Heterogeneous graphs contain multiple types of entities and relations,which are capable of modeling complex interactions.Embedding on heterogeneous graphs has become an essential tool for analyzing and understanding such graphs.Although these meticulously designed methods make progress,they are limited by model design and computational resources,making it difficult to scale to large-scale heterogeneous graph data and hindering the application and promotion of these methods.In this paper,we propose Restage,a relation structure-aware hierarchical heterogeneous graph embedding framework.Under this framework,embedding only a smaller-scale graph with existing graph representation learning methods is sufficient to obtain node representations on the original heterogeneous graph.We consider two types of relation structures in heterogeneous graphs:interaction relations and affiliation relations.Firstly,we design a relation structure-aware coarsening method to successively coarsen the original graph to the top-level layer,resulting in a smaller-scale graph.Secondly,we allow any unsupervised representation learning methods to obtain node embeddings on the top-level graph.Finally,we design a relation structure-aware refinement method to successively refine the node embeddings from the top-level graph back to the original graph,obtaining node embeddings on the original graph.Experimental results on three public heterogeneous graph datasets demonstrate the enhanced scalability of representation learning methods by the proposed Restage.On another large-scale graph,the speed of existing representation learning methods is increased by up to eighteen times at most.
In the domain of knowledge graph embedding,conventional approaches typically transform entities and relations into continuous vector spaces.However,parameter efficiency becomes increasingly crucial when dealing with large-scale knowledge graphs that contain vast numbers of entities and relations.In particular,resource-intensive embeddings often lead to increased computational costs,and may limit scalability and adaptability in practical environ-ments,such as in low-resource settings or real-world applications.This paper explores an approach to knowledge graph representation learning that leverages small,reserved entities and relation sets for parameter-efficient embedding.We introduce a hierarchical attention network designed to refine and maximize the representational quality of embeddings by selectively focusing on these reserved sets,thereby reducing model complexity.Empirical assessments validate that our model achieves high performance on the benchmark dataset with fewer parameters and smaller embedding dimensions.The ablation studies further highlight the impact and contribution of each component in the proposed hierarchical attention structure.
Zhen-Yu ChenFeng-Chi LiuXin WangCheng-Hsiung LeeChing-Sheng Lin
Exo-atmospheric vehicles are constrained by limited maneuverability,which leads to the contradiction between evasive maneuver and precision strike.To address the problem of Integrated Evasion and Impact(IEI)decision under multi-constraint conditions,a hierarchical intelligent decision-making method based on Deep Reinforcement Learning(DRL)was proposed.First,an intelligent decision-making framework of“DRL evasion decision”+“impact prediction guidance decision”was established:it takes the impact point deviation correction ability as the constraint and the maximum miss distance as the objective,and effectively solves the problem of poor decisionmaking effect caused by the large IEI decision space.Second,to solve the sparse reward problem faced by evasion decision-making,a hierarchical decision-making method consisting of maneuver timing decision and maneuver duration decision was proposed,and the corresponding Markov Decision Process(MDP)was designed.A detailed simulation experiment was designed to analyze the advantages and computational complexity of the proposed method.Simulation results show that the proposed model has good performance and low computational resource requirement.The minimum miss distance is 21.3 m under the condition of guaranteeing the impact point accuracy,and the single decision-making time is 4.086 ms on an STM32F407 single-chip microcomputer,which has engineering application value.
Leliang RENWeilin GUOYong XIANZhenyu LIUDaqiao ZHANGShaopeng LI
The increase in the utilization of infrared heat detection technology in military applications necessitates research on composites with improved thermal transmission performance and microwave absorption capabilities.This study satisfactorily fabricated a series of MoS_(2)/BN-xyz composites(which were characterized by the weight ratio of MoS_(2)to BN,denoted by xy:z)through chemical vapor depos-ition,which resulted in their improved thermal stability and thermal transmission performance.The results show that the remaining mass of MoS_(2)/BN-101 was as high as 69.25wt%at 800℃under air atmosphere,and a temperature difference of 31.7℃was maintained between the surface temperature and the heating source at a heating temperature of 200℃.Furthermore,MoS_(2)/BN-301 exhibited an im-pressive minimum reflection loss value of-32.21 dB at 4.0 mm and a wide effective attenuation bandwidth ranging from 9.32 to 18.00 GHz(8.68 GHz).Therefore,these simplified synthesized MoS_(2)/BN-xyz composites demonstrate great potential as highly efficient con-tenders for the enhancement of microwave absorption performance and thermal conductance.
Managing high-flux waste heat with controllable device working temperature is becoming challenging and critical for the artificial intelligence,communications,electric vehicles,defense and aerospace sectors.Spray cooling,which combines forced convection with phase-change latent heat of working fluids,is promising for high flux heat dissipation.Most of the previous studies on spray cooling enhancement adopted high spray flow rates to strengthen forced convection for high critical heat flux(CHF),leading to a low heat transfer coefficient(HTC).Micro/nanostructured surfaces can enhance boiling,but bubbles inside the structures tend to form a vapor blanket,which can deteriorate heat transfer.This work demonstrates simultaneous enhancement of CHF and HTC in spray cooling by improving both evaporation and liquid film boiling on three-dimensional(3D)ordered hierarchical micro/nano-structured surface.The hierarchical micro/nano-structured surface is designed to coordinate the transport of spray droplets,capillary liquid films,and boiling bubbles to enhance spray cooling performance.Boiling inversion where superheat decreases with increasing heat flux is observed,leading to an ultra-high HTC due to the simultaneous promotion of bubble nucleation and evaporation.Unprecedented CHF is obtained by overcoming the liquid–vapor counterflow,i.e.,synergistically facilitating bubble escape and liquid permeation.A record-breaking heat transfer performance of spray cooling is achieved with a maximum heat flux of1273 W/cm^(2)and an HTC of 443.7 kW/(m^(2)K)over a 1 cm^(2)heating area.?2024 Science China Press.Published by Elsevier B.V.and Science China Press.All rights are reserved,including those for text and data mining,AI training,and similar technologies.
Anion exchange membrane water electrolyzers(AEMWEs)are emerging as a promising technology due to the high performance and low cost.However,the development of highly active and stable non-precious metal-based catalysts for the anodic oxygen evolution reaction(OER)remains a great challenge.In this study,we present a top-down construction strategy for anode design,resulting in a hierarchical NiFe layered double hydroxide(LDH)/N-doped Co/nickel foam(NF)electrode synthesized via a hydrothermal-gas phase nitridation–electrodeposition method.This electrode features NiFe LDH nanoplates grown on N-doped Co nanowires supported by nickel foam substrates.The NiFe LDH/N-doped Co/NF electrode demonstrates exceptional performance,achieving a current density of 100 mA·cm^(-2) at a low overpotential of 262 mV with minimal attenuation of just 7 mV NiFe LDH/N-doped Co/NF RuO_(2)/NF NiFe LDH/N-doped Co/NF after 100 h of operation.When assembled into an AEMWE,the system requires only 1.63 V to achieve a current density of 1 A·cm^(-2),surpassing the performance of most reported catalysts.The N-doped Co nanowires are shown to enhance both activity and stability by increasing the electrode’s surface area and reinforcing the catalyst–support interaction.
