DEEP NEURAL NETWORKS COMBINING MULTI-TASK LEARNING FOR SOLVING DELAY INTEGRO-DIFFERENTIAL EQUATIONS

Deep neural networks(DNNs)are effective in solving both forward and inverse problems for nonlinear partial differential equations(PDEs).However,conventional DNNs are not effective in handling problems such as delay differential equations(DDEs)and delay integrodifferential equations(DIDEs)with constant delays,primarily due to their low regularity at delayinduced breaking points.In this paper,a DNN method that combines multi-task learning(MTL)which is proposed to solve both the forward and inverse problems of DIDEs.The core idea of this approach is to divide the original equation into multiple tasks based on the delay,using auxiliary outputs to represent the integral terms,followed by the use of MTL to seamlessly incorporate the properties at the breaking points into the loss function.Furthermore,given the increased training dificulty associated with multiple tasks and outputs,we employ a sequential training scheme to reduce training complexity and provide reference solutions for subsequent tasks.This approach significantly enhances the approximation accuracy of solving DIDEs with DNNs,as demonstrated by comparisons with traditional DNN methods.We validate the effectiveness of this method through several numerical experiments,test various parameter sharing structures in MTL and compare the testing results of these structures.Finally,this method is implemented to solve the inverse problem of nonlinear DIDE and the results show that the unknown parameters of DIDE can be discovered with sparse or noisy data.

Concurrent Two-Scale Topology Optimization of Thermoelastic Structures Using a M-VCUT Level Set Based Model of Microstructures

By analyzing the results of compliance minimization of thermoelastic structures,we observed that microstructures play an important role in this optimization problem.Then,we propose to use a multiple variable cutting(M-VCUT)level set-based model of microstructures to solve the concurrent two-scale topology optimization of thermoelastic structures.A microstructure is obtained by combining multiple virtual microstructures that are derived respectively from multiple microstructure prototypes,thus giving more diversity of microstructure and more flexibility in design optimization.The effective mechanical properties of microstructures are computed in an off-line phase by using the homogenization method,and then a mapping relationship between the design variables and the effective properties is established,which gives a data-driven model of microstructure.In the online phase,the data-driven model is used in the finite element analysis to improve the computational efficiency.The compliance minimization problem is considered,and the results of numerical examples prove that the proposed method is effective.

Evolutionary game dynamics of combining two different aspiration-driven update rules in structured populations

In evolutionary games,most studies on finite populations have focused on a single updating mechanism.However,given the differences in individual cognition,individuals may change their strategies according to different updating mechanisms.For this reason,we consider two different aspiration-driven updating mechanisms in structured populations:satisfied-stay unsatisfied shift(SSUS)and satisfied-cooperate unsatisfied defect(SCUD).To simulate the game player’s learning process,this paper improves the particle swarm optimization algorithm,which will be used to simulate the game player’s strategy selection,i.e.,population particle swarm optimization(PPSO)algorithms.We find that in the prisoner’s dilemma,the conditions that SSUS facilitates the evolution of cooperation do not enable cooperation to emerge.In contrast,SCUD conditions that promote the evolution of cooperation enable cooperation to emerge.In addition,the invasion of SCUD individuals helps promote cooperation among SSUS individuals.Simulated by the PPSO algorithm,the theoretical approximation results are found to be consistent with the trend of change in the simulation results.

First‑principles study on electronic structure,optical and magnetic properties of rare earth elements X(X=Sc,Y,La,Ce,Eu)doped with two‑dimensional GaSe

The electronic structure,magnetic,and optical properties of two-dimensional(2D)GaSe doped with rare earth elements X(X=Sc,Y,La,Ce,Eu)were calculated using the first-principles plane wave method based on den-sity functional theory.The results show that intrinsic 2D GaSe is a p-type nonmagnetic semiconductor with an indi-rect bandgap of 2.6611 eV.The spin-up and spin-down channels of Sc-,Y-,and La-doped 2D GaSe are symmetric,they are non-magnetic semiconductors.The magnetic moments of Ce-and Eu-doped 2D GaSe are 0.908μ_(B)and 7.163μ_(B),which are magnetic semiconductors.Impurity energy levels appear in both spin-up and spin-down chan-nels of Eu-doped 2D GaSe,which enhances the probability of electron transition.Compared with intrinsic 2D GaSe,the static dielectric constant of the doped 2D GaSe increases,and the polarization ability is strengthened.The ab-sorption spectrum of the doped 2D GaSe shifts in the low-energy direction,and the red-shift phenomenon occurs,which extends the absorption spectral range.The optical reflection coefficient of the doped 2D GaSe is improved in the low energy region,and the improvement of Eu-doped 2D GaSe is the most obvious.

Graph Convolutional Networks Embedding Textual Structure Information for Relation Extraction

Deep neural network-based relational extraction research has made significant progress in recent years,andit provides data support for many natural language processing downstream tasks such as building knowledgegraph,sentiment analysis and question-answering systems.However,previous studies ignored much unusedstructural information in sentences that could enhance the performance of the relation extraction task.Moreover,most existing dependency-based models utilize self-attention to distinguish the importance of context,whichhardly deals withmultiple-structure information.To efficiently leverage multiple structure information,this paperproposes a dynamic structure attention mechanism model based on textual structure information,which deeplyintegrates word embedding,named entity recognition labels,part of speech,dependency tree and dependency typeinto a graph convolutional network.Specifically,our model extracts text features of different structures from theinput sentence.Textual Structure information Graph Convolutional Networks employs the dynamic structureattention mechanism to learn multi-structure attention,effectively distinguishing important contextual features invarious structural information.In addition,multi-structure weights are carefully designed as amergingmechanismin the different structure attention to dynamically adjust the final attention.This paper combines these featuresand trains a graph convolutional network for relation extraction.We experiment on supervised relation extractiondatasets including SemEval 2010 Task 8,TACRED,TACREV,and Re-TACED,the result significantly outperformsthe previous.

3D Road Network Modeling and Road Structure Recognition in Internet of Vehicles

Internet of Vehicles (IoV) is a new system that enables individual vehicles to connect with nearby vehicles,people, transportation infrastructure, and networks, thereby realizing amore intelligent and efficient transportationsystem. The movement of vehicles and the three-dimensional (3D) nature of the road network cause the topologicalstructure of IoV to have the high space and time complexity.Network modeling and structure recognition for 3Droads can benefit the description of topological changes for IoV. This paper proposes a 3Dgeneral roadmodel basedon discrete points of roads obtained from GIS. First, the constraints imposed by 3D roads on moving vehicles areanalyzed. Then the effects of road curvature radius (Ra), longitudinal slope (Slo), and length (Len) on speed andacceleration are studied. Finally, a general 3D road network model based on road section features is established.This paper also presents intersection and road section recognition methods based on the structural features ofthe 3D road network model and the road features. Real GIS data from a specific region of Beijing is adopted tocreate the simulation scenario, and the simulation results validate the general 3D road network model and therecognitionmethod. Therefore, thiswork makes contributions to the field of intelligent transportation by providinga comprehensive approach tomodeling the 3Droad network and its topological changes in achieving efficient trafficflowand improved road safety.

Synthesis,crystal structures,and antibacterial activities of two zinc(Ⅱ) complexes bearing 5⁃phenyl⁃1H⁃pyrazole group

A novel compound(H_(2)L)SCN(5⁃methyl⁃3⁃phenyl⁃2H⁃pyrazol⁃1⁃ium thiocyanate)has been obtained by the reaction of thiosemicarbazide with benzoylacetone in ethanol.Two zinccomplexes[Zn(HL)_(2)(NCS)(CH_(3)COO)](1)and[Zn_(2)(L)_(2)(HL)_(2)(NCS)_(2)]_(2)·2CH_(3)OH(2)have been synthesized by the coordination reactions of Zn(OAc)_(2)·2H_(2)O or ZnCl_(2)with(H_(2)L)SCN under reflux conditions.Elemental analyses and single⁃crystal X⁃ray diffraction have con⁃firmed the structures of the synthesized compounds.The(H_(2)L)SCN ligand and complex 1 pertain to the triclinic sys⁃tem with space group P1,while complex 2 belongs to the monoclinic system with space group P2_(1)/n.Additionally,the antibacterial activities of the compounds were evaluated in vitro using the agar diffusion method against the bac⁃terial strains(Candida albicans,Staphylococcus aureus,and Escherichia coli).The results showed that the ligand exhibited relatively good antibacterial activities against the bacteria,and the complexes possessed stronger antibac⁃terial activities against the same bacteria than the free ligand.CCDC:2190252,(H2L)SCN;2190253,1;2190256,2.

Control of light-matter interactions in two-dimensional materials with nanoparticle-on-mirror structures

Light–matter interactions in two-dimensional(2D)materials have been the focus of research since the discovery of graphene.The light–matter interaction length in 2D materials is,however,much shorter than that in bulk materials owing to the atomic nature of 2D materials.Plasmonic nanostructures are usually integrated with 2D materials to enhance the light–matter interactions,offering great opportunities for both fundamental research and technological applications.Nanoparticle-on-mirror(NPo M)structures with extremely confined optical fields are highly desired in this aspect.In addition,2D materials provide a good platform for the study of plasmonic fields with subnanometer resolution and quantum plasmonics down to the characteristic length scale of a single atom.A focused and up-to-date review article is highly desired for a timely summary of the progress in this rapidly growing field and to encourage more research efforts in this direction.In this review,we will first introduce the basic concepts of plasmonic modes in NPo M structures.Interactions between plasmons and quasi-particles in 2D materials,e.g.,excitons and phonons,from weak to strong coupling and potential applications will then be described in detail.Related phenomena in subnanometer metallic gaps separated by 2D materials,such as quantum tunneling,will also be touched.We will finally discuss phenomena and physical processes that have not been understood clearly and provide an outlook for future research.We believe that the hybrid systems of2D materials and NPo M structures will be a promising research field in the future.

Syntheses,crystal structures,and quantum chemistry calculation of two Ni(Ⅱ)coordination polymers

Two new coordination polymers,[Ni(Hpdc)(bib)(H_(2)O)]_(n)(1)and{[Ni(bib)_(3)](ClO_(4))_(2)}_(n)(2),were prepared by mixing Ni^(2+),3,5⁃pyrazoledicarboxylic acid(H3pdc)/p⁃nitrobenzoic acid and 1,4⁃bis(imidazol⁃1⁃ylmethyl)butane(bib)by a hydrothermal method,respectively.X⁃ray crystallography reveals a 2D network constructed by six⁃coordinated Ni(Ⅱ)centers,bib,and Hpdc2-ligands in complex 1,while a 2D network is built by Ni(Ⅱ)and bib ligands in 2.Furthermore,the quantum⁃chemical calculations have been performed on‘molecular fragments’extracted from the crystal structure of 1 using the PBE0/LANL2DZ method in Gaussian 16 and the VASP program.CCDC:2343794,1;2343798,2.

A three-dimensional co-continuous network structure polymer electrolyte with efficient ion transport channels enabling ultralong-life all solid-state lithium metal batteries

Solid polymer electrolytes(SPEs)have emerged as one of the most promising candidates for the construction of solid-state lithium batteries due to their excellent flexibility,scalability,and interface compatibility with electrodes.Herein,a novel all-solid polymer electrolyte(PPLCE)was fabricated by the copolymer network of liquid crystalline monomers and poly(ethylene glycol)dimethacrylate(PEGDMA)acts as a structural frame,combined with poly(ethylene glycol)diglycidyl ether short chain interspersed serving as mobile ion transport entities.The preparaed PPLCEs exhibit excellent mechanical property and out-standing electrochemical performances,which is attributed to their unique three-dimensional cocontinuous structure,characterized by a cross-linked semi-interpenetrating network and an ionic liquid phase,resulting in a distinctive nanostructure with short-range order and long-range disorder.Remarkably,the addition of PEGDMA is proved to be critical to the comprehensive performance of the PPLCEs,which effectively modulates the microscopic morphology of polymer networks and improves the mechanical properties as well as cycling stability of the solid electrolyte.When used in a lithiumion symmetrical battery configuration,the 6 wt%-PPLCE exhibites super stability,sustaining operation for over 2000 h at 30 C,with minimal and consistent overpotential of 50 mV.The resulting Li|PPLCE|LFP solid-state battery demonstrates high discharge specific capacities of 160.9 and 120.1 mA h g^(-1)at current densities of 0.2 and 1 C,respectively.Even after more than 300 cycles at a current density of 0.2 C,it retaines an impressive 73.5%capacity.Moreover,it displayes stable cycling for over 180 cycles at a high current density of 0.5C.The super cycle stability may promote the application for ultralong-life all solid-state lithium metal batteries.

Two novel Cu(Ⅰ)complexes based on diimide and bisphosphine ligands:synthesis,crystal structure and spectroscopic properties

Two Cu(Ⅰ)complexes[Cu(Bphen)(dppBz)]ClO_(4)·2CH_(3)OH(1)and[Cu_(2)(Bphen)_(2)(dpppda)]BF_(4)(2){Bphen=4,7-diphenyl-1,10-phenanthroline,dppBz=1,2-Bis(diphenylphosphino)benzene,dpppda=N1,N1,N4,N4-tetrakis[(diphenylphosphino)methyl]-1,4-benzenediamin}were synthesized using a one-pot method.X-ray crystallography was used to elucidate their crystal structures and photophysical properties.A series of characterization tests including elemental analysis,NMR,FT-IR,UV-Vis absorption spectroscopy,fluorescence spectroscopy,thermal gravimetric analysis and terahertz time-domain spectroscopy(THz-TDS)were used to further investigate their properties.The results show that complex 1 structure is mononuclear containing two solvent molecules per unit cell,while complex 2 structure is binuclear containing two metal centers per unit cell.According to photophysical properties and density functional theory(DFT)calculations,their luminescence properties can be attributed to metal-to-ligand charge transfer(MLCT).Both complexes have a unique stability,which is confirmed by thermal gravimetric analysis.

Topological Structure-Modulated Collagen Carbon as Two-in-One Energy Storage Configuration toward Ultrahigh Power and Energy Density

Efficient energy storage devices with suitable electrode materials,that integrate high power and high energy,are the crucial requisites of the renewable power source,which have unwrapped new possibilities in the sustainable development of energy and the environment.Herein,a facile collagen microstructure modulation strategy is proposed to construct a nitrogen/oxygen dual-doped hierarchically porous carbon fiber with ultrahigh specific surface area(2788 m^(2)g^(-1))and large pore volume(4.56 cm^(3)g^(-1))via local microfibrous breakage/disassembly of natural structured proteins.Combining operando spectroscopy and density functional theory unveil that the dual-heteroatom doping could effectively regulate the electronic structure of carbon atom framework with enhanced electric conductivity and electronegativity as well as decreased diffusion resistance in favor of rapid pseudocapacitive-dominated Li^(+)-storage(353 mAh g^(-1)at 10 A g^(-1)).Theoretical calculations reveal that the tailored micro-/mesoporous structures favor the rapid charge transfer and ion storage,synergistically realizing high capacity and superior rate performance for NPCF-H cathode(75.0 mAh g^(-1)at 30 A g^(-1)).The assembled device with NPCF-H as both anode and cathode achieves extremely high energy density(200 Wh kg^(-1))with maximum power density(42600 W kg^(-1))and ultralong lifespan(80%capacity retention over 10000 cycles).

Analysis of Urban Agglomeration Network Structure Based on Baidu Migration Data: A Case Study of the Guangdong-Hong Kong-Macao Greater Bay Urban Agglomeration

The inter-city linkage heat data provided by Baidu Migration is employed as a characterization of inter-city linkages in order to facilitate the study of the network linkage characteristics and hierarchical structure of urban agglomeration in the Greater Bay Area through the use of social network analysis method.This is the inaugural application of big data based on location services in the study of urban agglomeration network structure,which represents a novel research perspective on this topic.The study reveals that the density of network linkages in the Greater Bay Area urban agglomeration has reached 100%,indicating a mature network-like spatial structure.This structure has given rise to three distinct communities:Shenzhen-Dongguan-Huizhou,Guangzhou-Foshan-Zhaoqing,and Zhuhai-Zhongshan-Jiangmen.Additionally,cities within the Greater Bay Area urban agglomeration play different roles,suggesting that varying development strategies may be necessary to achieve staggered development.The study demonstrates that large datasets represented by LBS can offer novel insights and methodologies for the examination of urban agglomeration network structures,contingent on the appropriate mining and processing of the data.

Redefinable neural network for structured light array

Neural networks have provided faster and more straightforward solutions for laser modulation.However,their effectiveness when facing diverse structured lights and various output resolutions remains vulnerable because of the specialized end-to-end training and static model.Here,we propose a redefinable neural network(RediNet),realizing customized modulation on diverse structured light arrays through a single general approach.The network input format features a redefinable dimension designation,which ensures RediNet wide applicability and removes the burden of processing pixel-wise light distributions.The prowess of originally generating arbitrary-resolution holograms with a fixed network is first demonstrated.The versatility is showcased in the generation of 2D/3D foci arrays,Bessel and Airy beam arrays,(perfect)vortex beam arrays,and even snowflake-intensity arrays with arbitrarily built phase functions.A standout application is producing multichannel compound vortex beams,where RediNet empowers a spatial light modulator(SLM)to offer comprehensive multiplexing functionalities for free-space optical communication.Moreover,RediNet has the hitherto highest efficiency,only consuming 12 ms(faster than the mainstream SLM framerate of 60 Hz)for a 1000^(2)-resolution holograph,which is critical in real-time required scenarios.Considering the fine resolution,high speed,and unprecedented universality,RediNet can serve extensive applications,such as next-generation optical communication,parallel laser direct writing,and optical traps.

Transition of plasticity and fracture mode of Zr-Al-Ni-Cu bulk metallic glasses with network structures

Effect of network structure on plasticity and fracture mode of Zr？Al？Ni？Cu bulk metallic glasses （BMGs） was investigated. The microstructures of transversal and longitudinal sections were exposed by chemical etching and observed by scanning electron microscopy （SEM）. The mechanical properties were examined by room-temperature uniaxial compression test. The results show that both plasticity and fracture mode are significantly affected by the network structure and the alteration occurs when the size of the network structure reaches up to a critical value. When the cell size （dc） of the network structure is ~3μm, Zr-based BMGs characterize in plasticity that decreases with increasingdc. The fracture mode gradually transforms from single 45° shear fracture to double 45° shear fracture and then cleavage fracture with increasingdc. In addition, the mechanisms of the transition of the plasticity and the fracture mode for these Zr-based BMGs are also discussed.

Analysis on the Composition and Structure of Branches of Two Kinds of Tree Shapes in Korla Fragrant Pear

ObjectiveThe thesis aims at investigating the distribution and structural characteristics of various branches in canopy of Korla fragrant pear. MethodStatistic work and analysis were conducted on the numbers and distribution characteristics of various branches in each cubic lattice by using the canopy cellular method. ResultThe results showed that： The total number of scaffold branches of evacuation layered tree shape was 97, which mainly distributed in the lower layer and middle part of the canopy; the total number of scaffold branches of open-center tree shape was 94, which mainly distributed in the lower layer and middle part of the canopy. The total number of annual branches of evacuation layered tree shape was 3 920, which mainly distributed in the middle layer and outer part of the canopy; and the total number of annual branches of the open-center tree shape was 3 183, which mainly distributed in middle layer and outer part of the canopy. The total number of perennial branches of evacuation layered tree shape was 2 184, which mainly distributed in lower layer and outer part of the canopy; the total number of perennial branches of open-center tree shape was 1 444, which mainly distributed in middle layer and outer part of the canopy. ConclusionThe total number and the distribution positions of scaffold branches in the canopy of each tree shape were basically the same. The total numbers of annual branches of the two kinds of tree shapes were different, but the distribution positions were basically the same. The total numbers and the distribution positions of perennial branches in the canopy of the two kinds of tree shapes were different.

STRUCTURE OPTIMIZATION STRATEGY OF NORMALIZED RBF NETWORKS

Aimed at studying normali zed radial basis function network (NRBFN), this paper introduces the subtractiv e clustering based on a mountain function to construct the initial structure of NR BFN, adopts singular value decomposition (SVD) to analyze the relationship betwe en neural nodes of the hidden layer and singular values, cumulative contribution ratio, index vector, and optimizes the structure of NRBFN. Finally, simulation and performance comparison show that the algorithm is feasible and effective.

EXPERIMENTAL STUDY ON COHERENT VORTEX STRUCTURES IN DIFFERENTIALLY ROTATING QUASI TWO DIMENSIONAL ZONAL FLOW

An experimental system for forming a rotating paraboloid shaped shallow water with a free surface was conducted to study coherent vortex structures in a differentially rotating quasi two dimensional zonal flow.Flow visualization and laser light scattering techniques were used to obtain the information of spatial flow patterns.Experimental results show that the coexistence of Coriolis effect and strong shear in latitudinal zones may lead to formation of coherent vortices.Power spectra analysis and photographs which were taken in a reference frame rotating with the observed vortices also justified the emergence,drift and evolution of persistent vortices on the large scale.Locked vortex state manifests the cyclone and anticyclone asymmetry.

Regulatable Orthotropic 3D Hybrid Continuous Carbon Networks for Efficient Bi-Directional Thermal Conduction

Vertically oriented carbon structures constructed from low-dimen-sional carbon materials are ideal frameworks for high-performance thermal inter-face materials(TIMs).However,improving the interfacial heat-transfer efficiency of vertically oriented carbon structures is a challenging task.Herein,an orthotropic three-dimensional(3D)hybrid carbon network(VSCG)is fabricated by depositing vertically aligned carbon nanotubes(VACNTs)on the surface of a horizontally oriented graphene film(HOGF).The interfacial interaction between the VACNTs and HOGF is then optimized through an annealing strategy.After regulating the orientation structure of the VACNTs and filling the VSCG with polydimethylsi-loxane(PDMS),VSCG/PDMS composites with excellent 3D thermal conductive properties are obtained.The highest in-plane and through-plane thermal conduc-tivities of the composites are 113.61 and 24.37 W m^(-1)K^(-1),respectively.The high contact area of HOGF and good compressibility of VACNTs imbue the VSCG/PDMS composite with low thermal resistance.In addition,the interfacial heat-transfer efficiency of VSCG/PDMS composite in the TIM performance was improved by 71.3%compared to that of a state-of-the-art thermal pad.This new structural design can potentially realize high-performance TIMs that meet the need for high thermal conductivity and low contact thermal resistance in interfacial heat-transfer processes.

Robot-Oriented 6G Satellite-UAV Networks: Requirements, Paradigm Shifts, and Case Studies

Networked robots can perceive their surroundings, interact with each other or humans,and make decisions to accomplish specified tasks in remote/hazardous/complex environments. Satelliteunmanned aerial vehicle(UAV) networks can support such robots by providing on-demand communication services. However, under traditional open-loop communication paradigm, the network resources are usually divided into user-wise mostly-independent links,via ignoring the task-level dependency of robot collaboration. Thus, it is imperative to develop a new communication paradigm, taking into account the highlevel content and values behind, to facilitate multirobot operation. Inspired by Wiener’s Cybernetics theory, this article explores a closed-loop communication paradigm for the robot-oriented satellite-UAV network. This paradigm turns to handle group-wise structured links, so as to allocate resources in a taskoriented manner. It could also exploit the mobility of robots to liberate the network from full coverage,enabling new orchestration between network serving and positive mobility control of robots. Moreover,the integration of sensing, communications, computing and control would enlarge the benefit of this new paradigm. We present a case study for joint mobile edge computing(MEC) offloading and mobility control of robots, and finally outline potential challenges and open issues.