arXiv subjects

Tong Yang

Publications and source records attributed to Tong Yang.

At least 181 records · Page 10Linked to original sources

arXiv2022

Anchor DETR: Query Design for Transformer-Based Object Detection

In this paper, we propose a novel query design for the transformer-based object detection. In previous transformer-based detectors, the object queries are a set of learned embeddings. However, each learned embedding does not have an explicit physical meaning and we cannot explain where it will focus on. It is difficult to optimize as the prediction slot of each object query does not have a specific mode. In other words, each object query will not focus on a specific region. To solved these problems, in our query design, object queries are based on anchor points, which are widely used in CNN-based detectors. So each object query focuses on the objects near the anchor point. Moreover, our query design can predict multiple objects at one position to solve the difficulty: "one region, multiple objects". In addition, we design an attention variant, which can reduce the memory cost while achieving similar or better performance than the standard attention in DETR. Thanks to the query design and the attention variant, the proposed detector that we called Anchor DETR, can achieve better performance and run faster than the DETR with 10$\times$ fewer training epochs. For example, it achieves 44.2 AP with 19 FPS on the MSCOCO dataset when using the ResNet50-DC5 feature for training 50 epochs. Extensive experiments on the MSCOCO benchmark prove the effectiveness of the proposed methods. Code is available at \url{https://github.com/megvii-research/AnchorDETR}.

cs.CV↗

arXiv2021

High-throughput Discovery and Intelligent Design of 2D Functional Materials for Various Applications

Novel technologies and new materials are in high demand for future energy-efficient electronic devices to overcome the fundamental limitations of miniaturization of current silicon-based devices. Two-dimensional (2D) materials show promising applications in the next generation devices because they can be tailored on the specific property that a technology is based on, and be compatible with other technologies, such as the silicon-based (opto)electronics. Although the number of experimentally discovered 2D materials is growing, the speed is very slow and only a few dozen 2D materials have been synthesized or exfoliated since the discovery of graphene. Recently, a novel computational technique, dubbed "high-throughput computational materials design", becomes a burgeoning area of materials science, which is the combination of the quantum-mechanical theory, materials genome, and database construction with intelligent data mining. This new and powerful tool can greatly accelerate the discovery, design and application of 2D materials by creating database containing a large amount of 2D materials with calculated fundamental properties, and then intelligently mining (via high-throughput automation or machine learning) the database in the search of 2D materials with the desired properties for particular applications, such as energy conversion, electronics, spintronics, and optoelectronics.

physics.comp-ph↗

arXiv2021

Co-mining: Self-Supervised Learning for Sparsely Annotated Object Detection

Object detectors usually achieve promising results with the supervision of complete instance annotations. However, their performance is far from satisfactory with sparse instance annotations. Most existing methods for sparsely annotated object detection either re-weight the loss of hard negative samples or convert the unlabeled instances into ignored regions to reduce the interference of false negatives. We argue that these strategies are insufficient since they can at most alleviate the negative effect caused by missing annotations. In this paper, we propose a simple but effective mechanism, called Co-mining, for sparsely annotated object detection. In our Co-mining, two branches of a Siamese network predict the pseudo-label sets for each other. To enhance multi-view learning and better mine unlabeled instances, the original image and corresponding augmented image are used as the inputs of two branches of the Siamese network, respectively. Co-mining can serve as a general training mechanism applied to most of modern object detectors. Experiments are performed on MS COCO dataset with three different sparsely annotated settings using two typical frameworks: anchor-based detector RetinaNet and anchor-free detector FCOS. Experimental results show that our Co-mining with RetinaNet achieves 1.4%~2.1% improvements compared with different baselines and surpasses existing methods under the same sparsely annotated setting. Code is available at https://github.com/megvii-research/Co-mining.

cs.CV↗

arXiv2021

Selective hydrogenation improves interface properties of high-k dielectrics on 2D semiconductors

The integration of high-k dielectrics with two-dimensional (2D) semiconductors is a critical step towards high-performance nanoelectronics, which however remains challenging due to high density of interface states and the damage to the monolayer 2D semiconductors. In this study, we propose a selective hydrogenation strategy to improve the interface properties while do not affect the 2D semiconductors. Using the interface of monolayer MoS2 and silicon nitride as an example, we show substantially improved interface properties for electronic applications after the interfacial hydrogenation, as evidenced by reduced inhomogeneous charge redistribution, increased band offset, and untouched electronic properties of MoS2. Interestingly, this hydrogenation process selectively occurs only at the silicon nitride surface and is compatible with the current semiconductor fabrication process. We further show that this strategy is general and applicable to other interfaces between high-k dielectrics and 2D semiconductors such as HfO2 on the monolayer MoS2. Our results demonstrate a simple yet viable way to improve the interfacial properties for integrating many high-k dielectrics on a broad range of two-dimensional transition metal disulfide semiconductors.

cond-mat.mtrl-sci↗

arXiv2021

Analysis on Tollmien-Schlichting wave in the Prandtl-Hartmann Regime

In this paper, we study the instability induced by the Tollmien-Schlichting wave governed by the MHD system in the Prandtl-Hartmann regime. The interaction of the inviscid mode and viscous mode that leads to the instability is analyzed by the introduction of a new decomposition of the Orr-Sommerfeld operator on the velocity and magnetic fields. The critical Gevrey index for the instability is justified by constructing the growing mode in the essential frequency and it is shown to be the same as the incompressible Navier-Stokes equations in the Prandtl regime. This result justifies rigorously the physical understanding that the transverse magnetic field to the boundary in the Prandtl-Hartmann regime has no extra stabilizing effect on the Tollmien-Schlichting wave.

math.AP↗

arXiv2021

The Boltzmann equation for plane Couette flow

In the paper, we study the plane Couette flow of a rarefied gas between two parallel infinite plates at $y=\pm L$ moving relative to each other with opposite velocities $(\pm αL,0,0)$ along the $x$-direction. Assuming that the stationary state takes the specific form of $F(y,v_x-αy,v_y,v_z)$ with the $x$-component of the molecular velocity sheared linearly along the $y$-direction, such steady flow is governed by a boundary value problem on a steady nonlinear Boltzmann equation driven by an external shear force under the homogeneous non-moving diffuse reflection boundary condition. In case of the Maxwell molecule collisions, we establish the existence of spatially inhomogeneous non-equilibrium stationary solutions to the steady problem for any small enough shear rate $α>0$ via an elaborate perturbation approach using Caflisch's decomposition together with Guo's $L^\infty\cap L^2$ theory. The result indicates the polynomial tail at large velocities for the stationary distribution. Moreover, the large time asymptotic stability of the stationary solution with an exponential convergence is also obtained and as a consequence the nonnegativity of the steady profile is justified.

math.AP↗

arXiv2021

Precise Layer-Dependent Electronic Structure of MBE-Grown PtSe$_2$

Two-dimensional (2D) platinum diselenide (PtSe$_2$) has received significant attention for 2D transistor applications due to its high mobility. Here, using molecular beam epitaxy, we investigate the growth of 2D PtSe$_2$ on highly oriented pyrolytic graphite (HOPG) and unveil their electronic properties via X-ray photoelectron spectroscopy, Raman spectra, and scanning tunnelling microscopy/spectroscopy as well as density functional theory (DFT) calculations. PtSe$_2$ adopts a layer-by-layer growth mode on HOPG and shows a decreasing band gap with increasing layer number. For the layer numbers from one to four, PtSe$_2$ has band gaps of $2.0 \pm 0.1$, $1.1 \pm 0.1$, $0.6 \pm 0.1$ and $0.20 \pm 0.1$ eV, respectively, and becomes semimetal from the fifth layer. DFT calculations reproduce the layer-dependent evolution of both the band gap and band edges, suggest an indirect band-gap structure, and elucidate the underlying physics at the atomic level.

cond-mat.mtrl-sci↗

arXiv2021

Points as Queries: Weakly Semi-supervised Object Detection by Points

We propose a novel point annotated setting for the weakly semi-supervised object detection task, in which the dataset comprises small fully annotated images and large weakly annotated images by points. It achieves a balance between tremendous annotation burden and detection performance. Based on this setting, we analyze existing detectors and find that these detectors have difficulty in fully exploiting the power of the annotated points. To solve this, we introduce a new detector, Point DETR, which extends DETR by adding a point encoder. Extensive experiments conducted on MS-COCO dataset in various data settings show the effectiveness of our method. In particular, when using 20% fully labeled data from COCO, our detector achieves a promising performance, 33.3 AP, which outperforms a strong baseline (FCOS) by 2.0 AP, and we demonstrate the point annotations bring over 10 points in various AR metrics.

cs.CV↗

arXiv2021

Validity of Prandtl expansions for steady MHD in the Sobolev framework

This paper is concerned with the vanishing viscosity and magnetic resistivity limit for the two-dimensional steady incompressible MHD system on the half plane with no-slip boundary condition on velocity field and perfectly conducting wall condition on magnetic field. We prove the nonlinear stability of shear flows of Prandtl type with nondegenerate tangential magnetic field, but without any positivity or monotonicity assumption on the velocity field. It is in sharp contrast to the steady Navier-Stokes equations and reflects the stabilization effect of magnetic field. Unlike the unsteady MHD system, we manage the degeneracy on the boundary caused by no-slip boundary condition and obtain the estimates of solutions by introducing an intrinsic weight function and some good auxiliary functions.

math.AP↗

arXiv2021

You Only Look One-level Feature

This paper revisits feature pyramids networks (FPN) for one-stage detectors and points out that the success of FPN is due to its divide-and-conquer solution to the optimization problem in object detection rather than multi-scale feature fusion. From the perspective of optimization, we introduce an alternative way to address the problem instead of adopting the complex feature pyramids - {\em utilizing only one-level feature for detection}. Based on the simple and efficient solution, we present You Only Look One-level Feature (YOLOF). In our method, two key components, Dilated Encoder and Uniform Matching, are proposed and bring considerable improvements. Extensive experiments on the COCO benchmark prove the effectiveness of the proposed model. Our YOLOF achieves comparable results with its feature pyramids counterpart RetinaNet while being $2.5\times$ faster. Without transformer layers, YOLOF can match the performance of DETR in a single-level feature manner with $7\times$ less training epochs. With an image size of $608\times608$, YOLOF achieves 44.3 mAP running at 60 fps on 2080Ti, which is $13\%$ faster than YOLOv4. Code is available at \url{https://github.com/megvii-model/YOLOF}.

cs.CV↗

arXiv2021

Phase diagram and superlattice structures of monolayer phosphorus carbide (P$_x$C$_{1-x}$)

Phase stability and properties of two-dimensional phosphorus carbide, P$_x$C$_{1-x}$, are investigated using the first-principles method in combination with cluster expansion and Monte Carlo simulation. Monolayer P$_x$C$_{1-x}$ is found to be a phase separating system which indicates difficulty in fabricating monolayer P$_x$C$_{1-x}$ or crystalline P$_x$C$_{1-x}$ thin films. Nevertheless, a bottom-up design approach is used to determine the stable structures of P$_x$C$_{1-x}$ of various compositions which turn out to be superlattices consisting of alternating carbon and phosphorus nanoribbons along the armchair direction. Results of first-principles calculations indicate that once these structures are produced, they are mechanically and thermodynamically stable. All the ordered structures are predicted to be semiconductors, with band gap ranging from 0.2 to 1.2 eV. In addition, the monolayer P$_x$C$_{1-x}$ are predicted to have high carrier mobility, and high optical absorption in the ultraviolet region which shows a red-shift as the P:C ratio increases. These properties make 2D P$_x$C$_{1-x}$ promising materials for applications in electronics and optoelectronics.

cond-mat.mtrl-sci↗

arXiv2021

Tunable Rashba spin-orbit coupling and its interplay with multiorbital effect and magnetic ordering at oxide interfaces

The complex oxide heterostructures such as LaAlO3/SrTiO3 (LAO/STO) interface are paradigmatic platforms to explore emerging multi-degrees of freedom coupling and the associated exotic phenomena. In this study, we reveal the effects of multiorbital and magnetic ordering on Rashba spin-orbit coupling (SOC) at the LAO/STO (001) interface. Based on first-principles calculations, we show that the Rashba spin splitting near the conduction band edge can be tuned substantially by the interfacial insulator-metal transition due to the multiorbital effect of the lowest t_2g bands. We further unravel a competition between Rashba SOC and intrinsic magnetism, in which the Rashba SOC induced spin polarization is suppressed by the interfacial magnetic ordering. These results deepen our understanding of intricate electronic and magnetic reconstruction at the perovskite oxide interfaces and shed light on the engineering of oxide heterostructures for all-oxides-based spintronic devices.

cond-mat.mtrl-sci↗

arXiv2020

Generator Pyramid for High-Resolution Image Inpainting

Inpainting high-resolution images with large holes challenges existing deep learning based image inpainting methods. We present a novel framework -- PyramidFill for high-resolution image inpainting task, which explicitly disentangles content completion and texture synthesis. PyramidFill attempts to complete the content of unknown regions in a lower-resolution image, and synthesis the textures of unknown regions in a higher-resolution image, progressively. Thus, our model consists of a pyramid of fully convolutional GANs, wherein the content GAN is responsible for completing contents in the lowest-resolution masked image, and each texture GAN is responsible for synthesizing textures in a higher-resolution image. Since completing contents and synthesising textures demand different abilities from generators, we customize different architectures for the content GAN and texture GAN. Experiments on multiple datasets including CelebA-HQ, Places2 and a new natural scenery dataset (NSHQ) with different resolutions demonstrate that PyramidFill generates higher-quality inpainting results than the state-of-the-art methods. To better assess high-resolution image inpainting methods, we will release NSHQ, high-quality natural scenery images with high-resolution 1920$\times$1080.

cs.CV↗

arXiv2020

Well-posedness of the MHD boundary layer system in Gevrey function space without Structural Assumption

We establish the well-posedness of the MHD boundary layer system in Gevrey function space without any structural assumption. Compared to the classical Prandtl equation, the loss of tangential derivative comes from both the velocity and magnetic fields that are coupled with each other. By observing a new type of cancellation mechanism in the system for overcoming the loss derivative degeneracy, we show that the MHD boundary layer system is well-posed with Gevrey index up to $3/2$ in both two and three dimensional spaces.

math.AP↗

arXiv2020

Spectral Clustering with Smooth Tiny Clusters

Spectral clustering is one of the most prominent clustering approaches. The distance-based similarity is the most widely used method for spectral clustering. However, people have already noticed that this is not suitable for multi-scale data, as the distance varies a lot for clusters with different densities. State of the art(ROSC and CAST ) addresses this limitation by taking the reachability similarity of objects into account. However, we observe that in real-world scenarios, data in the same cluster tend to present in a smooth manner, and previous algorithms never take this into account. Based on this observation, we propose a novel clustering algorithm, which con-siders the smoothness of data for the first time. We first divide objects into a great many tiny clusters. Our key idea is to cluster tiny clusters, whose centers constitute smooth graphs. Theoretical analysis and experimental results show that our clustering algorithm significantly outperforms state of the art. Although in this paper, we singly focus on multi-scale situations, the idea of data smoothness can certainly be extended to any clustering algorithms

cs.LG↗

arXiv2020

NagE: Non-Abelian Group Embedding for Knowledge Graphs

We demonstrated the existence of a group algebraic structure hidden in relational knowledge embedding problems, which suggests that a group-based embedding framework is essential for designing embedding models. Our theoretical analysis explores merely the intrinsic property of the embedding problem itself hence is model-independent. Motivated by the theoretical analysis, we have proposed a group theory-based knowledge graph embedding framework, in which relations are embedded as group elements, and entities are represented by vectors in group action spaces. We provide a generic recipe to construct embedding models associated with two instantiating examples: SO3E and SU2E, both of which apply a continuous non-Abelian group as the relation embedding. Empirical experiments using these two exampling models have shown state-of-the-art results on benchmark datasets.

cs.AI↗

arXiv2020

Space- and Computationally-Efficient Set Reconciliation via Parity Bitmap Sketch (PBS)

Set reconciliation is a fundamental algorithmic problem that arises in many networking, system, and database applications. In this problem, two large sets A and B of objects (bitcoins, files, records, etc.) are stored respectively at two different network-connected hosts, which we name Alice and Bob respectively. Alice and Bob communicate with each other to learn $AΔB$, the difference between A and B, and as a result the reconciled set $A\bigcup B$. Current set reconciliation schemes are based on either Invertible Bloom Filters (IBF) or Error-Correction Codes (ECC). The former has a low computational complexity of O(d), where d is the cardinality of $AΔB$, but has a high communication overhead that is several times larger than the theoretical minimum. The latter has a low communication overhead close to the theoretical minimum, but has a much higher computational complexity of $O(d^2)$. In this work, we propose Parity Bitmap Sketch (PBS), an ECC- based set reconciliation scheme that gets the better of both worlds: PBS has both a low computational complexity of O(d) just like IBF-based solutions and a low communication overhead of roughly twice the theoretical minimum. A separate contribution of this work is a novel rigorous analytical framework that can be used for the precise calculation of various performance metrics and for the near-optimal parameter tuning of PBS.

cs.DB↗

arXiv2020

Variance Regularization for Accelerating Stochastic Optimization

While nowadays most gradient-based optimization methods focus on exploring the high-dimensional geometric features, the random error accumulated in a stochastic version of any algorithm implementation has not been stressed yet. In this work, we propose a universal principle which reduces the random error accumulation by exploiting statistic information hidden in mini-batch gradients. This is achieved by regularizing the learning-rate according to mini-batch variances. Due to the complementarity of our perspective, this regularization could provide a further improvement for stochastic implementation of generic 1st order approaches. With empirical results, we demonstrated the variance regularization could speed up the convergence as well as stabilize the stochastic optimization.

cs.LG↗