arXiv subjects

Jian Wang

Publications and source records attributed to Jian Wang.

At least 631 records · Page 35Linked to original sources

arXiv2023

Discrete scale invariance of the quasi-bound states at atomic vacancies in a topological material

Recently, log-periodic quantum oscillations have been detected in topological materials zirconium pentatelluride (ZrTe5) and hafnium pentatelluride (HfTe5), displaying intriguing discrete scale invariance (DSI) characteristic. In condensed materials, the DSI is considered to be related to the quasi-bound states formed by massless Dirac fermions with strong Coulomb attraction, offering a feasible platform to study the long-pursued atomic-collapse phenomenon. Here, we demonstrate that a variety of atomic vacancies in the topological material HfTe5 can host the geometric quasi-bound states with DSI feature, resembling the artificial supercritical atom collapse. The density of states of these quasi-bound states are enhanced and the quasi-bound states are spatially distributed in the "orbitals" surrounding the vacancy sites, which are detected and visualized by low-temperature scanning tunneling microscope/spectroscopy (STM/S). By applying the perpendicular magnetic fields, the quasi-bound states at lower energies become wider and eventually invisible, meanwhile the energies of quasi-bound states move gradually towards the Fermi energy (EF). These features are consistent with the theoretical prediction of a magnetic-field-induced transition from supercritical to subcritical states. The direct observation of geometric quasi-bound states sheds light on the deep understanding of the DSI in quantum materials.

cond-mat.mtrl-sci↗

arXiv2023

MetaMorph: Learning Metamorphic Image Transformation With Appearance Changes

This paper presents a novel predictive model, MetaMorph, for metamorphic registration of images with appearance changes (i.e., caused by brain tumors). In contrast to previous learning-based registration methods that have little or no control over appearance-changes, our model introduces a new regularization that can effectively suppress the negative effects of appearance changing areas. In particular, we develop a piecewise regularization on the tangent space of diffeomorphic transformations (also known as initial velocity fields) via learned segmentation maps of abnormal regions. The geometric transformation and appearance changes are treated as joint tasks that are mutually beneficial. Our model MetaMorph is more robust and accurate when searching for an optimal registration solution under the guidance of segmentation, which in turn improves the segmentation performance by providing appropriately augmented training labels. We validate MetaMorph on real 3D human brain tumor magnetic resonance imaging (MRI) scans. Experimental results show that our model outperforms the state-of-the-art learning-based registration models. The proposed MetaMorph has great potential in various image-guided clinical interventions, e.g., real-time image-guided navigation systems for tumor removal surgery.

cs.CV↗

arXiv2023

Exploring Adversarial Attacks on Neural Networks: An Explainable Approach

Deep Learning (DL) is being applied in various domains, especially in safety-critical applications such as autonomous driving. Consequently, it is of great significance to ensure the robustness of these methods and thus counteract uncertain behaviors caused by adversarial attacks. In this paper, we use gradient heatmaps to analyze the response characteristics of the VGG-16 model when the input images are mixed with adversarial noise and statistically similar Gaussian random noise. In particular, we compare the network response layer by layer to determine where errors occurred. Several interesting findings are derived. First, compared to Gaussian random noise, intentionally generated adversarial noise causes severe behavior deviation by distracting the area of concentration in the networks. Second, in many cases, adversarial examples only need to compromise a few intermediate blocks to mislead the final decision. Third, our experiments revealed that specific blocks are more vulnerable and easier to exploit by adversarial examples. Finally, we demonstrate that the layers $Block4\_conv1$ and $Block5\_cov1$ of the VGG-16 model are more susceptible to adversarial attacks. Our work could provide valuable insights into developing more reliable Deep Neural Network (DNN) models.

cs.LG↗

arXiv2023

Universal behavior of magnon-mediated spin transport in disordered NM/NM/FI heterostructure

We numerically investigate the magnon-mediated spin transport and scaling behaviors of a nonmagnetic metal/nonmagnetic metal/ferromagnetic insulator (NM/NM/FI) heterostructure in the presence of Anderson disorders. For the two-dimensional (2D) NM/NM/FI system, an enhancement of spin conductance in the weak disorder regime is found due to the increasing of the interfacial density of states (DOS) at NM/FI interface. As a result, a new scaling regime is uncovered in the metallic regime where the spin conductance fluctuation rms(G) scales linearly with the average spin conductance , independent of system parameters such as Fermi energies and temperatures. The competition between the disorder-enhanced interfacial DOS and disorder-suppressed spin transport results in a non-monotonic dependence of average spin conductance on disorder strength. In the localized regime, the variance of lnG for different system parameters follows a universal function that depends linearly on the average of logarithm of spin conductance .The distribution of lnG in the localized regime is non-Gaussian whose deviation from Gaussian can be characterized by the third and fourth order cumulants of lnG, denoted as k3 and k4, respectively. We find that k3~<-lnG>^3/2 and k4~ ^2. This suggests that the spin conductance mediated by magnon in NM/NM/FI hybrid systems belongs to a different universality class. For fixed , the lnG distribution in the localized regime for different system parameters is found to collapse onto a single curve, suggesting the distribution P(lnG; ) is a universal function that depends only on . Furthermore, spin thermopower S in the presence of disorder is studied and the variance of lnS is found to scale linearly with for different system parameters in the localized regime. Spin conductance and spin thermopower for 1D NM/NM/FI are also studied.

cond-mat.mes-hall↗

arXiv2023

Temporal Segment Transformer for Action Segmentation

Recognizing human actions from untrimmed videos is an important task in activity understanding, and poses unique challenges in modeling long-range temporal relations. Recent works adopt a predict-and-refine strategy which converts an initial prediction to action segments for global context modeling. However, the generated segment representations are often noisy and exhibit inaccurate segment boundaries, over-segmentation and other problems. To deal with these issues, we propose an attention based approach which we call \textit{temporal segment transformer}, for joint segment relation modeling and denoising. The main idea is to denoise segment representations using attention between segment and frame representations, and also use inter-segment attention to capture temporal correlations between segments. The refined segment representations are used to predict action labels and adjust segment boundaries, and a final action segmentation is produced based on voting from segment masks. We show that this novel architecture achieves state-of-the-art accuracy on the popular 50Salads, GTEA and Breakfast benchmarks. We also conduct extensive ablations to demonstrate the effectiveness of different components of our design.

cs.CV↗

arXiv2023

Planckian Dissipation and non-Ginzburg-Landau Type Upper Critical Field in Bi2201

Resistivity and Hall effect measurements have been carried out on a micro-fabricated bridge of Bi2201 single crystal at low temperatures down to 0.4 K under high magnetic fields. When superconductivity is crashed by a high magnetic field, the recovered "normal state" resistivity still shows a linear temperature dependence in low temperature region. Combining with the effective mass and the charge carrier density, we get a linear scattering rate $1/τ= αk_{B} T/\hbar$ with $0.77<α<1.16$, which gives a strong evidence of the Planckian dissipation. Furthermore, our results reveal a new type of temperature dependence of upper critical field, $H_{c2}(T)=H^*\sqrt{(1-t)/(t+0.154)}$, which is totally different from the expectation of the Ginzburg-Landau theory, and suggests uncondensed Cooper pairs above $H_{c2}(T)$ line.

cond-mat.supr-con↗

arXiv2023

Flexoelectricity-stabilized ferroelectric phase with enhanced reliability in ultrathin La:HfO2 films

Doped HfO2 thin films exhibit robust ferroelectric properties even for nanometric thicknesses, are compatible with current Si technology and thus have great potential for the revival of integrated ferroelectrics. Phase control and reliability are core issues for their applications. Here we show that, in (111)-oriented 5%La:HfO2 (HLO) epitaxial thin films deposited on (La0.3Sr0.7)(Al0.65Ta0.35)O3 substrates, the flexoelectric effect, arising from the strain gradient along the films normal, induces a rhombohedral distortion in the otherwise Pca21 orthorhombic structure. Density functional calculations reveal that the distorted structure is indeed more stable than the pure Pca21 structure, when applying an electric field mimicking the flexoelectric field. This rhombohedral distortion greatly improves the fatigue endurance of HLO thin films by further stabilizing the metastable ferroelectric phase against the transition to the thermodynamically stable non-polar monoclinic phase during repetitive cycling. Our results demonstrate that the flexoelectric effect, though negligibly weak in bulk, is crucial to optimize the structure and properties of doped HfO2 thin films with nanometric thicknesses for integrated ferroelectric applications.

cond-mat.mtrl-sci↗

arXiv2023

Differentiable Rotamer Sampling with Molecular Force Fields

Molecular dynamics is the primary computational method by which modern structural biology explores macromolecule structure and function. Boltzmann generators have been proposed as an alternative to molecular dynamics, by replacing the integration of molecular systems over time with the training of generative neural networks. This neural network approach to MD samples rare events at a higher rate than traditional MD, however critical gaps in the theory and computational feasibility of Boltzmann generators significantly reduce their usability. Here, we develop a mathematical foundation to overcome these barriers; we demonstrate that the Boltzmann generator approach is sufficiently rapid to replace traditional MD for complex macromolecules, such as proteins in specific applications, and we provide a comprehensive toolkit for the exploration of molecular energy landscapes with neural networks.

physics.comp-ph↗

arXiv2023

Promoting Cooperation in Multi-Agent Reinforcement Learning via Mutual Help

Multi-agent reinforcement learning (MARL) has achieved great progress in cooperative tasks in recent years. However, in the local reward scheme, where only local rewards for each agent are given without global rewards shared by all the agents, traditional MARL algorithms lack sufficient consideration of agents' mutual influence. In cooperative tasks, agents' mutual influence is especially important since agents are supposed to coordinate to achieve better performance. In this paper, we propose a novel algorithm Mutual-Help-based MARL (MH-MARL) to instruct agents to help each other in order to promote cooperation. MH-MARL utilizes an expected action module to generate expected other agents' actions for each particular agent. Then, the expected actions are delivered to other agents for selective imitation during training. Experimental results show that MH-MARL improves the performance of MARL both in success rate and cumulative reward.

cs.LG↗

arXiv2023

Analytic two-loop master integrals for $tW$ production at hadron colliders. Part II

We present analytic results of the two-loop master integrals for hadronic $tW$ production that contain two massive propagators. For the planar integral family, we succeed in constructing the canonical basis, so the results are written in terms of multiple polylogarithms. The master integrals in the non-planar integral families have been calculated by Taylor series expansion around $m_W^2=0$. Even with this simplification, there are multiple square roots involved in the differential equations, which can not be rationalized simultaneously. We find that a proper linear combination of the integral basis alleviates the problem so that the results are expressed by multiple polylogarithms or single integrals over multiple polylogarithms up to weight four. These single integrals can be evaluated using the classic Newton-Cotes formulas. Our analytic results of master integrals would be used in the computation of the two-loop amplitudes for $tW$ production.

hep-ph↗

arXiv2023

Reconfigurable integrated full-dimensional optical lattice generator

Optical lattices with periodic potentials have attracted great attention in modern optics and photonics, enabling extensive applications in atomic manipulation, optical trapping, optical communications, imaging, sensing, etc. In the last decade, the generation of optical lattices has been widely investigated by various approaches such as multi-plane-wave interferometer, beam superposition, spatial light modulators, nanophotonic circuits, etc. However, all of the previous state-of-the-art works are restricted to only one or two dimensions of the light field, which cannot fulfill the increasing demand on complex light manipulation. Full-dimensional and dynamic control of the light field, including spatial amplitude, phase and polarization, is quite challenging and indispensable for the generation of sophisticated optical lattices. Here, we propose and demonstrate a reconfigurable integrated full-dimensional optical lattice generator, i.e. a photonic emitting array (PEA) enabling reconfigurable and full-dimensional manipulation of optical lattices, in which 4x4 photonic emitting units (PEUs) with 64 thermo-optic microheaters are densely integrated on a silicon chip. By engineering each PEU precisely with independent and complete control of optical properties of amplitude, phase and polarization, various optical vortex lattices, cylindrical vector beam lattices, and vector vortex beam lattices can be generated and reconfigured in the far field. The demonstrated integrated optical lattice generator paves the way for the miniaturization, full-dimensional control and enhanced flexibility of complex light manipulation.

physics.optics↗

arXiv2023

Low Entropy Communication in Multi-Agent Reinforcement Learning

Communication in multi-agent reinforcement learning has been drawing attention recently for its significant role in cooperation. However, multi-agent systems may suffer from limitations on communication resources and thus need efficient communication techniques in real-world scenarios. According to the Shannon-Hartley theorem, messages to be transmitted reliably in worse channels require lower entropy. Therefore, we aim to reduce message entropy in multi-agent communication. A fundamental challenge is that the gradients of entropy are either 0 or infinity, disabling gradient-based methods. To handle it, we propose a pseudo gradient descent scheme, which reduces entropy by adjusting the distributions of messages wisely. We conduct experiments on two base communication frameworks with six environment settings and find that our scheme can reduce message entropy by up to 90% with nearly no loss of cooperation performance.

cs.MA↗

arXiv2023

Improving Zero-Shot Coordination Performance Based on Policy Similarity

Over these years, multi-agent reinforcement learning has achieved remarkable performance in multi-agent planning and scheduling tasks. It typically follows the self-play setting, where agents are trained by playing with a fixed group of agents. However, in the face of zero-shot coordination, where an agent must coordinate with unseen partners, self-play agents may fail. Several methods have been proposed to handle this problem, but they either take a lot of time or lack generalizability. In this paper, we firstly reveal an important phenomenon: the zero-shot coordination performance is strongly linearly correlated with the similarity between an agent's training partner and testing partner. Inspired by it, we put forward a Similarity-Based Robust Training (SBRT) scheme that improves agents' zero-shot coordination performance by disturbing their partners' actions during training according to a pre-defined policy similarity value. To validate its effectiveness, we apply our scheme to three multi-agent reinforcement learning frameworks and achieve better performance compared with previous methods.

cs.MA↗

arXiv2023

Minimum degree thresholds for Hamilton $(\ell,k-\ell)$-cycles in $k$-uniform hypergraphs

Let $n>k>\ell$ be positive integers. We say a $k$-uniform hypergraph $\mathcal{H}$ contains a Hamilton $(\ell,k-\ell)$-cycle if there is a partition $(L_0,R_0,L_1,R_1,\ldots,L_{t-1},R_{t-1})$ of $V(\mathcal{H})$ with $|L_i|=\ell$, $|R_i|=k-\ell$ such that $L_i\cup R_i$ and $R_i\cup L_{i+1}$ (subscripts module $t$) are all edges of $\mathcal{H}$ for $i=0,1,\ldots,t-1$. In the present paper, we determine the tight minimum $\ell$-degree condition that guarantees the existence of a Hamilton $(\ell,k-\ell)$-cycle in every $k$-uniform $n$-vertex hypergraph for $k\geq 7$, $k/2\leq \ell\leq k-1$ and sufficiently large $n\in k\mathbb{N}$.

math.CO↗

arXiv2023

FE-TCM: Filter-Enhanced Transformer Click Model for Web Search

Constructing click models and extracting implicit relevance feedback information from the interaction between users and search engines are very important to improve the ranking of search results. Using neural network to model users' click behaviors has become one of the effective methods to construct click models. In this paper, We use Transformer as the backbone network of feature extraction, add filter layer innovatively, and propose a new Filter-Enhanced Transformer Click Model (FE-TCM) for web search. Firstly, in order to reduce the influence of noise on user behavior data, we use the learnable filters to filter log noise. Secondly, following the examination hypothesis, we model the attraction estimator and examination predictor respectively to output the attractiveness scores and examination probabilities. A novel transformer model is used to learn the deeper representation among different features. Finally, we apply the combination functions to integrate attractiveness scores and examination probabilities into the click prediction. From our experiments on two real-world session datasets, it is proved that FE-TCM outperforms the existing click models for the click prediction.

cs.IR↗

arXiv2023

Third-order intrinsic anomalous Hall effect with generalized semiclassical theory

The linear intrinsic anomalous Hall effect (IAHE) and second-order IAHE have been intensively investigated in time-reversal broken systems. However, as one of the important members of the nonlinear Hall family, the investigation of third-order IAHE remains absent due to the lack of an appropriate theoretical approach, although the third-order extrinsic AHE has been studied within the framework of first- and second-order semiclassical theory. Herein, we generalize the semiclassical theory for Bloch electrons under the uniform electric field up to the third-order using wavepacket method and based on which we predict that the third-order IAHE can also occur in time-reversal broken systems. Same as the second-order IAHE, we find the band geometric quantity, the second-order field-dependent Berry curvature arising from the second-order field-induced positional shift, plays a pivotal role to observe this effect. Moreover, with symmetry analysis, we find that the third-order IAHE, as the leading contribution, is supported by 15 time-reversal broken 3D magnetic point groups (MPGs), corresponding to a wide class of antiferromagnetic (AFM) materials. Guided by the symmetry arguments, a two-band model is chosen to demonstrate the generalized theory. Furthermore, the generalized third-order semiclassical theory depends only on the properties of Bloch bands, implying that it can also be employed to explore the IAHE in realistic AFM materials, by combining with first-principles calculations.

cond-mat.mes-hall↗

arXiv2023

Quantum Fluctuation of the Quantum Geometric Tensor and its Manifestation as Intrinsic Hall Signatures in Time-Reversal Invariant Systems

In time-reversal invariant systems, all charge Hall effects predicted so far are extrinsic effects due to the dependence on the relaxation time. We explore intrinsic Hall signatures by studying quantum noise spectrum of the Hall current in time-reversal invariant systems, and discover intrinsic thermal Hall noises in both linear and nonlinear regimes. As the band geometric characteristics, quantum geometric tensor and Berry curvature play critical roles in various Hall effects, so are their quantum fluctuations. It is found that the thermal Hall noise in linear order of the electric field is purely intrinsic, and the second-order thermal Hall noise has both intrinsic and extrinsic contributions. In particular, the intrinsic part of the second-order thermal Hall noise is a manifestation of the quantum fluctuation of quantum geometric tensor, which widely exists as long as Berry curvature is nonzero. These intrinsic thermal Hall noises provide direct measurable means to band geometric information, including Berry curvature related quantities and quantum fluctuation of quantum geometric tensor.

cond-mat.mes-hall↗

arXiv2023

Four-vertex traces of finite sets

Let $[n]=X_1\cup X_2\cup X_3$ be a partition with $\lfloor\frac{n}{3}\rfloor \leq |X_i|\leq \lceil\frac{n}{3}\rceil$ and define $\mathcal{G}=\{G\subset [n]\colon |G\cap X_i|\leq 1, 1\leq i\leq 3\}$. It is easy to check that the trace $\mathcal{G}_{\mid Y}:=\{G\cap Y\colon G\in \mathcal{G}\}$ satisfies $|\mathcal{G}_{\mid Y}|\leq 12$ for all 4-sets $Y\subset [n]$. For $n\geq 25$ it is proven that whenever $\mathcal{F}\subset 2^{[n]}$ satisfies $|\mathcal{F}|>|\mathcal{G}|$ then $|\mathcal{F}_{\mid C}|\geq 13$ for some $C\subset [n]$, $|C|=4$. Several further results of a similar flavor are established as well.

math.CO↗