The research on mining hotspot topic in online public opinion is of great significance for improving social management efficiency and promoting economic development. In order to overcome the problems of low accuracy, ...
详细信息
Context: Deep learning has been widely used in Autonomous Driving Systems (ADS). Though significant progress has been made regarding their efficiency and accuracy, uncertainty remains a critical factor affecting ADS s...
详细信息
Context: Deep learning has been widely used in Autonomous Driving Systems (ADS). Though significant progress has been made regarding their efficiency and accuracy, uncertainty remains a critical factor affecting ADS safety. Such uncertainties are often due to environmental noise and/or imperfect algorithm structures. Studies on uncertainty quantification mostly focus on single classification tasks and overlook how uncertainties propagate from the perception to downstream decision-making, studying of which is critical, as the interplay between perception and decision-making can significantly impact the overall safety of ADS. Objectives: We quantify and understand the uncertainty propagation from sensor data to deep learning models, as well as its impact on ADS safety. Methods: We present an empirical study that quantifies both aleatoric and epistemic uncertainties and assesses how such uncertainties propagate and impact ADS safety under various sensor noise conditions. We also investigate the suitability of two epistemic uncertainty quantification methods (i.e., MC Dropout and Deep Ensembles) to ADS tasks and their cost-effectiveness in selecting highly-uncertain samples. Results: Results show that increased noise can significantly increase uncertainty and degrade model performance, thereby compromising decision-making and potentially impacting ADS safety. Both MC Dropout and Deep Ensembles effectively measure the model's epistemic uncertainty, with MC Dropout showing higher correlation with ADS safety, and saving time and computational costs. Moreover, there are significant differences in the highly-uncertain samples they identified. Conclusion: Our results show the importance of considering uncertainty propagation to ensure the ADS safety. Compared to Deep Ensembles, MC Dropout's efficiency makes it a more suitable choice in the context of ADS.
In precise fuel circuit systems, the filtration of particulate impurities seriously affects the efficiency and service life of various components. For filtration process intensification of high-pressure fuel laser per...
详细信息
In precise fuel circuit systems, the filtration of particulate impurities seriously affects the efficiency and service life of various components. For filtration process intensification of high-pressure fuel laser perforated filters, the two-phase flow characteristics in filters is studied. The size, position and number of filtration holes are taken as optimization variables, and take the filtration efficiency and flow pressure drop as optimization objectives. Computational fluid dynamics (CFD) is used to simulate the two-phase motion of continuous phase and discrete particles in a periodic unit. Artificial neural networks (ANN) are utilized for objectives prediction, and the NSGA-II genetic algorithm is employed for multi-objective optimization, resulting in the Pareto front solution set. Furtherly, the reasonable solution is selected by introducing TOPSIS to ensure that two optimization indexes are relatively smaller and balanced. The optimized filter element scheme allows the filter to have a pressure drop of less than 3.2 MPa under high pressure and a filtration efficiency of over 80% for spherical particle impurities with a diameter of 5 μm or more.
The earthworm-based vermiremediation facilitated with benign chemicals such as nano zero-valent iron(nZVI)is a promising approach for the remediation of a variety of soil contaminants including *** themost toxic cyano...
详细信息
The earthworm-based vermiremediation facilitated with benign chemicals such as nano zero-valent iron(nZVI)is a promising approach for the remediation of a variety of soil contaminants including *** themost toxic cyanotoxin,microcystin-LR(MC-LR)enter soil via runoff,irrigated surface water and sewage,and the application of cyanobacterial biofertilizers as part of the sustainable agricultural *** in such remediation systems must sustain the potential risk from both nZVI and *** the present study,earthworms(Eisenia fetida)were exposed up to 14 days to MC-LR and nZVI(individually and inmixture),and the toxicity was investigated at both the organismal andmetabolic levels,including growth,tissue damage,oxidative stress,metabolic response and gut *** showed that co-exposure of MC-LR and nZVI is less potent to earthworms than that of separate *** observations in the co-exposure group revealed only minor epidermal brokenness,and KEGG enrichment analysis showed that co-exposure induced earthworms to regulate glutathione biosynthesis for detoxification and reduced adverse effects from *** combined use of nZVI promoted the growth and reproduction of soil and earthworm gut bacteria(e.g.,Sphingobacterium and Acinetobacter)responsible for the degradation of MC-LR,whichmight explain the observed antagonism between nZVI and MC-LR in earthworm *** study suggests the beneficial use of nZVI to detoxify pollutants in earthworm-based vermiremediation systems where freshwater containing cyanobacterial blooms is frequently used to irrigate soil and supply water for the growth and metabolism of earthworms.
In this paper, we develop a new one-stage O(N log N) algorithm to generate a rank-minimized H-2-representation of electrically large volume integral equations (VIEs), which significantly reduces the CPU run time of st...
详细信息
In this paper, we develop a new one-stage O(N log N) algorithm to generate a rank-minimized H-2-representation of electrically large volume integral equations (VIEs), which significantly reduces the CPU run time of state-ofthe-art algorithms for completing the same task. Unlike existing two-stage algorithms, this new algorithm requires only one stage to build nested cluster bases. The cluster basis is obtained directly from the interaction between a cluster and its admissible clusters composed of real or auxiliary ones that cover all interaction directions. Furthermore, the row and column pivots of the resultant low- rank representation are chosen from the source and observer points in an analytical way without the need for numerically finding them. This further speeds up the computation. Numerical experiments on a suite of electrically large 3D scattering problems have demonstrated the efficiency and accuracy of the proposed new algorithm.
Machine-learned force fields (MLFFs) promise to offer a computationally efficient alternative to ab initio simulations for complex molecular systems. However, ensuring their generalizability beyond training data is cr...
详细信息
Machine-learned force fields (MLFFs) promise to offer a computationally efficient alternative to ab initio simulations for complex molecular systems. However, ensuring their generalizability beyond training data is crucial for their wide application in studying solid materials. This work investigates the ability of a graph neural network (GNN)-based MLFF, trained on Lennard-Jones Argon, to describe solid-state phenomena not explicitly included during training. The MLFF's performance is assessed in predicting phonon density of states (PDOS) for a perfect face-centered cubic (FCC) crystal structure at both zero and finite temperatures. Additionally, vacancy migration rates and energy barriers are evaluated in an imperfect crystal using direct molecular dynamics (MD) simulations and the string method. Notably, vacancy configurations are absent from the training data. These results demonstrate the MLFF's capability to capture essential solid-state properties with good agreement to reference data, even for unseen configurations. Data engineering strategies are further discussed to enhance the generalizability of MLFFs. The proposed set of benchmark tests and workflow for evaluating MLFF performance in describing perfect and imperfect crystals pave the way for reliable application of MLFFs in studying complex solid-state materials.
作者：
wang, yifanTuo, XinlinYe, GangTsinghua Univ
Inst Nucl & New Energy Technol Collaborat Innovat Ctr Adv Nucl Energy Technol Beijing 100084 Peoples R China Tsinghua Univ
Dept Chem Engn Key Lab Adv Mat MOE Beijing 100084 Peoples R China
Developing aramid paper with superb mechanical and dielectric properties is of great significance for honeycomb fabrication and electrical insulating applications. This study presents a facile strategy, by exploiting ...
详细信息
Developing aramid paper with superb mechanical and dielectric properties is of great significance for honeycomb fabrication and electrical insulating applications. This study presents a facile strategy, by exploiting self-healing heterocyclic aramid nanofibers (HANFs) as interfacial cement, to fabricate new hybrid aramid paper with exceptional mechanical and dielectric properties. The HANFs generated by the polymerization-induced self-assembly (PISA) approach exhibit intrinsic interfacial compatibility and hydrogen-bonding driven dynamic fusion behaviors under thermal annealing, enabling the formation of a distinctive concrete-like structure with poly(pphenylene terephthalamide) (PPTA) short fibers and pulps. Substantially reinforced mechanical properties of the hybrid aramid paper, particularly in terms of tensile strength (132.8 MPa, an 18-fold rise) and internal bond strength (358 J/m2, a 13-fold rise), are achieved. Meanwhile, the hybrid aramid paper demonstrates superior dielectric properties with the dielectric breakdown strength increasing from 12.4 kV/mm to 32.4 kV/mm, as well as remarkable flame-resistant properties. Overall, this study offers an adaptable HANF-directed strategy to consolidate hybrid aramid building blocks through engineering their interfacial bonding, opening up opportunities for fabricating new-generation hybrid aramid paper for demanding industrial applications.
PurposeIn the field of engineering, the fractional moments of random variables play a crucial role and are widely utilized. They are applied in various areas such as structural reliability assessment and analysis, stu...
详细信息
PurposeIn the field of engineering, the fractional moments of random variables play a crucial role and are widely utilized. They are applied in various areas such as structural reliability assessment and analysis, studying the response characteristics of random vibration systems and optimizing signal processing and control systems. This study focuses on calculating the fractional moments of positive random variables encountered in engineering. This study focuses on calculating the fractional moments of positive random variables encountered in ***/methodology/approachBy integrating Laplace transforms with fractional derivatives, both analytical and practical numerical solutions are derived. Furthermore, specific practical application methods are *** approach allows for the stable and highly accurate calculation of fractional moments based on the integer moments of random variables. Data experiments included in this study demonstrate the effectiveness of this method in solving fractional moment calculations in engineering. Compared to traditional methods, the proposed method offers significant advantages in stability and accuracy, which can further advance research in the engineering field that employs fractional ***/value(1) Accuracy: Although the proposed method does involve some error, its error level is significantly lower than traditional methods, such as the Taylor expansion method. (2) Stability: The computational error of the proposed method is not only minimal but also remains stable within a narrow range as the fractional order varies. (3) Efficiency: Compared to the widely used Taylor expansion method, the proposed method requires only a minimal number of integer-order moments to achieve the desired results. Additionally, it avoids convergence issues during computation, greatly reducing computational resource requirements. (4) Simplicity: The application steps of the proposed method are very straightforward,
Non-extreme flood disasters caused by urban fluvial flooding can disrupt and impact the operation of urban road traffic systems. This is particularly evident in the influence on the path selection behavior of network ...
详细信息
Non-extreme flood disasters caused by urban fluvial flooding can disrupt and impact the operation of urban road traffic systems. This is particularly evident in the influence on the path selection behavior of network users and the resulting changes in the equilibrium state of the road network. Consequently, the network cannot maintain its original performance, leading to disturbances and interruptions. Therefore, this study proposes a novel stochastic traffic assignment model to simulate and analyze such scenarios. The model proposed in this study introduces a path cost expression that incorporates two stochastic terms, effectively capturing the perceived objective costs for different types of users under non-extreme flooding: flood risk and travel time, as well as the subjective cost factors of the users. Additionally, this study introduces a new criterion to classify paths into acceptable and unacceptable categories. Users will abandon unacceptable paths deemed too dangerous and will choose paths only from their set of acceptable paths until the road network reaches an equilibrium state. The corresponding set of acceptable paths will dynamically change based on the risk sensitivity of different types of users and the prevailing flood conditions. The model developed in this study can effectively analyze the impact of non-extreme floods on the path selection behavior of users with different risk sensitivities and simulates the evolution of the road network's equilibrium state as users instinctively avoid risks. This research provides valuable insights for stakeholders in the operation, management, maintenance, and restoration of road networks under non-extreme flood conditions.
Previous studies showed the crucial effects of morphology and size of polymer vesicles on applications. Herein, we report an efficient approach for producing amphiphilic block copolymer (BCP) vesicles with various siz...
详细信息
Previous studies showed the crucial effects of morphology and size of polymer vesicles on applications. Herein, we report an efficient approach for producing amphiphilic block copolymer (BCP) vesicles with various sizes via tunable distributions of solvophilic monomer units of acrylic acid (AA) on the amphiphilic BCP in reversible addition-fragmentation chain transfer (RAFT)-mediated dispersion polymerization-induced self-assembly. The distribution of solvophilic block was facilely regulated by RAFT agents with two structures and copolymerized with solvophobic monomer (styrene, St). Transmission electron microscopy and dynamic light scattering were utilized for the characterization of the morphology and size of self-assemblies. As a consequence, PAA-b-PSt-b-PSt vesicles with sizes ranging from 85 +/- 8 nm to 692 +/- 59 nm were produced through the distribution of BCP. In addition, large compound vesicles, porous nanospheres, and multilamellar vesicles were observed in the system. Besides, we also combined the experiments with theoretical calculations to demonstrate the effect of different AA distributions in the block copolymer chains on assembly behavior, providing a new strategy for the efficient preparation of polymer nanoparticles.
暂无评论