Complex Structural Models Research Articles

Deep learning from a statistical perspective

As one of the most rapidly developing artificial intelligence techniques, deep learning has been applied in various machine learning tasks and has received great attention in data science and statistics. Regardless of the complex model structure, deep neural networks can be viewed as a nonlinear and nonparametric generalization of existing statistical models. In this review, we introduce several popular deep learning models including convolutional neural networks, generative adversarial networks, recurrent neural networks, and autoencoders, with their applications in image data, sequential data and recommender systems. We review the architecture of each model and highlight their connections and differences compared with conventional statistical models. In particular, we provide a brief survey of the recent works on the unique overparameterization phenomenon, which explains the strengths and advantages of using an extremely large number of parameters in deep learning. In addition, we provide a practical guidance on optimization algorithms, hyperparameter tuning, and computing resources.

Influence of different Processing Parameter on distortion and Residual Stress of Inconel 718 Alloys Fabricated by Selective Laser Melting (SLM)

Anstract Selective Laser Melting (SLM) is one of additive manufacturing techniques which enable to build a complex structure model layer by layer with 3D CAD software. However, a higher research cost makes it hardly carry out by a traditional methodology, the best way to solve the problem is using simulation software. This paper aims to find an optimal processing parameters combination for a specimen with the smallest distortion and lowest residual stress through Simufact Additive (SA) software. Simulation results under an optimal processing parameters, which led to a smallest value of distortion and residual stress, was the combination of scan power with 300W, scan speed with 1.3m/s, scan interval with one spot diameter (0.12mm) and heat treatment holding time with 4h. In addition, the calculation results provide a novel study method to verify the influence of different processing parameters on Inconel 718 alloys fabricated by SLM.

Structural modeling of a forest cluster using discrete mathematics

Economic modeling allows studying the development trends of clusters and planning the introduction of control actions in this socio-economic system to achieve the necessary stabilization and development trends. The cluster can be described as a set of information about its elements and the relationships between them and can be represented as a graph, where the vertices are the agents, and the directed edges are connections between them. Cluster structural formations can be considered as social graphs containing information about heterogeneous factors and the relations between them. Complex structural modeling of the forest cluster allows creating a formal representation of the tasks of identifying the objects (socio-economic systems), analyzing their complexity, coherence, stability, and development scenarios. To do this, we applied the apparatus of discrete mathematics. Signed graphs make it possible to formally make forecasts of the development or trajectory of the simulated system in the phase space of its variables (factors) based on information about its structure and development programs by means of approximating them with pieces of trajectories of impulse processes in signed digraphs. Modeling with the help of the graph theory makes it possible to remove uncertainty associated with predicting the development of a complex system and to propose the option of controlling a stochastic process.

Should Health Care Demand Interpretable Artificial Intelligence or Accept "Black Box" Medicine?

Health care applications of artificial intelligence (AI) have recently emerged. Artificial intelligence approaches, such as deep learning, rely on vast amounts of data and complex model structures ...

Automatic scaled boundary finite element method for three-dimensional elastoplastic analysis

This paper introduces an automatic way of performing 3D static and dynamic elastoplastic analyses in the framework of the scaled boundary finite element method (SBFEM), which only requires the boundary discretization and thus can provide high flexibility in automatic mesh generation. The input models in this paper are described by Standard Tessellation Language (STL) format due to its simplicity and popularity in computer-aided design. The automatic mesh generation from any input geometry is achieved by utilizing the octree decomposition algorithm and boundary trimming. An efficient approach for 2D static image-based elastoplastic analysis based on SBFEM is extended to 3D static and dynamic elastoplastic analyses in this present work. In this improved approach, the return mapping algorithm is only required to be performed at the scaling center of each subdomain in the yield zone. Constant elastoplastic constitutive matrix and internal stresses are used within each yielded subdomain as well. This will greatly simplify the implementation of elastoplastic formulation and reduce the costs involved in the elastoplastic analysis as the return mapping algorithm is computationally expensive. Meanwhile, stabilization matrix is also introduced in the elastoplastic stiffness matrix to eliminate the spurious modes. Numerical examples are presented in this paper to show the feasibility and accuracy of the proposed approach as well as its capability of modelling complex structures in practical applications.

Regularized Structural Equation Modeling to Detect Measurement Bias: Evaluation of Lasso, Adaptive Lasso, and Elastic Net

ABSTRACT Correct detection of measurement bias could help researchers revise models or refine psychological scales. Measurement bias detection can be viewed as a variable-selection problem, in which biased items are optimally selected from a set of items. This study investigated a number of regularization methods: ridge, lasso, elastic net (enet) and adaptive lasso (alasso), in comparison with maximum likelihood estimation (MLE) for detecting various forms of measurement bias in regard to a continuous violator using restricted factor analysis. Particularly, complex structural equation models with relatively small sample sizes were the study focus. Through a simulation study and an empirical example, results indicated that the enet outperformed other methods in small samples for identifying biased items. The alasso yielded low false positive rates for non-biased items outside of a high number of biased items. MLE performed well for the overall estimation of biased items.

The Community Land Model Version 5: Description of New Features, Benchmarking, and Impact of Forcing Uncertainty

AbstractThe Community Land Model (CLM) is the land component of the Community Earth System Model (CESM) and is used in several global and regional modeling systems. In this paper, we introduce model developments included in CLM version 5 (CLM5), which is the default land component for CESM2. We assess an ensemble of simulations, including prescribed and prognostic vegetation state, multiple forcing data sets, and CLM4, CLM4.5, and CLM5, against a range of metrics including from the International Land Model Benchmarking (ILAMBv2) package. CLM5 includes new and updated processes and parameterizations: (1) dynamic land units, (2) updated parameterizations and structure for hydrology and snow (spatially explicit soil depth, dry surface layer, revised groundwater scheme, revised canopy interception and canopy snow processes, updated fresh snow density, simple firn model, and Model for Scale Adaptive River Transport), (3) plant hydraulics and hydraulic redistribution, (4) revised nitrogen cycling (flexible leaf stoichiometry, leaf N optimization for photosynthesis, and carbon costs for plant nitrogen uptake), (5) global crop model with six crop types and time‐evolving irrigated areas and fertilization rates, (6) updated urban building energy, (7) carbon isotopes, and (8) updated stomatal physiology. New optional features include demographically structured dynamic vegetation model (Functionally Assembled Terrestrial Ecosystem Simulator), ozone damage to plants, and fire trace gas emissions coupling to the atmosphere. Conclusive establishment of improvement or degradation of individual variables or metrics is challenged by forcing uncertainty, parametric uncertainty, and model structural complexity, but the multivariate metrics presented here suggest a general broad improvement from CLM4 to CLM5.

The use of reduced models in the reliability analysis of deformable systems with dynamic vibration dampers

The transformations of the initial design models of complex constructions and structures as the multi-degrees of freedom (DOF) systems is advantageous to use in order to reduce the amount of computational procedures in their dynamic analysis. This becomes particularly relevant in the reliability analysis requiring multiple recalculations for taking into account the stochastic properties of system parameters and impacts. The paper presents a technique based on known approaches of reduction the original multi-DOF system to an equivalent generalized design model with one degree of freedom. Various combinations of the equivalence conditions for the original system and the generalized model, including equalities of their natural frequencies, characteristic displacements or kinetic energies, were used to determine the modelling lumped mass and the corresponding rigidity of the design model. An algorithm for the probabilistic properties evaluation of the design model characteristics has been developed. The verification of the correctness of the proposed model in the area of tuning in to the workload frequency is performed. The quantitative results of solving the model problems in reliability calculations of multi-DOF structures with dynamic vibration dampers (DVD) are presented. The proposed approach allows us to simplify the process of selecting a single-mass DVD taking into account reliability requirements.

Directional transmission of ultra-high frequency acoustic signals based on metamaterial structure

Cell photoacoustic detection of acoustic signals has serious problems of transmission loss and difficulty in acquisition. Based on the acoustic transmission characteristics of acoustic supermaterials, the acoustic wave directional transmission model is designed using COMSOL software, and a finite element simulation experiment was carried out. Experiments show that the model designed in this paper has good sound wave focusing and acoustic wave directional transmission effects. According to the specific application in different environments, this paper designs a variety of transmission models with different structures and carries out comparative experiments to verify the influence of complex model structures on the directional transmission of sound waves. At the same time, the experimental analysis of the acoustic wave directional transmission effect of the model under different frequency segments is carried out, and the possibility of acoustic wave logic operation is verified. Models can be applied to high frequency acoustic signal acquisition and sound wave transmission in complex environments. These studies have important significance for the development of cell photoacoustic detection technology and the application of metamaterial structures and have good practical merit.

Modeling the positioning process of lifting “crane –balloon

In hard-to-reach places, the performance of transportation and lifting operations with cargo is usually associated with significant financial costs and is a complex engineering task. One of the solutions can be the use of a lifting crane. It is a balloon filled with “light” gas, equipped with a cable system (sling) for suspending the load on the ball and connecting the ball with three drive motors reliably fixed on the ground, which rotate the drums with wound cables. The purpose of the work is to perform a mathematical description and simulation of the movement of objects of the electromechanical system of a "crane ball" with an automated electric drive, to assess the possibility of ensuring accurate positioning of the load when using a direct current electric drive with a subordinate control system. In the electromechanical system under consideration, the relocatable load has five degrees of freedom; dynamic processes are described by a system of five nonlinear differential Lagrange equations of the second kind in the corresponding coordinates. The linearization of the equations is performed by expanding into a Taylor series, limiting the representation of nonlinearity to the first two members of the series. Positioning of the ball and the load in space is provided by changing the lengths of the cables holding the ball. Therefore, during the synthesis of the positional control system for the electric drives of the drums, the dependences of the required cable lengths were determined as a function of the given coordinates of the ball (load) in the x, y, z coordinate space. Taking into account the accepted assumptions and requirements for limiting the oscillations of the cargo, a rather complex structural simulation model was obtained, containing several computational units interconnected in accordance with the system operation algorithm. The proposed mathematical model of the object and transient processes carried out during the Matlab structural simulation showed the ability to solve the load positioning problem by using a three-loop subordinate control system (control loops for the electromagnetic momentum, engine rotational speed and load position) for electric drums with cables holding the ball.

The relationship between the intrinsic Čech and persistence distortion distances for metric graphs

Metric graphs are meaningful objects for modeling complex structures that arise in many real-world applications, such as road networks, river systems, earthquake faults, blood vessels, and filamentary structures in galaxies. To study metric graphs in the context of comparison, we are interested in determining the relative discriminative capabilities of two topology-based distances between a pair of arbitrary finite metric graphs: the persistence distortion distance and the intrinsic Cech distance. We explicitly show how to compute the intrinsic Cech distance between two metric graphs based solely on knowledge of the shortest systems of loops for the graphs. Our main theorem establishes an inequality between the intrinsic Cech and persistence distortion distances in the case when one of the graphs is a bouquet graph and the other is arbitrary. The relationship also holds when both graphs are constructed via wedge sums of cycles and edges.

Using Integrative Data Analysis to Investigate School Climate Across Multiple Informants.

Survey research frequently involves the collection of data from multiple informants. Results, however, are usually analyzed by informant group, potentially ignoring important relationships across groups. When the same construct(s) are measured, integrative data analysis (IDA) allows pooling of data from multiple sources into one data set to examine information from multiple perspectives within the same analysis. Here, the IDA procedure is demonstrated via the examination of pooled data from student and teacher school climate surveys. This study contributes to the sparse literature regarding IDA applications in the social sciences, specifically in education. It also lays the groundwork for future educational researchers interested in the practical applications of the IDA framework to empirical data sets with complex model structures.

An effective parametric model reduction technique for uncertainty propagation analysis in structural dynamics

An efficient formulation for uncertainty propagation analysis of complex structural models is presented. The formulation is based on parametric reduced-order models. Fixed-interface normal modes and interface modes are approximated in terms of a set of support points in the uncertain parameter space. The potential time-consuming step of computing the modes for different values of the model parameters needs to be performed only at the support points. Based on these approximate modes, reduced-order matrices can be updated efficiently during the simulation process associated with the uncertainty propagation analysis. The effectiveness of the proposed parametric model reduction technique is demonstrated by means of two numerical examples.

Model discovery for studies of surface morphological modifications based on Kuramoto-Sivashinsky dynamics.

A wide range of observations in studies of surfaces exposed to ion beams can be explained and analyzed successfully by continuum models of the Kuramoto-Sivashinsky type. Despite certain progress in the theoretical understanding of the model parameters on the basis of atomistic models, much of the applications are based on phenomenological determination of several unknown quantities. In this work a numerical tool is discussed and investigated, which allows us to determine model coefficients and complex model structures from experimental findings. The method resembles known approaches in machine learning and data-driven reconstruction techniques. To keep the discussion on a fundamental level, numerical simulations are conducted by employing a scaled test model. The reconstruction technique is demonstrated for this model system and shows a high accuracy in recovering input parameters for situations without beam noise. As an application to an unknown system to be explored, the algorithm is then applied to a system with lognormal distributed ion bombardment. The impact of the beam fluctuations in the proposed model are discussed. Perspectives of the numerical algorithm for an analysis of experimental data are addressed.

How many ferrocene units of multi-ferrocenyl complexes can react with the electrode?

Ferrocenylcarboxylic acid Fe, Co, and Ni complexes were synthesized as model complexes to investigate how many ferrocene units of multi-ferrocenyl system can react with the electrode. The molecular structures of model complexes were characterized by X-ray single-crystal diffraction, the diffusion coefficient of the complexes and ferrocene was determined by the diffusion-ordered spectroscopy and Einstein–Stokes equation, the electrode reaction numbers of ferrocene units were determined by cyclic voltammetry and the Randles–Sevcik equation. A model of electrode reaction of multi-ferrocenyl system was proposed in this paper, which was used to explore how many ferrocene units of multi-ferrocenyl complexes can react with the electrode.

Multiclass heterogeneous domain adaptation via bidirectional ECOC projection

Heterogeneous domain adaptation aims to exploit the source domain data to train a prediction model for the target domain with different input feature space. Current methods either map the data points from different domains with different feature space to a common latent subspace or use asymmetric projections for learning the classifier. However, these learning methods separate common space learning and shared classifier training. This may lead complex model structure and more parameters to be determined. To appropriately address this problem, we propose a novel bidirectional ECOC projection method, named HDA-ECOC, for heterogeneous domain adaptation. The proposed method projects the inputs and outputs (labels) of two domains into a common ECOC coding space, such that, the common space learning and the shared classifier training can be performed simultaneously. Then, classification of the target testing sample can be directly addressed by an ECOC decoding. Moreover, the unlabeled target data is exploited by estimating the two domains projected instances consistency through a maximum mean discrepancy (MMD) criterion. We formulate this method as a dual convex minimization problem and propose an alternating optimization algorithm for solving it. For performance evaluation, experiments are performed on cross-lingual text classification and cross-domain digital image classification with heterogeneous feature space. The experimental results demonstrate that the proposed method is effective and efficient in solving the heterogeneous domain adaptation problems.

Estimation of peak wind response of building using regression analysis

The maximum along-wind displacement of a considerable amount of building under simulated wind loads is computed with the aim to produce a simple prediction model using multiple regression analysis with variables transformation. The Shinozuka and Newmark methods are used to simulate the turbulent wind and to calculate the dynamic response, respectively. In order to evaluate the prediction performance of the regression model with longer degree of determination, two complex structural models were analyzed dynamically. In addition, the prediction model proposed is used to estimate and compare the maximum response of two test buildings studied with wind loads by other authors. Finally, it was proved that the prediction model is reliable to estimate the maximum displacements of structures subjected to the wind loads.

A Dual Attention Network with Semantic Embedding for Few-Shot Learning

Despite recent success of deep neural networks, it remains challenging to efficiently learn new visual concepts from limited training data. To address this problem, a prevailing strategy is to build a meta-learner that learns prior knowledge on learning from a small set of annotated data. However, most of existing meta-learning approaches rely on a global representation of images and a meta-learner with complex model structures, which are sensitive to background clutter and difficult to interpret. We propose a novel meta-learning method for few-shot classification based on two simple attention mechanisms: one is a spatial attention to localize relevant object regions and the other is a task attention to select similar training data for label prediction. We implement our method via a dual-attention network and design a semantic-aware meta-learning loss to train the meta-learner network in an end-to-end manner. We validate our model on three few-shot image classification datasets with extensive ablative study, and our approach shows competitive performances over these datasets with fewer parameters. For facilitating the future research, code and data split are available: https://github.com/tonysy/STANet-PyTorch

Abstract 779: Unraveling exosome-enabled cancer signaling: An integrated genomic approach

Abstract Exosomes play important roles in mediating intercellular communication by transferring various cargos (lipids, RNAs and proteins) into recipients. Emerging evidence has indicated strong implication of exosomes and their RNA cargos in human and animal health including cancers, however, little is known about the mechanisms. In this study, we focus on a specific class of exosomal cargo, microRNAs, to explore new machineries underlying exosome-controlled microRNA secretion and trafficking and subsequently, to assess the regulatory impacts of cancerous noncoding RNAs to surrounding cells or remote tissues in terms of signaling transduction. Particularly, we developed a series of novel bioinformatics techniques that address important problems related to sequence motif finding (primarily based on short RNAs) and gene regulation network analysis (modeling complex structures that reflect multifaceted RNA cooperation and competition). Applying public exosomal microRNA sequence data and TCGA cancer transcriptomic profiles on various types of cancer, we have identified several candidate RNA motifs that might be responsible for guiding the transporter-mediated microRNA loading into exosomes, along with corresponding gene networks that are controlled by those circulating intruders. For example, in human colon cancer SW620 case, three 4-mer sequence motif patterns including [AU][CUG][UG]G, [CGU][UA][GU]G, [AUG][CG]U[UG] are highly enriched among exosomal microRNA sequences (with > 85% coverage, adj.p-value <=1.83E-07). Cell transfection experiments were performed on mutants of the constituent motifs versus their wild-type sequences, and have confirmed motifs, e.g., GGUG, were associated with microRNA loading into exosomes, mostly specific to a certain type of cancer. Results from network analysis further shows such exosomal microRNAs can regulate key signaling processes related to TGF-beta, MAPK, Neurotrophin, and Ras, in the recipient cells through interacting with significant numbers of target genes participating in those pathways. Moreover, the dynamic and conditional aspect of the network analysis enables the identification of microRNA regulatory modules as well as the characterization of regulatory transition across stages along cancer progression. Overall, we demonstrate here that exosome-mediated microRNAs regulation could be an effective channel for cancer cells to communicate with surroundings; and the packaging process is selective depending on unique sequence properties. More importantly, we present an integrative approach using computational modeling and genomic data mining to facilitate the discovery of disease-related exosome and microRNA functions, which can be generally useful for generating new insights and novel intervention tools. Citation Format: Juan Cui, Jiang Shu, Tian Gao, Haluk Dogan. Unraveling exosome-enabled cancer signaling: An integrated genomic approach [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 779.

When the fear of dentist is relevant for more than one's oral health. A structural equation model of dental fear, self-esteem, oral-health-related well-being, and general well-being.

PurposeTo develop and test a complex model that captures the individuals’ general well-being and the specific oral-health-related well-being. We were specifically interested, as a specific research question, if self-esteem, dental fear, and the oral health-related well-being are credible predictors for the general well-being.Patients and methodsA one-time associative research design measured dental-specific anxiety, self-esteem, oral-health-related specific well-being, and general well-being in 281 participants, 3rd and 6th year dental students (MAge =22.59 years, SDAge =3.13; 55% females), which completed a battery of relevant questionnaires: the Dental Fear Survey, the Rosenberg Self-Image Scale, the short form of Oral Health Impact Profile, and the Flourishing Scale. The data were subject to structural equation modeling in order to validate potential pathways of influence hypothesized based on previous evidence from the literature.ResultsWe developed and tested a complex structural equations model, in which dental fear influences both the specific oral-health-related well-being and the persons’ self-esteem. In turn, self-esteem mediates the influence pathways between dental fear and oral-health-specific well-being, on the one hand, and the overall well-being, on the other hand.ConclusionOur research contributes directly to strengthening the theoretical basis for future interdisciplinary research, by providing, first, a tested and replicable model that surpasses the simple correlation or prediction, and second, empirical evidence for the significant mutual interdependence between psychological experiences, eg, self-esteem, and the two main aspects of well-being, ie, specific and general. From a practical, clinical viewpoint, our research provides further insights and justification for the importance of educating the patient, on all levels, from the individual clinical practice to community programs and public oral health policies, with respect to the importance of oral health.