arXiv subjects

Ming Lu

Publications and source records attributed to Ming Lu.

At least 217 records · Page 12Linked to original sources

arXiv2019

Learned Quality Enhancement via Multi-Frame Priors for HEVC Compliant Low-Delay Applications

Networked video applications, e.g., video conferencing, often suffer from poor visual quality due to unexpected network fluctuation and limited bandwidth. In this paper, we have developed a Quality Enhancement Network (QENet) to reduce the video compression artifacts, leveraging the spatial and temporal priors generated by respective multi-scale convolutions spatially and warped temporal predictions in a recurrent fashion temporally. We have integrated this QENet as a standard-alone post-processing subsystem to the High Efficiency Video Coding (HEVC) compliant decoder. Experimental results show that our QENet demonstrates the state-of-the-art performance against default in-loop filters in HEVC and other deep learning based methods with noticeable objective gains in Peak-Signal-to-Noise Ratio (PSNR) and subjective gains visually.

eess.IV↗

arXiv2019

Singularity categories of representations of algebras over local rings

Let $Λ$ be a finite-dimensional algebra with finite global dimension, $R_k=K[X]/(X^k)$ be the $\mathcal{Z}$-graded local ring with $k\geq1$, and $Λ_k=Λ\otimes_K R_k$. We consider the singularity category $\mathcal{D}_{sg}(\mathrm{mod}^\mathcal{Z}(Λ_k))$ of the graded modules over $Λ_k$. It is showed that there is a tilting object in $\mathcal{D}_{sg}(\mathrm{mod}^\mathcal{Z}(Λ_k))$ such that its endomorphism algebra is isomorphic to the triangular matrix algebra $T_{k-1}(Λ)$ with coefficients in $Λ$ and there is a triangulated equivalence between $\mathcal{D}_{sg}(\mathrm{mod}^{\mathcal{Z}/k\mathcal{Z}}(Λ))$ and the root category of $T_{k-1}(Λ)$. Finally, a classification of $Λ_k$ up to the Cohen-Macaulay representation type is given.

math.RT↗

arXiv2019

Neural Video Compression using Spatio-Temporal Priors

The pursuit of higher compression efficiency continuously drives the advances of video coding technologies. Fundamentally, we wish to find better "predictions" or "priors" that are reconstructed previously to remove the signal dependency efficiently and to accurately model the signal distribution for entropy coding. In this work, we propose a neural video compression framework, leveraging the spatial and temporal priors, independently and jointly to exploit the correlations in intra texture, optical flow based temporal motion and residuals. Spatial priors are generated using downscaled low-resolution features, while temporal priors (from previous reference frames and residuals) are captured using a convolutional neural network based long-short term memory (ConvLSTM) structure in a temporal recurrent fashion. All of these parts are connected and trained jointly towards the optimal rate-distortion performance. Compared with the High-Efficiency Video Coding (HEVC) Main Profile (MP), our method has demonstrated averaged 38% Bjontegaard-Delta Rate (BD-Rate) improvement using standard common test sequences, where the distortion is multi-scale structural similarity (MS-SSIM).

eess.IV↗

arXiv2018

Electronic scattering, focusing, and resonance by a spherical barrier in Weyl semimetals

We solve the Weyl electron scattered by a spherical step potential barrier. Tuning the incident energy and the potential radius, one can enter both quasiclassical and quantum regimes. Transport features related to far-field currents and integrated cross sections are studied to reveal the preferred forward scattering. In the quasiclassical regime, a strong focusing effect along the incident spherical axis is found in addition to optical caustic patterns. In the quantum regime, at energies of successive angular momentum resonances, a polar aggregation of electron density is found inside the potential. The findings will be useful in transport studies and electronic lens applications in Weyl systems.

cond-mat.mes-hall↗

arXiv2017

Gorenstein homological properties of tensor rings

Let $R$ be a two-sided noetherian ring and $M$ be a nilpotent $R$-bimodule, which is finitely generated on both sides. We study Gorenstein homological properties of the tensor ring $T_R(M)$. Under certain conditions, the ring $R$ is Gorenstein if and only if so is $T_R(M)$. We characterize Gorenstein projective $T_R(M)$-modules in terms of $R$-modules.

math.RA↗

arXiv2017

An silicon nanocrystal laser

Silicon lasers have been the most challenging element for the monolithic integrated Si photonics. Here we report the first successful all-Si laser at room temperature based on silicon nanocrystals (Si NCs) with high optical gains. The active Si NC layer was made from hydrogen silsesquioxane (HSQ) that had undergone a phase separation annealing, followed by high-pressure hydrogen passivation. Using the nanoimprint technique, a second-order Bragg grating was made on the active layer as the distributed feedback (DFB) laser cavity. This Si NC DFB laser device was optically pumped with femtosecond pulses. Criteria for lasing action, i.e., the threshold effect, the polarization dependence and the significant spectral narrowing of stimulated emission were evidently fulfilled, which indicated the realization of an all-Si laser.

physics.optics↗

arXiv2017

RON: Reverse Connection with Objectness Prior Networks for Object Detection

We present RON, an efficient and effective framework for generic object detection. Our motivation is to smartly associate the best of the region-based (e.g., Faster R-CNN) and region-free (e.g., SSD) methodologies. Under fully convolutional architecture, RON mainly focuses on two fundamental problems: (a) multi-scale object localization and (b) negative sample mining. To address (a), we design the reverse connection, which enables the network to detect objects on multi-levels of CNNs. To deal with (b), we propose the objectness prior to significantly reduce the searching space of objects. We optimize the reverse connection, objectness prior and object detector jointly by a multi-task loss function, thus RON can directly predict final detection results from all locations of various feature maps. Extensive experiments on the challenging PASCAL VOC 2007, PASCAL VOC 2012 and MS COCO benchmarks demonstrate the competitive performance of RON. Specifically, with VGG-16 and low resolution 384X384 input size, the network gets 81.3% mAP on PASCAL VOC 2007, 80.7% mAP on PASCAL VOC 2012 datasets. Its superiority increases when datasets become larger and more difficult, as demonstrated by the results on the MS COCO dataset. With 1.5G GPU memory at test phase, the speed of the network is 15 FPS, 3X faster than the Faster R-CNN counterpart.

cs.CV↗

arXiv2017

Physics Inspired Optimization on Semantic Transfer Features: An Alternative Method for Room Layout Estimation

In this paper, we propose an alternative method to estimate room layouts of cluttered indoor scenes. This method enjoys the benefits of two novel techniques. The first one is semantic transfer (ST), which is: (1) a formulation to integrate the relationship between scene clutter and room layout into convolutional neural networks; (2) an architecture that can be end-to-end trained; (3) a practical strategy to initialize weights for very deep networks under unbalanced training data distribution. ST allows us to extract highly robust features under various circumstances, and in order to address the computation redundance hidden in these features we develop a principled and efficient inference scheme named physics inspired optimization (PIO). PIO's basic idea is to formulate some phenomena observed in ST features into mechanics concepts. Evaluations on public datasets LSUN and Hedau show that the proposed method is more accurate than state-of-the-art methods.

cs.CV↗

arXiv2017

Cohen-Macaulay Auslander algebras of gentle algebras

For any gentle algebra $Λ=KQ/\langle I\rangle$, following Kalck, we describe the quiver and the relations for its Cohen-Macaulay Auslander algebra $\mathrm{Aus}(\mathrm{Gproj}Λ)$ explicitly, and obtain some properties, such as $Λ$ is representation-finite if and only if $\mathrm{Aus}(\mathrm{Gproj}Λ)$ is; if $Q$ has no loop and any indecomposable $Λ$-module is uniquely determined by its dimension vector, then any indecomposable $\mathrm{Aus}(\mathrm{Gproj}Λ)$-module is uniquely determined by its dimension vector.

math.RT↗

arXiv2017

Gorenstein defect categories of triangular matrix algebras

We apply the technique of recollement to study the Gorenstein defect categories of triangular matrix algebras. First, we construct a left recollement of Gorenstein defect categories for a triangular matrix algebra under some conditions, using it, we give a categorical interpretation of the Gorenstein properties of the triangular matrix algebra obtained by X-W. Chen, B. L. Xiong and P. Zhang respectively. Second, under some additional conditions, a recollement of Gorenstein defect categories for a triangular matrix algebra is constructed. As an application, for a special kind of triangular matrix algebras, which are called simple gluing algebras, we describe their singularity categories and Gorenstein defect categories.

math.RT↗

arXiv2017

Gorenstein properties of simple gluing algebras

Let $A=KQ_A/I_A$ and $B=KQ_B/I_B$ be two finite-dimensional bound quiver algebras, fix two vertices $a\in Q_A$ and $b\in Q_B$. We define an algebra $Λ=KQ_Λ/I_Λ$, which is called a simple gluing algebra of $A$ and $B$, where $Q_Λ$ is from $Q_A$ and $Q_B$ by identifying $a$ and $b$, $I_Λ=\langle I_A,I_B\rangle$. We prove that $Λ$ is Gorenstein if and only if $A$ and $B$ are Gorenstein, and describe the Gorenstein projective modules, singularity category, Gorenstein defect category and also Cohen-Macaulay Auslander algebra of $Λ$ from the corresponding ones of $A$ and $B$.

math.RT↗

arXiv2016

Singularity categories of some 2-CY-tilted algebras

We define a class of finite-dimensional Jacobian algebras, which are called (simple) polygon-tree algebras, as a generalization of cluster-tilted algebras of type $\D$. They are $2$-CY-tilted algebras. Using a suitable process of mutations of quivers with potentials (which are also BB-mutations) inducing derived equivalences, and one-pointed (co)extensions which preserve singularity equivalences, we find a connected selfinjective Nakayama algebra whose stable category is equivalent to the singularity category of a simple polygon-tree algebra. Furthermore, we also give a classification of algebras of this kind up to representation type.

math.RT↗

arXiv2016

Higgs amplitude mode in massless Dirac fermion systems

The Higgs amplitude mode in superconductors is the condensed matter analogy of Higgs bosons in particle physics. We investigate the time evolution of Higgs amplitude mode in massless Dirac systems, induced by a weak quench of an attractive interaction. We find that the Higgs amplitude mode in the half-filling honeycomb lattice has a logarithmic decaying behaviour, qualitatively different from the $1/\sqrt{t}$ decay in the normal superconductors. Our study is also extended to the doped cases in honeycomb lattice. As for the 3D Dirac semimetal at half filling, we obtain an undamped oscillation of the amplitude mode. Our finding is not only an important supplement to the previous theoretical studies on normal fermion systems, but also provide an experimental signature to characterize the superconductivity in 2D or 3D Dirac systems.

cond-mat.supr-con↗

arXiv2015

Desingularization of quiver Grassmannians for Gentle algebras

Following [20], a desingularization of arbitrary quiver Grassmannians for finite dimensional Gorenstein projective modules of 1-Gorenstein gentle algebras is constructed in terms of quiver Grassmannians for their Cohen-Macaulay Auslander algebras.

math.RT↗

arXiv2015

Singularity categories of skewed-gentle algebras

Let $K$ be an algebraically closed field. Let $(Q,Sp,I)$ be a skewed-gentle triple, $(Q^{sg},I^{sg})$ and $(Q^g,I^{g})$ be its corresponding skewed-gentle pair and associated gentle pair respectively. It proves that the skewed-gentle algebra $KQ^{sg}/< I^{sg}>$ is singularity equivalent to $KQ/< I>$. Moreover, we use $(Q,Sp,I)$ to describe the singularity category of $KQ^g/< I^g>$. As a corollary, we get that $\mathrm{gldim} KQ^{sg}/< I^{sg}><\infty$ if and only if $\mathrm{gldim} KQ/< I><\infty$ if and only if $\mathrm{gldim} KQ^{g}/< I^{g}><\infty$.

math.RT↗

arXiv2014

Generation of Ensembles of Individually Resolvable Nitrogen Vacancies Using Nanometer-Scale Apertures in Ultrahigh-Aspect Ratio Planar Implantation Masks

A central challenge in developing magnetically coupled quantum registers in diamond is the fabrication of nitrogen vacancy (NV) centers with localization below ~20 nm to enable fast dipolar interaction compared to the NV decoherence rate. Here, we demonstrate the targeted, high throughput formation of NV centers using masks with a thickness of 270 nm and feature sizes down to ~1 nm. Super-resolution imaging resolves NVs with a full-width maximum distribution of $26\pm7$ nm and a distribution of NV-NV separations of $16\pm5$ nm.

cond-mat.mtrl-sci↗

arXiv2014

Photon transport enhanced by transverse Anderson localization in disordered superlattices

One of the daunting challenges in optical physics is to accurately control the flow of light at the subwavelength scale, by patterning the optical medium one can design anisotropic media. The light transport can also be significantly affected by Anderson localization, namely the wave localization in a disordered medium, a ubiquitous phenomenon in wave physics. Here we report the photon transport and collimation enhanced by transverse Anderson localization in chip-scale dispersion engineered anisotropic media. We demonstrate a new type of anisotropic photonic structure in which diffraction is nearly completely arrested by cascaded resonant tunneling through transverse guided resonances. By perturbing the shape of more than 4,000 scatterers in these superlattices we add structural disordered in a controlled manner and uncover the mechanism of disorder-induced transverse localization at the chip-scale. Arrested spatial divergence is captured in the power-law scaling, along with exponential asymmetric mode profiles and enhanced collimation bandwidth for increasing disorder. With increasing disorder, we observe the crossover from cascaded guided resonances into the transverse localization regime, beyond the ballistic and diffusive transport of photons.

physics.optics↗

arXiv2014

The singularity categories of the Cluster-tilted algebras of Dynkin type

In this paper, we use the stable categories of some selfinjective algebras to describe the singularity categories of the cluster-tilted algebras of Dynkin type. Furthermore, in this way, we settle the problem of singularity equivalence classification of the cluster-tilted algebra of type $A$, $D$ and $E$ respectively.

math.RT↗