning Multi-task Scale-associated Deep Side Outputs for Object Skeleton Extraction in Natural Images

Object skeletons are useful for object representation and object detection. They are complementary to the object contour, and provide extra information, such as how object scale (thickness) varies among object parts. But object skeleton extraction from natural images is very challenging, because it requires the extractor to be able to capture both local and non-local image context in order to determine the scale of each skeleton pixel. In this paper, we present a novel fully convolutional network with multiple scale-associated side outputs to address this problem. By observing the relationship between the receptive field sizes of the different layers in the network and the skeleton scales they can capture, we introduce two scale-associated side outputs to each stage of the network. The network is trained by multi-task learning, where one task is skeleton localization to classify whether a pixel is a skeleton pixel or not, and the other is skeleton scale prediction to regress the scale of each skeleton pixel. Supervision is imposed at different stages by guiding the scale-associated side outputs toward the groundtruth skeletons at the appropriate scales. The responses of the multiple scale-associated side outputs are then fused in a scale-specific way to detect skeleton pixels using multiple scales effectively. Our method achieves promising results on two skeleton extraction datasets, and significantly outperforms other competitors. Additionally, the usefulness of the obtained skeletons and scales (thickness) are verified on two object detection applications: Foreground object segmentation and object proposal detection.

cs.CV↗

arXiv2017

Stable Equilibria of Anisotropic Particles on Substrates: a Generalized Winterbottom Construction

We present a new approach for predicting stable equilibrium shapes of crystalline islands on flat substrates, as commonly occur through solid-state dewetting of thin films. The new theory is a generalization of the widely used Winterbottom construction i.e., an extension of the Wulff construction for particles on substrates). This approach is equally applicable to cases where the crystal surface energy is isotropic, weakly anisotropic, strongly anisotropic and "cusped". We demonstrate that, unlike in the classical Winterbottom approach, multiple equilibrium island shapes may be possible when the surface energy is strongly anisotropic. We analyze these shapes through perturbation analysis, by calculating the first and second variations of the total free energy functional with respect to contact locations and island shape. Based on this analysis, we find the necessary conditions for stable equilibria and exploit this through a generalization of the Winterbottom construction to identify all possible stable equilibrium shapes. Finally, we propose a dynamical evolution method based on surface diffusion mass transport to determine whether all of the stable equilibrium shapes are dynamically accessible. Applying this approach, we demonstrate that islands with different initial shapes may evolve into different stationary shapes and show that these dynamically-determined stationary states correspond to the predicted stable equilibrium shapes, as obtained from the generalized Winterbottom construction.

cond-mat.mtrl-sci↗

arXiv2017

A Fixed-Point Model for Pancreas Segmentation in Abdominal CT Scans

Deep neural networks have been widely adopted for automatic organ segmentation from abdominal CT scans. However, the segmentation accuracy of some small organs (e.g., the pancreas) is sometimes below satisfaction, arguably because deep networks are easily disrupted by the complex and variable background regions which occupies a large fraction of the input volume. In this paper, we formulate this problem into a fixed-point model which uses a predicted segmentation mask to shrink the input region. This is motivated by the fact that a smaller input region often leads to more accurate segmentation. In the training process, we use the ground-truth annotation to generate accurate input regions and optimize network weights. On the testing stage, we fix the network parameters and update the segmentation results in an iterative manner. We evaluate our approach on the NIH pancreas segmentation dataset, and outperform the state-of-the-art by more than 4%, measured by the average Dice-Sørensen Coefficient (DSC). In addition, we report 62.43% DSC in the worst case, which guarantees the reliability of our approach in clinical applications.

cs.CV↗

arXiv2017

Almost perfect matchings in $k$-partite $k$-graphs

The minimum co-degree threshold for a perfect matching in a $k$-graph with $n$ vertices was determined by Rödl, Ruciński and Szemerédi for the case when $n\equiv 0\pmod k$. Recently, Han resolved the remaining cases when $n \not\equiv 0\pmod k$, establishing a conjecture of Rödl, Ruciński and Szemerédi. In this paper, we determine the minimum co-degree threshold for almost perfect matchings in $k$-partite $k$-graphs, answering a question of Rödl and Ruciński.

math.CO↗

arXiv2017

Continuous gravitational wave searches with pulsar timing arrays: Maximization versus marginalization over pulsar phase parameters

Resolvable Supermassive Black Hole Binaries are promising sources for Pulsar Timing Array based gravitational wave searches. Search algorithms for such targets must contend with the large number of so-called pulsar phase parameters in the joint log-likelihood function of the data. We compare the localization accuracy for two approaches: Maximization over the pulsar phase parameters (MaxPhase) against marginalization over them (AvPhase). Using simulated data from a pulsar timing array with 17 pulsars, we find that for weak and moderately strong signals, AvPhase outperforms MaxPhase significantly, while they perform comparably for strong signals.

astro-ph.IM↗

arXiv2017

Pulsar Timing Array Based Search for Supermassive Black Hole Binaries in the Square Kilometer Array Era

The advent of next generation radio telescope facilities, such as the Square Kilometer Array (SKA), will usher in an era where a Pulsar Timing Array (PTA) based search for gravitational waves (GWs) will be able to use hundreds of well timed millisecond pulsars rather than the few dozens in existing PTAs. A realistic assessment of the performance of such an extremely large PTA must take into account the data analysis challenge posed by an exponential increase in the parameter space volume due to the large number of so-called pulsar phase parameters. We address this problem and present such an assessment for isolated supermassive black hole binary (SMBHB) searches using a SKA era PTA containing $10^3$ pulsars. We find that an all-sky search will be able to confidently detect non-evolving sources with redshifted chirp mass of $10^{10}$ $M_\odot$ out to a redshift of about $28$ (corresponding to a rest-frame chirp mass of $3.4\times 10^{8}$ $M_\odot$). We discuss the important implications that the large distance reach of a SKA era PTA has on GW observations from optically identified SMBHB candidates. If no SMBHB detections occur, a highly unlikely scenario in the light of our results, the sky-averaged upper limit on strain amplitude will be improved by about three orders of magnitude over existing limits.

astro-ph.IM↗

arXiv2017

Phonon linewidth due to electron-phonon interactions with strong forward scattering in FeSe thin films on oxide substrates

The discovery of an enhanced superconducting transition temperature $T_c$ in monolayers of FeSe grown on several oxide substrates has opened a new route to high-$T_c$ superconductivity through interface engineering. One proposal for the origin of the observed enhancement is an electron-phonon (e-ph) interaction across the interface that peaked at small momentum transfers. In this paper, we examine the implications of such a coupling on the phononic properties of the system. We show that a strong forward scattering leads to a sizable broadening of phonon lineshape, which may result in charge instabilities at long-wavelengths. However, we further find that the inclusion of Coulombic screening significantly reduces the phonon broadening. Our results show that one might not expect anomalously broad phonon linewidths in the FeSe interface systems, despite the fact that the e-ph interaction has a strong peak in the forward scattering (small $q$) direction.

cond-mat.supr-con↗

arXiv2017

EmotioNet Challenge: Recognition of facial expressions of emotion in the wild

This paper details the methodology and results of the EmotioNet challenge. This challenge is the first to test the ability of computer vision algorithms in the automatic analysis of a large number of images of facial expressions of emotion in the wild. The challenge was divided into two tracks. The first track tested the ability of current computer vision algorithms in the automatic detection of action units (AUs). Specifically, we tested the detection of 11 AUs. The second track tested the algorithms' ability to recognize emotion categories in images of facial expressions. Specifically, we tested the recognition of 16 basic and compound emotion categories. The results of the challenge suggest that current computer vision and machine learning algorithms are unable to reliably solve these two tasks. The limitations of current algorithms are more apparent when trying to recognize emotion. We also show that current algorithms are not affected by mild resolution changes, small occluders, gender or age, but that 3D pose is a major limiting factor on performance. We provide an in-depth discussion of the points that need special attention moving forward.

cs.CV↗

arXiv2017

Deep Collaborative Learning for Visual Recognition

Deep neural networks are playing an important role in state-of-the-art visual recognition. To represent high-level visual concepts, modern networks are equipped with large convolutional layers, which use a large number of filters and contribute significantly to model complexity. For example, more than half of the weights of AlexNet are stored in the first fully-connected layer (4,096 filters). We formulate the function of a convolutional layer as learning a large visual vocabulary, and propose an alternative way, namely Deep Collaborative Learning (DCL), to reduce the computational complexity. We replace a convolutional layer with a two-stage DCL module, in which we first construct a couple of smaller convolutional layers individually, and then fuse them at each spatial position to consider feature co-occurrence. In mathematics, DCL can be explained as an efficient way of learning compositional visual concepts, in which the vocabulary size increases exponentially while the model complexity only increases linearly. We evaluate DCL on a wide range of visual recognition tasks, including a series of multi-digit number classification datasets, and some generic image classification datasets such as SVHN, CIFAR and ILSVRC2012. We apply DCL to several state-of-the-art network structures, improving the recognition accuracy meanwhile reducing the number of parameters (16.82% fewer in AlexNet).

cs.CV↗

arXiv2016

Visualizing dispersive features in 2D image via minimum gradient method

We developed a minimum gradient based method to track ridge features in 2D image plot, which is a typical data representation in many momentum resolved spectroscopy experiments. Through both analytic formulation and numerical simulation, we compare this new method with existing DC (distribution curve) based and higher order derivative based analyses. We find that the new method has good noise resilience and enhanced contrast especially for weak intensity features, meanwhile preserves the quantitative local maxima information from the raw image. An algorithm is proposed to extract 1D ridge dispersion from the 2D image plot, whose quantitative application to angle-resolved photoemission spectroscopy measurements on high temperature superconductors is demonstrated.

physics.data-an↗

arXiv2016

The Kelmans-Seymour conjecture IV: a proof

A well known theorem of Kuratowski in 1932 states that a graph is planar if, and only if, it does not contain a subdivision of $K_5$ or $K_{3,3}$. Wagner proved in 1937 that if a graph other than $K_5$ does not contain any subdivision of $K_{3,3}$ then it is planar or it admits a cut of size at most 2. Kelmans and, independently, Seymour conjectured in the 1970s that if a graph does not contain any subdivision of $K_5$ then it is planar or it admits a cut of size at most 4. In this paper, we give a proof of the Kelmans-Seymour conjecture. We also discuss several related results and problems.

math.CO↗