Searcharxiv⌕ Search

arXiv subjects

Yang Jiao

Publications and source records attributed to Yang Jiao.

At least 91 records · Page 5Linked to original sources
arXiv2023

Computational Design of Anisotropic Stealthy Hyperuniform Composites with Engineered Directional Scattering Properties

Disordered hyperuniform materials are an emerging class of exotic amorphous states of matter that endow them with singular physical properties. Here, we generalize the Fourier-space based numerical construction procedure for designing {\it isotropic} disordered hyperuniform two-phase heterogeneous materials (i.e., composites) developed by Chen and Torquato [Acta Mater. {\bf 142}, 152 (2018)] to {\it anisotropic} microstructures by explicitly incorporating the {\it vector-dependent} spectral density function ${\tilde χ}_{_V}({\bf k})$ of {\it arbitrary form} that is realizable. We demonstrate the utility of the procedure by generating a wide spectrum of {\it anisotropic} stealthy hyperuniform (SHU) microstructures with ${\tilde χ}_{_V}({\bf k}) = 0$ for ${\bf k} \in Ω$. We show how different exclusion-region shapes with various discrete symmetries and varying size affect the resulting statistically anisotropic microstructures as a function of the and phase volume fraction. We find that, among other properties, the directional hyperuniform behaviors imposed by the shape asymmetry (or anisotropy) of certain exclusion regions give rise to distinct anisotropic structures and degree of uniformity in the distribution of the phases on intermediate and large length scales along different directions. Moreover, while the anisotropic exclusion regions impose strong constraints on the {\it global} symmetry of the resulting media, they can still possess almost isotropic {\it local} structures. Our construction algorithm enables one to control the statistical anisotropy of composite microstructures which is crucial to engineering directional optical, transport and mechanical properties of two-phase composite media.

cond-mat.mtrl-sci↗
arXiv2023

The role of receptor uniformity in multivalent binding

Multivalency is prevalent in various biological systems and applications due to the superselectivity that arises from the cooperativity of multivalent binding. Traditionally, it was thought that weaker individual binding would improve the selectivity in multivalent targeting. Here using analytical mean field theory and Monte Carlo simulations, we discover that for receptors that are highly uniformly distributed, the highest selectivity occurs at an intermediate binding energy and can be significantly greater than the weak binding limit. This is caused by an exponential relationship between the bound fraction and receptor concentration, which is influenced by both the strength and combinatorial entropy of binding. Our findings not only provide new guidelines for the rational design of biosensors using multivalent nano-particles but also introduce a new perspective in understanding biological processes involving multivalency.

cond-mat.soft↗
arXiv2023

MSMDFusion: Fusing LiDAR and Camera at Multiple Scales with Multi-Depth Seeds for 3D Object Detection

Fusing LiDAR and camera information is essential for achieving accurate and reliable 3D object detection in autonomous driving systems. This is challenging due to the difficulty of combining multi-granularity geometric and semantic features from two drastically different modalities. Recent approaches aim at exploring the semantic densities of camera features through lifting points in 2D camera images (referred to as seeds) into 3D space, and then incorporate 2D semantics via cross-modal interaction or fusion techniques. However, depth information is under-investigated in these approaches when lifting points into 3D space, thus 2D semantics can not be reliably fused with 3D points. Moreover, their multi-modal fusion strategy, which is implemented as concatenation or attention, either can not effectively fuse 2D and 3D information or is unable to perform fine-grained interactions in the voxel space. To this end, we propose a novel framework with better utilization of the depth information and fine-grained cross-modal interaction between LiDAR and camera, which consists of two important components. First, a Multi-Depth Unprojection (MDU) method with depth-aware designs is used to enhance the depth quality of the lifted points at each interaction level. Second, a Gated Modality-Aware Convolution (GMA-Conv) block is applied to modulate voxels involved with the camera modality in a fine-grained manner and then aggregate multi-modal features into a unified space. Together they provide the detection head with more comprehensive features from LiDAR and camera. On the nuScenes test benchmark, our proposed method, abbreviated as MSMDFusion, achieves state-of-the-art 3D object detection results with 71.5% mAP and 74.0% NDS, and strong tracking results with 74.0% AMOTA without using test-time-augmentation and ensemble techniques. The code is available at https://github.com/SxJyJay/MSMDFusion.

cs.CV↗
arXiv2023

Spheroid Model for Molecular Packing in Crystalline Phase

Dense packing of particles has provided important models to study the structure of matter in various systems such as liquid, glassy and crystalline phase, etc. The simplest sphere packing models are able to represent and capture salient properties of the building blocks for covalent, metallic and ionic crystals; it however becomes insufficient to reflect the broken symmetry of the commonly anisotropic molecules in complex molecular crystals. Here we develop spheroid models with the minimal degree of anisotropy, which serve as a simple geometrical representation for a rich spectrum of molecules--including both isotropic and anisotropic, convex and concave ones--in crystalline phases. Our models are determined via an inverse packing approach: given a molecular crystal, an optimal spheroid model is constructed using a contact diagram, which depicts packing relationship between neighboring molecules within the crystal. The spheroid models are capable of accurately capturing the broken symmetry and characterizing the equivalent volume of molecules in the crystalline phases. Our model also allows to retrieve such molecular information from poor-quality crystal X-ray diffraction data that otherwise would be simply discarded.

cond-mat.soft↗
arXiv2023

Asynchronous Distributed Bilevel Optimization

Bilevel optimization plays an essential role in many machine learning tasks, ranging from hyperparameter optimization to meta-learning. Existing studies on bilevel optimization, however, focus on either centralized or synchronous distributed setting. The centralized bilevel optimization approaches require collecting massive amount of data to a single server, which inevitably incur significant communication expenses and may give rise to data privacy risks. Synchronous distributed bilevel optimization algorithms, on the other hand, often face the straggler problem and will immediately stop working if a few workers fail to respond. As a remedy, we propose Asynchronous Distributed Bilevel Optimization (ADBO) algorithm. The proposed ADBO can tackle bilevel optimization problems with both nonconvex upper-level and lower-level objective functions, and its convergence is theoretically guaranteed. Furthermore, it is revealed through theoretic analysis that the iteration complexity of ADBO to obtain the $ε$-stationary point is upper bounded by $\mathcal{O}(\frac{1}{ε^2})$. Thorough empirical studies on public datasets have been conducted to elucidate the effectiveness and efficiency of the proposed ADBO.

cs.LG↗
arXiv2023

Stock Market Prediction via Deep Learning Techniques: A Survey

Existing surveys on stock market prediction often focus on traditional machine learning methods instead of deep learning methods. This motivates us to provide a structured and comprehensive overview of the research on stock market prediction. We present four elaborated subtasks of stock market prediction and propose a novel taxonomy to summarize the state-of-the-art models based on deep neural networks. In addition, we also provide detailed statistics on the datasets and evaluation metrics commonly used in the stock market. Finally, we point out several future directions by sharing some new perspectives on stock market prediction.

q-fin.GN↗
arXiv2022

Distributed Distributionally Robust Optimization with Non-Convex Objectives

Distributionally Robust Optimization (DRO), which aims to find an optimal decision that minimizes the worst case cost over the ambiguity set of probability distribution, has been widely applied in diverse applications, e.g., network behavior analysis, risk management, etc. However, existing DRO techniques face three key challenges: 1) how to deal with the asynchronous updating in a distributed environment; 2) how to leverage the prior distribution effectively; 3) how to properly adjust the degree of robustness according to different scenarios. To this end, we propose an asynchronous distributed algorithm, named Asynchronous Single-looP alternatIve gRadient projEction (ASPIRE) algorithm with the itErative Active SEt method (EASE) to tackle the distributed distributionally robust optimization (DDRO) problem. Furthermore, a new uncertainty set, i.e., constrained D-norm uncertainty set, is developed to effectively leverage the prior distribution and flexibly control the degree of robustness. Finally, our theoretical analysis elucidates that the proposed algorithm is guaranteed to converge and the iteration complexity is also analyzed. Extensive empirical studies on real-world datasets demonstrate that the proposed method can not only achieve fast convergence, and remain robust against data heterogeneity as well as malicious attacks, but also tradeoff robustness with performance.

cs.LG↗
arXiv2022

CPMLHO:Hyperparameter Tuning via Cutting Plane and Mixed-Level Optimization

The hyperparameter optimization of neural network can be expressed as a bilevel optimization problem. The bilevel optimization is used to automatically update the hyperparameter, and the gradient of the hyperparameter is the approximate gradient based on the best response function. Finding the best response function is very time consuming. In this paper we propose CPMLHO, a new hyperparameter optimization method using cutting plane method and mixed-level objective function.The cutting plane is added to the inner layer to constrain the space of the response function. To obtain more accurate hypergradient,the mixed-level can flexibly adjust the loss function by using the loss of the training set and the verification set. Compared to existing methods, the experimental results show that our method can automatically update the hyperparameters in the training process, and can find more superior hyperparameters with higher accuracy and faster convergence.

cs.LG↗
arXiv2022

Design and Control of the "TransBoat": A Transformable Unmanned Surface Vehicle for Overwater Construction

This paper presents the TransBoat, a novel omnidirectional unmanned surface vehicle (USV) with a magnetbased docking system for overwater construction with wave disturbances. This is the first such USV that can build overwater structures by transporting modules. The TransBoat incorporates two features designed to reject wave disturbances. First, the TransBoat's expandable body structure can actively transform from a mono-hull into a multi-hull for stabilization in turbulent environments by extending its four outrigger hulls. Second, a real-time nonlinear model predictive control (NMPC) scheme is proposed for all shapes of the TransBoat to enhance its maneuverability and resist disturbance to its movement, based on a nonlinear dynamic model. An experimental approach is proposed to identify the parameters of the dynamic model, and a subsequent trajectory tracking test validates the dynamics, NMPC controller and system mobility. Further, docking experiments identify improved performance in the expanded form of the TransBoat compared with the contracted form, including an increased success rate (of ~ 10%) and reduced docking time (of ~ 40 s on average). Finally, a bridge construction test verifies our system design and the NMPC control method.

cs.RO↗
arXiv2022

Disordered Hyperuniform Solid State Materials

Disordered hyperuniform (DHU) states are recently discovered exotic states of condensed matter. DHU systems are similar to liquids or glasses in that they are statistically isotropic and lack conventional long-range translational and orientational order. On the other hand, they completely suppress normalized infinite-wavelength density fluctuations like crystals, and in this sense possess a hidden long-range order. Very recently, there are several exciting discoveries of disordered hyperuniformity in solid-state materials, including amorphous carbon nanotubes, amorphous 2D silica, amorphous graphene, defected transition metal dichalcogenides, defected pentagonal 2D materials, and medium/high-entropy alloys. It has been found the DHU states of these materials often possess a significantly lower energy than other disorder models, and can lead to unique electronic and thermal transport properties, which resulted from mechanisms distinct from those identified for their crystalline counterparts. For example, DHU states can enhance electronic transport in 2D amorphous silica; DHU medium/high-entropy alloys realize the Vegard's law, and possess enhanced electronic band gaps and thermal transport at low temperatures. These unique properties open up many promising potential device applications in optoelectronics and thermoelectrics. Here, we provide a focused review on these important new developments of hyperuniformity in solid-state materials, taking an applied and ``materials'' perspective, which complements the existing reviews on hyperuniformity in physical systems and photonic materials. Future directions and outlook are also provided, with a focus on the design and discovery of DHU quantum materials for quantum information science and engineering.

cond-mat.mtrl-sci↗
arXiv2022

Accurate non-empirical range-separated hybrid van der Waals density functional for complex molecular problems, solids, and surfaces

We introduce a new, general-purpose, range-separated hybrid van der Waals density \ph{functional, termed vdW-DF-ahbr,} within the non-empirical vdW-DF method [JPCM 32, 393001 (2020)]. It combines correlation from vdW-DF2 with a screened Fock exchange that is fixed by \ph{a new model of exchange effects} in the density-explicit vdW-DF2-b86r functional [PRB 89, 121103(R) (2014)]. The new vdW-DF2-ahbr prevents spurious exchange binding and has a small-density-gradient form set from many-body perturbation analysis. It is accurate for \ph{bulk as well as layered materials} and it systematically and significantly improves the performance of present vdW-DFs for molecular problems. Importantly, vdW-DF2-ahbr also outperforms present-standard (dispersion-corrected) range-separated hybrids on a broad collection of noncovalent-interaction benchmark sets, while at the same time successfully mitigating the density-driven errors that often affect the description of molecular transition states and isomerization calculations. vdW-DF2-ahbr furthermore improves on state of the art density functional theory approaches by 1) correctly predicting both the substrate structure and the site preference for CO adsorption on Pt(111), 2) outperforming existing non-empirical vdW-DFs for the description of CO$_2$ adsorption in both a functionalized and in a simple metal-organic framework, and 3) being highly accurate \ph{for the} set of base-pair interactions in a model of DNA assembly.

physics.chem-ph↗
arXiv2022

Exclusion Volumes of Convex Bodies in High Space Dimensions: Applications to Virial Coefficients and Continuum Percolation

Using the concepts of mixed volumes and quermassintegrals of convex geometry, we derive an exact formula for the exclusion volume for a general convex body that applies in any space dimension, including both the rotationally-averaged exclusion volume and with the same orientation. We show that the sphere minimizes the dimensionless exclusion volume $v_{ex}(K)/v(K)$ among all convex bodies, whether randomly oriented or uniformly oriented, for any $d$, where $v(K)$ is the volume of $K$. When the bodies have the same orientation, the simplex maximizes the dimensionless exclusion volume for any $d$ with a large-$d$ asymptotic scaling behavior of $2^{2d}/d^{3/2}$, which is to be contrasted with the scaling of $2^d$ for the sphere. We present explicit formulas for quermassintegrals for many nonspherical convex bodies as well as as well as lower-dimensional bodies. These results are utilized to determine the rotationally-averaged exclusion volume for these shapes for dimensions 2 through 12. While the sphere is the shape possessing the minimal dimensionless exclusion volume, among the convex bodies considered that are sufficiently compact, the simplex possesses the maximal dimensionless exclusion volume with a scaling behavior of $2^{1.6618\ldots d}$. We also determine the corresponding second virial coefficient $B_2(K)$ of the aforementioned hard hyperparticles and compute estimates of the continuum percolation threshold $η_c$ derived previously by the authors. We conjecture that overlapping spheres possess the maximal value of $η_c$ among all identical nonzero-volume convex overlapping bodies for $d \ge 2$, randomly or uniformly oriented, and that, among all identical, oriented nonzero-volume convex bodies, overlapping simplices have the minimal value of $η_c$ for $d\ge 2$.

cond-mat.stat-mech↗
arXiv2022

Disordered Hyperuniform Quasi-1D Materials

Carbon nanotubes are quasi-one-dimensional systems that possess superior transport, mechanical, optical, and chemical properties. In this work, we generalize the notion of disorder hyperuniformity, a recently discovered exotic state of matter with hidden long-range order, to quasi-one-dimensional materials. As a proof of concept, we then apply the generalized framework to quantify the density fluctuations in amorphous carbon nanotubes containing randomly distributed Stone-Wales defects. We demonstrate that all of these amorphous nanotubes are hyperuniform, i.e., the infinite-wavelength density fluctuations of these systems are completely suppressed, regardless of the diameter, rolling axis, number of rolling sheets, and defect fraction of the nanotubes. We find that these amorphous nanotubes are energetically more stable than nanotubes with periodically distributed Stone-Wales defects. Moreover, certain semiconducting defect-free carbon nanotubes become metallic as sufficiently large amounts of defects are randomly introduced. This structural study of amorphous nanotubes strengthens our fundamental understanding of these systems, and suggests possible exotic physical properties, as endowed by their disordered hyperuniformity. Our findings also shed light on the effect of dimensionality reduction on the hyperuniformity property of materials.

cond-mat.stat-mech↗
arXiv2022

MORE: Multi-Order RElation Mining for Dense Captioning in 3D Scenes

3D dense captioning is a recently-proposed novel task, where point clouds contain more geometric information than the 2D counterpart. However, it is also more challenging due to the higher complexity and wider variety of inter-object relations contained in point clouds. Existing methods only treat such relations as by-products of object feature learning in graphs without specifically encoding them, which leads to sub-optimal results. In this paper, aiming at improving 3D dense captioning via capturing and utilizing the complex relations in the 3D scene, we propose MORE, a Multi-Order RElation mining model, to support generating more descriptive and comprehensive captions. Technically, our MORE encodes object relations in a progressive manner since complex relations can be deduced from a limited number of basic ones. We first devise a novel Spatial Layout Graph Convolution (SLGC), which semantically encodes several first-order relations as edges of a graph constructed over 3D object proposals. Next, from the resulting graph, we further extract multiple triplets which encapsulate basic first-order relations as the basic unit, and construct several Object-centric Triplet Attention Graphs (OTAG) to infer multi-order relations for every target object. The updated node features from OTAG are aggregated and fed into the caption decoder to provide abundant relational cues, so that captions including diverse relations with context objects can be generated. Extensive experiments on the Scan2Cap dataset prove the effectiveness of our proposed MORE and its components, and we also outperform the current state-of-the-art method. Our code is available at https://github.com/SxJyJay/MORE.

cs.CV↗
arXiv2022

Task-aware Similarity Learning for Event-triggered Time Series

Time series analysis has achieved great success in diverse applications such as network security, environmental monitoring, and medical informatics. Learning similarities among different time series is a crucial problem since it serves as the foundation for downstream analysis such as clustering and anomaly detection. It often remains unclear what kind of distance metric is suitable for similarity learning due to the complex temporal dynamics of the time series generated from event-triggered sensing, which is common in diverse applications, including automated driving, interactive healthcare, and smart home automation. The overarching goal of this paper is to develop an unsupervised learning framework that is capable of learning task-aware similarities among unlabeled event-triggered time series. From the machine learning vantage point, the proposed framework harnesses the power of both hierarchical multi-scale sequence autoencoders and Gaussian Mixture Model (GMM) to effectively learn the low-dimensional representations from the time series. Finally, the obtained similarity measure can be easily visualized for explaining. The proposed framework aspires to offer a stepping stone that gives rise to a systematic approach to model and learn similarities among a multitude of event-triggered time series. Through extensive qualitative and quantitative experiments, it is revealed that the proposed method outperforms state-of-the-art methods considerably.

cs.LG↗
arXiv2022

vdW-DF-ahcx: a range-separated van der Waals density functional hybrid

Hybrid density functionals replace a fraction of an underlying generalized-gradient approximation (GGA) exchange description with a Fock-exchange component. Range-separated hybrids (RSHs) also effectively screen the Fock-exchange component and thus open the door for characterizations of metals and adsorption at metal surfaces. The RSHs are traditionally based on a robust GGA, such as PBE [PRL $\textbf{77}$, 3865 (1996)], for example, as implemented in the HSE design [JPC $\textbf{118}$, 8207 (2003)]. Here we define a RSH extension to the van der Waals density functional (vdW-DF) method [ROPP $\textbf{78}$, 066501 (2015)], launching vdW-DF-ahcx. We use an analytical-hole (AH) framework [JCP $\textbf{128}$, 194105 (2008)] to characterize the GGA-type exchange in the vdW-DF-cx version [PRB $\textbf{89}$, 075148 (2014)], isolate the short-ranged exchange component, and define the new RSH. We find that the performance vdW-DF-ahcx compares favorably to (dispersion-corrected) HSE for descriptions of bulk (broad molecular) properties. We also find that it provides accurate descriptions of noble-metal surface properties, including CO adsorption.

cond-mat.mtrl-sci↗
arXiv2021

Screening nature of the van der Waals density functional method: A review and analysis of the many-body physics foundation

We review the screening nature and many-body physics foundation of the van der Waals density functional (vdW-DF) method, a systematic approach to construct truly nonlocal exchange-correlation energy density functionals. To that end we define and focus on a class of consistent vdW-DF versions that adhere to the Lindhard screening logic of the full method formulation. The consistent-exchange vdW-DF-cx version and its spin extension represent the first examples of this class; In general, consistent vdW-DFs reflect a concerted expansion of a formal recast of the adiabatic-connection formula, an exponential summation of contributions to the local-field response, and the Dyson equation. We argue that the screening emphasis is essential because the exchange-correlation energy reflects an effective electrodynamics set by a long-range interaction. Two consequences are that 1) there are, in principle, no wiggle room in how one balances exchange and correlation, for example, in vdW-DF-cx, and that 2) consistent vdW-DFs have a formal structure that allows them to incorporate vertex-correction effects, at least in the case of levels that experience recoil-less interactions (for example, near the Fermi surface). We explore the extent to which the strictly nonempirical vdW-DF-cx formulation can serve as a systematic extension of the constraint-based semilocal functionals. For validation, we provide a complete survey of vdW-DF-cx performance for broad molecular processes and comparing to the quantum-chemistry calculations that are summarized in that paper. We also provide new vdW-DF-cx results for metal surface energies and work functions that we compare to experiment. Finally, we use the screening insight to separate the vdW-DF nonlocal-correlation term and present tools to compute and map the binding signatures.

cond-mat.mtrl-sci↗