arXiv subjects

Can Wang

Publications and source records attributed to Can Wang.

At least 181 records · Page 10Linked to original sources

arXiv2022

Numerical method for the Fokker-Planck equation of Brownian motion subordinated by inverse tempered stable subordinator with drift

In this work, based on the complete Bernstein function, we propose a generalized regularity analysis including maximal $\mathrm{L}^p$ regularity for the Fokker--Planck equation, which governs the subordinated Brownian motion with the inverse tempered stable subordinator that has a drift. We derive a generalized time--stepping finite element scheme based on the backward Euler convolution quadrature, and the optimal-order convergence of the numerical solutions is established using the proven solution regularity. Further, the analysis is generalized to more general diffusion equations. Numerical experiments are provided to support the theoretical results.

math.NA↗

arXiv2022

A sharp $α$-robust $L1$ scheme on graded meshes for two-dimensional time tempered fractional Fokker-Planck equation

In this paper, we are concerned with the numerical solution for the two-dimensional time fractional Fokker-Planck equation with tempered fractional derivative of order $α$. Although some of its variants are considered in many recent numerical analysis papers, there are still some significant differences. Here we first provide the regularity estimates of the solution. And then a modified $L$1 scheme inspired by the middle rectangle quadrature formula on graded meshes is employed to compensate for the singularity of the solution at $t\rightarrow 0^{+}$, while the five-point difference scheme is used in space. Stability and convergence are proved in the sence of $L^{\infty}$ norm, then a sharp error estimate $\mathscr{O}(τ^{\min\{2-α, rα\}})$ is derived on graded meshes. Furthermore, unlike the bounds proved in the previous works, the constant multipliers in our analysis do not blow up as the Caputo fractional derivative $α$ approaches the classical value of 1. Finally, we perform the numerical experiments to verify the effectiveness and convergence order of the presented algorithms.

math.NA↗

arXiv2022

Alignahead: Online Cross-Layer Knowledge Extraction on Graph Neural Networks

Existing knowledge distillation methods on graph neural networks (GNNs) are almost offline, where the student model extracts knowledge from a powerful teacher model to improve its performance. However, a pre-trained teacher model is not always accessible due to training cost, privacy, etc. In this paper, we propose a novel online knowledge distillation framework to resolve this problem. Specifically, each student GNN model learns the extracted local structure from another simultaneously trained counterpart in an alternating training procedure. We further develop a cross-layer distillation strategy by aligning ahead one student layer with the layer in different depth of another student model, which theoretically makes the structure information spread over all layers. Experimental results on five datasets including PPI, Coauthor-CS/Physics and Amazon-Computer/Photo demonstrate that the student performance is consistently boosted in our collaborative training framework without the supervision of a pre-trained teacher model. In addition, we also find that our alignahead technique can accelerate the model convergence speed and its effectiveness can be generally improved by increasing the student numbers in training. Code is available: https://github.com/GuoJY-eatsTG/Alignahead

cs.LG↗

arXiv2022

Knowledge Distillation with the Reused Teacher Classifier

Knowledge distillation aims to compress a powerful yet cumbersome teacher model into a lightweight student model without much sacrifice of performance. For this purpose, various approaches have been proposed over the past few years, generally with elaborately designed knowledge representations, which in turn increase the difficulty of model development and interpretation. In contrast, we empirically show that a simple knowledge distillation technique is enough to significantly narrow down the teacher-student performance gap. We directly reuse the discriminative classifier from the pre-trained teacher model for student inference and train a student encoder through feature alignment with a single $\ell_2$ loss. In this way, the student model is able to achieve exactly the same performance as the teacher model provided that their extracted features are perfectly aligned. An additional projector is developed to help the student encoder match with the teacher classifier, which renders our technique applicable to various teacher and student architectures. Extensive experiments demonstrate that our technique achieves state-of-the-art results at the modest cost of compression ratio due to the added projector.

cs.CV↗

arXiv2022

Strong light-matter interactions between gap plasmons and two-dimensional excitons at ambient condition in a deterministic way

Strong exciton-plasmon interaction between the layered two-dimensional (2D) semiconductors and gap plasmons shows a great potential to implement cavity quantum-electrodynamics in ambient condition. However, achieving a robust plasmon-exciton coupling with nanocavity is still very challenging, because the layer area is usually small with conventional approaches. Here, we report on a robust strong exciton-plasmon coupling between the gap mode of bowtie and the excitons in MoS$_2$ layers with gold-assisted mechanical exfoliation and the nondestructive wet transfer techniques for large-area layer. Benefiting from the ultrasmall mode volume and strong in-plane field, the estimated effective exciton number contributing to the coupling is largely reduced. With a corrected exciton transition dipole moment, the exciton numbers are extracted with 40 for the case of monolayer and 48 for 8 layers. Our work paves a way to realize the strong coupling with 2D materials with few excitons at room temperature.

cond-mat.mes-hall↗

arXiv2022

CLIP-NeRF: Text-and-Image Driven Manipulation of Neural Radiance Fields

We present CLIP-NeRF, a multi-modal 3D object manipulation method for neural radiance fields (NeRF). By leveraging the joint language-image embedding space of the recent Contrastive Language-Image Pre-Training (CLIP) model, we propose a unified framework that allows manipulating NeRF in a user-friendly way, using either a short text prompt or an exemplar image. Specifically, to combine the novel view synthesis capability of NeRF and the controllable manipulation ability of latent representations from generative models, we introduce a disentangled conditional NeRF architecture that allows individual control over both shape and appearance. This is achieved by performing the shape conditioning via applying a learned deformation field to the positional encoding and deferring color conditioning to the volumetric rendering stage. To bridge this disentangled latent representation to the CLIP embedding, we design two code mappers that take a CLIP embedding as input and update the latent codes to reflect the targeted editing. The mappers are trained with a CLIP-based matching loss to ensure the manipulation accuracy. Furthermore, we propose an inverse optimization method that accurately projects an input image to the latent codes for manipulation to enable editing on real images. We evaluate our approach by extensive experiments on a variety of text prompts and exemplar images and also provide an intuitive interface for interactive editing. Our implementation is available at https://cassiepython.github.io/clipnerf/

cs.CV↗

arXiv2022

Cross-Domain and Disentangled Face Manipulation with 3D Guidance

Face image manipulation via three-dimensional guidance has been widely applied in various interactive scenarios due to its semantically-meaningful understanding and user-friendly controllability. However, existing 3D-morphable-model-based manipulation methods are not directly applicable to out-of-domain faces, such as non-photorealistic paintings, cartoon portraits, or even animals, mainly due to the formidable difficulties in building the model for each specific face domain. To overcome this challenge, we propose, as far as we know, the first method to manipulate faces in arbitrary domains using human 3DMM. This is achieved through two major steps: 1) disentangled mapping from 3DMM parameters to the latent space embedding of a pre-trained StyleGAN2 that guarantees disentangled and precise controls for each semantic attribute; and 2) cross-domain adaptation that bridges domain discrepancies and makes human 3DMM applicable to out-of-domain faces by enforcing a consistent latent space embedding. Experiments and comparisons demonstrate the superiority of our high-quality semantic manipulation method on a variety of face domains with all major 3D facial attributes controllable-pose, expression, shape, albedo, and illumination. Moreover, we develop an intuitive editing interface to support user-friendly control and instant feedback. Our project page is https://cassiepython.github.io/cddfm3d/index.html

cs.CV↗

arXiv2022

Confidence-Aware Multi-Teacher Knowledge Distillation

Knowledge distillation is initially introduced to utilize additional supervision from a single teacher model for the student model training. To boost the student performance, some recent variants attempt to exploit diverse knowledge sources from multiple teachers. However, existing studies mainly integrate knowledge from diverse sources by averaging over multiple teacher predictions or combining them using other various label-free strategies, which may mislead student in the presence of low-quality teacher predictions. To tackle this problem, we propose Confidence-Aware Multi-teacher Knowledge Distillation (CA-MKD), which adaptively assigns sample-wise reliability for each teacher prediction with the help of ground-truth labels, with those teacher predictions close to one-hot labels assigned large weights. Besides, CA-MKD incorporates intermediate layers to stable the knowledge transfer process. Extensive experiments show that our CA-MKD consistently outperforms all compared state-of-the-art methods across various teacher-student architectures.

cs.LG↗

arXiv2022

Ruling out real-valued standard formalism of quantum theory

Standard quantum theory was formulated with complex-valued Schrodinger equations, wave functions, operators, and Hilbert spaces. Previous work attempted to simulate quantum systems using only real numbers by exploiting an enlarged Hilbert space. A fundamental question arises: are complex numbers really necessary in the standard formalism of quantum theory? To answer this question, a quantum game has been developed to distinguish standard quantum theory from its real-number analog by revealing a contradiction in the maximum game scores between a high-fidelity multi-qubit quantum experiment and players using only real-number quantum theory. Here, using superconducting qubits, we faithfully experimentally implement the quantum game based on entanglement swapping with a state-of-the-art fidelity of 0.952(1), which beats the real-number bound of 7.66 by 43 standard deviations. Our results disprove the real-number formulation and establish the indispensable role of complex numbers in the standard quantum theory.

quant-ph↗

arXiv2022

Pseudo-Labeled Auto-Curriculum Learning for Semi-Supervised Keypoint Localization

Localizing keypoints of an object is a basic visual problem. However, supervised learning of a keypoint localization network often requires a large amount of data, which is expensive and time-consuming to obtain. To remedy this, there is an ever-growing interest in semi-supervised learning (SSL), which leverages a small set of labeled data along with a large set of unlabeled data. Among these SSL approaches, pseudo-labeling (PL) is one of the most popular. PL approaches apply pseudo-labels to unlabeled data, and then train the model with a combination of the labeled and pseudo-labeled data iteratively. The key to the success of PL is the selection of high-quality pseudo-labeled samples. Previous works mostly select training samples by manually setting a single confidence threshold. We propose to automatically select reliable pseudo-labeled samples with a series of dynamic thresholds, which constitutes a learning curriculum. Extensive experiments on six keypoint localization benchmark datasets demonstrate that the proposed approach significantly outperforms the previous state-of-the-art SSL approaches.

cs.CV↗

arXiv2022

Strain-engineered high-temperature ferromagnetic Oxygen-substituted NaMnF3 from first principles

Using first-principles calculations, we investigated the magnetic, electronic, and structural properties of oxygen-substituted NaMnF3 (NaMnF1.5O1.5) with in-plane biaxial strain. For simplicity, a structure containing an oxygen octahedron is used to explore the underlying physical mechanism. We found that the oxygen octahedron induces a transition from an insulating antiferromagnet to a high-temperature half-metallic ferromagnet. More importantly, the Curie temperature can be significantly enhanced and even might reach room temperature by applying tensile strain. The changing trends of exchange coupling constants with the increasing biaxial tensile strain can be attributed to the cooperative effects of Jahn-Teller distortion and rotation distortion. It is expected that these findings can enrich the versatility of NaMnF3 and make it a promising candidate for spintronic applications.

cond-mat.mtrl-sci↗

arXiv2021

Ferromagnetic Enhancement in LaMnO3 Films with Release and Flexure

A variety of novel phenomena and functionalities emerge from lowering the dimensionality of materials and enriching the degrees of freedom in modulation. In this work, it is found that the saturation magnetization of LaMnO3 (LMO) films is largely enhanced by 56% after releasing from a brand-new phase of tetragonal strontium aluminate buffer layer, and is significantly increased by 92% with bending films to a curvature of 1 mm-1 using a water-assisted direct-transferring method. Meanwhile, the Curie temperature of LMO films has been improved by 13 K. High-resolution spherical aberration-corrected scanning transmission electron microscopy and first-principles calculations unambiguously demonstrate that the enhanced ferromagnetism is attributed to the strengthened Mn-O-Mn super-exchange interactions from the augmented characteristics of the unconventional P21/n structure caused by the out-of-plane lattice shrinking after strain releasing and increased flexure degree of freestanding LMO films. This work paves a way to achieve large-scale and crack-and-wrinkle-free freestanding films of oxides with largely improved functionalities.

cond-mat.mtrl-sci↗

arXiv2021

Enhanced valley polarization in WS$_2$/LaMnO$_3$ heterostructure

Monolayer transition metal dichalcogenides have attracted great attentions for potential applications in valleytronics. However, the valley polarization degree is usually not high because of the intervalley scattering. Here, we demonstrate a largely enhanced valley polarization up to 80\% in monolayer WS$_2$ under non-resonant excitation at 4.2 K using WS$_2$/LaMnO$_3$ thin film heterostructure, which is much higher than that for monolayer WS$_2$ on SiO$_2$/Si substrate with a valley polarization of 15\%. Furthermore, the greatly enhanced valley polarization can be maintained to a high temperature of about 160 K with a valley polarization of 53\%. The temperature dependence of valley polarization is strongly correlated with the thermomagnetic curve of LaMnO$_3$, indicating an exciton-magnon coupling between WS$_2$ and LaMnO$_3$. A simple model is introduced to illustrate the underlying mechanisms. The coupling of WS$_2$ and LaMnO$_3$ is further confirmed with an observation of two interlayer excitons with opposite valley polarizations in the heterostructure, resulting from the spin-orbit coupling induced splitting of the conduction bands in monolayer transition metal dichalcogenides. Our results provide a pathway to control the valleytronic properties of transition metal dichalcogenides by means of ferromagnetic van der Waals engineering, paving a way to practical valleytronic applications.

cond-mat.mtrl-sci↗

arXiv2021

Room-temperature ferromagnetism at an oxide/nitride interface

Heterointerfaces have led to the discovery of novel electronic and magnetic states because of their strongly entangled electronic degrees of freedom. Single-phase chromium compounds always exhibit antiferromagnetism following the prediction of Goodenough-Kanamori rules. So far, exchange coupling between chromium ions via hetero-anions has not been explored and the associated quantum states is unknown. Here we report the successful epitaxial synthesis and characterizations of chromium oxide (Cr2O3)-chromium nitride (CrN) superlattices. Room-temperature ferromagnetic spin ordering is achieved at the interfaces between these two antiferromagnets, and the magnitude of the effect decays with increasing layer thickness. First-principles calculations indicate that robust ferromagnetic spin interaction between Cr3+ ions via anion-hybridizations across the interface yields the lowest total energy. This work opens the door to fundamental understanding of the unexpected and exceptional properties of oxide-nitride interfaces and provides access to hidden phases at low-dimensional quantum heterostructures.

cond-mat.mtrl-sci↗

arXiv2021

Exchange coupling in synthetic anion-engineered chromia heterostructures

Control of magnetic states by external factors has garnered a mainstream status in spintronic research for designing low power consumption and fast-response information storage and processing devices. Previously, magnetic-cation substitution is the conventional means to induce ferromagnetism in an intrinsic antiferromagnet. Theoretically, the anion-doping is proposed to be another effect means to change magnetic ground states. Here we demonstrate the synthesis of high-quality single-phase chromium oxynitride thin films using in-situ nitrogen doping. Unlike antiferromagnetic monoanionic chromium oxide and nitride phases, chromium oxynitride exhibits a robust ferromagnetic and insulating state, as demonstrated by the combination of multiple magnetization probes and theoretical calculations. With increasing the nitrogen content, the crystal structure of chromium oxynitride transits from trigonal (R3c) to tetragonal (4mm) phase and its saturation magnetization reduces significantly. Furthermore, we achieve a large and controllable exchange bias field in the chromia heterostructures by synthetic anion engineering. This work reflects the anion engineering in functional oxides towards the potential applications in giant magnetoresistance and tunnelling junctions of modern magnetic sensors and read heads.

cond-mat.mtrl-sci↗

arXiv2021

Dynamics of anisotropic oxygen-ion migration in strained cobaltites

Orientation control of oxygen vacancy channel (OVC) is a highly desirable for tailoring oxygen diffusion as it serves fast transport channel in ion conductors, which is widespread exploited in solid-state fuel cells, catalysts, and ion-batteries. Direct observation of oxygen-ions hopping towards preferential vacant sites is a key to clarifying migration pathways. Here we report the anisotropic oxygen-ion migration mediated by strain in ultrathin cobaltites via in-situ thermal activation in an atomic-resolved transmission electron microscopy. Oxygen migration pathways are constructed on the basis of the atomic structure during the OVC switching, which is manifested as the vertical-to-horizontal OVC switching under tensile strain, but the horizontal-to-diagonal switching under compression. We evaluate the topotactic structural changes to OVC, determine the crucial role of tolerance factor for OVC stability and establish the strain-dependent phase diagram. Our work provides a practical guide for engineering OVC orientation that is applicable ionic-oxide electronics.

cond-mat.mtrl-sci↗

arXiv2021

Anisotropic electronic phase transition in CrN epitaxial thin films

Electronic phase transition in strongly correlated materials is extremely sensitive to the dimensionality and crystallographic orientations. Transition metal nitrides (TMNs) are seldom investigated due to the difficulty in fabricating the high-quality and stoichiometric single crystals. In this letter, we report the epitaxial growth and electronic properties of CrN films on different-oriented NdGaO3 (NGO) substrates. Astonishingly, the CrN films grown on (110)-oriented NGO substrates maintain a metallic phase, whereas the CrN films grown on (010)-oriented NGO substrates are semiconducting. We attribute the unconventional electronic transition in the CrN films to the strongly correlation with epitaxial strain. The effective modulation of bandgap by the anisotropic strain triggers the metal-to-insulator transition consequently. This work provides a convenient approach to modify the electronic ground states of functional materials using anisotropic strain and further stimulates the investigations of TMNs.

cond-mat.mtrl-sci↗

arXiv2021

A scalable tripartite Wigner's friend scenario

Wigner's friend thought experiment is intended to reveal the inherent tension between unitary evolution and measurement collapse. On the basis of Wigner's friend experiment, Brukner derives a no-go theorem for observer-independent facts. We construct an extended Wigner's friend scenario including three laboratories, namely, Alice's laboratory, Bob's laboratory and Charlie's laboratory, where Alice, Bob and Charlie are standing outside the laboratories while their friends are placed inside their own laboratories. We consider quantum simulation via Q\# quantum programming and also realize the primary quantum circuits using IBM quantum computers. Then, we calculate the probabilities and corresponding statistical uncertainties. It has been shown that the results of quantum simulation are clearly consistent with theoretical values, while it has a slightly higher error rates for the experimental results of quantum computers mainly because of a series of quantum gates, especially CNOT gates.

quant-ph↗