Concurrent Model Research Articles

Functional concurrent hidden Markov model

Functional concurrent hidden Markov model

Equivalence between LC-CRF and HMM, and Discriminative Computing of HMM-Based MPM and MAP

Practitioners have used hidden Markov models (HMMs) in different problems for about sixty years. Moreover, conditional random fields (CRFs) are an alternative to HMMs and appear in the literature as different and somewhat concurrent models. We propose two contributions: First, we show that the basic linear-chain CRFs (LC-CRFs), considered as different from HMMs, are in fact equivalent to HMMs in the sense that for each LC-CRF there exists an HMM—that we specify—whose posterior distribution is identical to the given LC-CRF. Second, we show that it is possible to reformulate the generative Bayesian classifiers maximum posterior mode (MPM) and maximum a posteriori (MAP), used in HMMs, as discriminative ones. The last point is of importance in many fields, especially in natural language processing (NLP), as it shows that in some situations dropping HMMs in favor of CRFs is not necessary.

Reaching for the Sky: Maximizing Deep Learning Inference Throughput on Edge Devices with AI Multi-Tenancy

The wide adoption of smart devices and Internet-of-Things (IoT) sensors has led to massive growth in data generation at the edge of the Internet over the past decade. Intelligent real-time analysis of such a high volume of data, particularly leveraging highly accurate deep learning (DL) models, often requires the data to be processed as close to the data sources (or at the edge of the Internet) to minimize the network and processing latency. The advent of specialized, low-cost, and power-efficient edge devices has greatly facilitated DL inference tasks at the edge. However, limited research has been done to improve the inference throughput (e.g., number of inferences per second) by exploiting various system techniques. This study investigates system techniques, such as batched inferencing, AI multi-tenancy, and cluster of AI accelerators, which can significantly enhance the overall inference throughput on edge devices with DL models for image classification tasks. In particular, AI multi-tenancy enables collective utilization of edge devices’ system resources (CPU, GPU) and AI accelerators (e.g., Edge Tensor Processing Units; EdgeTPUs). The evaluation results show that batched inferencing results in more than 2.4× throughput improvement on devices equipped with high-performance GPUs like Jetson Xavier NX. Moreover, with multi-tenancy approaches, e.g., concurrent model executions (CME) and dynamic model placements (DMP), the DL inference throughput on edge devices (with GPUs) and EdgeTPU can be further improved by up to 3× and 10×, respectively. Furthermore, we present a detailed analysis of hardware and software factors that change the DL inference throughput on edge devices and EdgeTPUs, thereby shedding light on areas that could be further improved to achieve high-performance DL inference at the edge.

Thermal and mass transfer aspects of MHD flow across an exponentially stretched sheet with chemical reaction

This investigation deals with the two-dimensional steady problem incorporating the magneto hydrodynamic effect with three distinct boundary layer streams of fluid across an exponentially stretched sheet under the influence of thermal radiation and chemical reactions. The concurrent flow model also accounts for the effect of Thermophoresis and Brownian movement, caused by temperature and concentration gradients between two boundaries, respectively. By using suitable similarity transformation variables, the nonlinear systems of governing PDEs are transformed into a system of nonlinear ODEs. The findings, obtained by using the Bvp4c strategy, are analysed and depicted graphically and through tables. Furthermore, the Nusselt number, Sherwood number and drag coefficient, which are crucial for industrial applications, are also shown in the table to demonstrate the impact of dimensionless physical parameters. In response to a steady rise in Lewis number and chemical reaction parameter, the heat transit rate tends to decline while the species transit rate increases.

Concurrent multi-material and multi-scale design optimization of fiber-reinforced composite material and structures for minimum structural compliance

With the development of additive manufacturing technology for fiber-reinforced composite materials, topology optimization of fiber-reinforced composite laminates involving multiple materials and variable stiffness is gaining increasing attention. This study proposes an effective methodology for a Multi-scale and Multi-material Composite Anisotropic Penalization (MMCAP) model to investigate multi-scale and multi-material design optimization of a fiber-reinforced variable stiffness (VS) composite structure to minimize structural compliance. In the concurrent MMCAP model, the macroscopic multi-material structural topology and microscopic discrete fiber laying angle selection are introduced as independent design variables and optimized simultaneously. The modified Solid Isotropic Material with Penalization (SIMP) and the Discrete Material Optimization (DMO) approaches are utilized at the macro- and micro-scales, respectively, to realize a clear macroscopic multi-material structural topology and microscopic specific discrete fiber laying angle selection. Multi-material fiber-reinforced plastic (FRP) materials, such as carbon fiber–reinforced plastic (CFRP) and glass fiber–reinforced plastic (GFRP), are considered two types of solid materials in terms of structural volume cost. Sensitivity analysis of the structural compliance concerning the variables of the two geometrical scales is performed using the analytical sensitivity analysis method. The DMO approach is utilized to couple two geometrical scales: macroscopic topology and microscopic material selection. The capabilities of the proposed MMCAP are demonstrated by concurrent multi-material and multi-scale design optimization of composite panels. The influence of the number of discrete fiber laying angles on the structural compliance and optimized topology configuration is also been discussed. Numerical studies showed that the proposed MMCAP scheme can effectively realize multi-material and multi-scale design optimization of fiber-reinforced composite structure with achieving a clear macroscopic multi-material structural topology and microscopic fiber laying angle. The proposed MMCAP scheme provides a new implementation strategy for lightweight, multi-material, and multi-scale design optimization of composite materials, considering the design and manufacturing collaboration through additive manufacturing technology.

A generic framework for representing and analyzing model concurrency

Recent results in language engineering simplify the development of tool-supported executable domain-specific modeling languages (xDSMLs), including editing (e.g., completion and error checking) and execution analysis tools (e.g., debugging, monitoring and live modeling). However, such frameworks are currently limited to sequential execution traces and cannot handle execution traces resulting from an execution semantics with a concurrency model supporting parallelism or interleaving. This prevents the development of concurrency analysis tools, like debuggers supporting the exploration of model executions resulting from different interleavings. In this paper, we present a generic framework to integrate execution semantics with either implicit or explicit concurrency models, to explore the possible execution traces of conforming models, and to define strategies for helping in the exploration of the possible executions. This framework is complemented with a protocol to interact with the resulting executions and hence to build advanced concurrency analysis tools. The approach has been implemented within the GEMOC Studio. We demonstrate how to integrate two representative concurrent meta-programming approaches (MoCCML/Java and Henshin), which use different paradigms and underlying foundations to define an xDSML’s concurrency model. We also demonstrate the ability to define an advanced concurrent omniscient debugger with the proposed protocol. The paper, thus, contributes key abstractions and an associated protocol for integrating concurrent meta-programming approaches in a language workbench, and dynamically exploring the possible executions of a model in the modeling workbench.

CG-enriched concurrent multi-scale modeling of dynamic surface interactions between discrete particles and solid continua

CG-enriched concurrent multi-scale modeling of dynamic surface interactions between discrete particles and solid continua

Suppress that algae: Mitigating the effects of harmful algal blooms through preemptive detection & suppression

Recently, Harmful Algal Blooms (HABs), e.g., toxic red tide and blue-green algae, have suffocated and plagued coastlines and inland water bodies leading to economic losses. Currently, batch data collection in many areas is inadequate and reactionary following regional fish kills. Therefore, there is an inherent need for a “smart” solution aimed at preemptive detection and mitigation of the impending super bloom through combination of concurrent measurements, modeling, and mitigation using suppression agents. Our hypothesis is if a Seek and Destroy Algal Mitigation System (SDAMS) is engineered with a multitude of capabilities, including remote algal parameter sensing, wireless data transmission, and mitigation using suppression agents, then it will preemptively detect and mitigate HABs. The SDAMS includes (i) a floatation device with a Wi-Fi microcontroller and five sensors for concurrent measurements, (ii) real-time data transmission to the cloud, (iii) data visualization; diagnostic and predictive analysis, and (iv) an algae suppression component. During laboratory testing, physical and chemical agents used for mitigation noticeably suppressed the algae in various ways. Algal suppression occurred by either reducing pH, increasing Dissolved Oxygen (DO), or exerting high mechanical properties. We conducted predictive analytics to quantify the influence of the suppression agent on algae and compared the green spectrum strength (indicating algal intensity) to the DO concentration. Using machine learning, a 4th order polynomial equation with 94% accuracy provided the best curve-fit to explain the green spectrum-to-DO relationship. This cost-effective solution can be applied to instantaneously suppress, or preemptively mitigate, HABs to minimize environmental impacts.

A method for constructing concurrent processing model of Xiangxi Hmong part-of-speech tagging

A method for constructing concurrent processing model of Xiangxi Hmong part-of-speech tagging

A Quality-Related Concurrent Dual-Latent Variable Model and Its Semi-Supervised Soft Sensor Applications

To provide accurate key variable prediction, data-driven soft sensing techniques have extracted much attention in recent years. Due to different control strategies in industrial processes, it is noticed that the variables in the control loops can be autocorrelated while the others may be static, which needs to be considered simultaneously. In this paper, a quality-related concurrent dual-latent variable (CDLV) model is proposed for soft sensing construction. Two different kinds of latent variables are adopted to learn quality-related dynamic information and quality-related static information respectively. Both quality-related variables are then applied for quality prediction purposes. On this basis, the CDLV model is extended to a semi-supervised form to provide a comprehensive description for the soft sensor design with insufficient quality information. The proposed models are demonstrated by two industrial cases which show superiority over other relative methods in the accuracy of key quality variables prediction.

A Concurrent Switching Model for Traffic Congestion Control

We introduce a new conservation-based approach for traffic coordination modeling and control in a network of interconnected roads (NOIR) with switching movement phase rotations at every NOIR junction. For modeling of traffic evolution, we first assume that the movement phase rotation is cyclic at every NOIR junction, but the duration of each movement phase can be arbitrarily commanded by traffic signals. Then, we propose a novel concurrent switching dynamics (CSD) with deterministic transitions among a finite number of states, representing the NOIR movement phases. We define the CSD control as a cyclic receding horizon optimization problem with periodic quadratic cost and constraints. More specifically, the cost is defined so that the traffic density is minimized and the boundary inflow is uniformly distributed over the boundary inlet roads, whereas the cost parameters are periodically changed with time. The constraints are linear and imposed by a trapezoidal fundamental diagram at every NOIR road so that traffic feasibility is assured and traffic over-saturation is avoided.

DIGITAL LITERACY AS A TOOL FOR FOSTERING ENGAGEMENT AND MOTIVATION IN ONLINE/HYBRID ESP CLASSES

Technology is inevitably connected to the teaching/learning process. It helps instructors, but at the same time, technology transfers responsibility for learning to students. Students can guide their learning at their own pace, direct their progress and have access to course content by participating in online/hybrid learning. That is why digital skills are essential in today’s education and society. But, what digital skills are needed to learn and thrive today? How about the skills needed to encourage engagement and motivation in learning? Digital literacy and confidence for everyone involved in education have become more vital than ever.This paper describes the results of a one-semester English for Information Technology course at the South East European University in Macedonia to determine if the core curriculum items provide motivation for students engaged in online/hybrid class. Also, it determines whether the students have the capabilities and skills to participate entirely in the digital society.The hybrid model discussed in this paper is the concurrent teaching-learning model. In this model, the author taught some students who were in person with her in the physical classroom, and at the same time, the teacher’s instruction was being streamed live through Google Meet with other students logged in at home. The teaching/learning process results will show why it is necessary to change our classroom practices in the modern education system.

Cognitive and Interpersonal Factors Affecting Social Adjustment of University Students in Pakistan

Cognitive and interpersonal factors play an important role in the social adjustment of students. Factors affecting the social adjustment of university students have been verified in different cultures. However, no study has tested a concurrent model with the study variables in the Pakistani context. This study aimed to investigate the effect of personal and interpersonal factors on the social adjustment of university students in Pakistan. Three hundred participants from the Azad Jammu and Kashmir regions of Pakistan responded on a questionnaire package containing self-reported measures on social self-efficacy, social anxiety, teachers' social support, and peers' social support. The results indicate that self-efficacy, teachers' support, and peer support have a significant positive effect on the social adjustment of university students, suggesting that an enhanced self-efficacy, and increased teachers' and peers' support would increase social adjustment. However, academic anxiety is inversely associated with social adjustment, suggesting that a higher anxiety level would result in a reduction in social adjustment. Students should be given more opportunities to enhance self-efficacy, obtain social support, and reduce social anxiety.

The Urgency of Mixed Methods Studies in Tourism Research For the Development of Tourist Destinations and Attractions: A Practical Application in Tanjungpinang City, Indonesia

Many literature reviews have found that most tourism and hospitality industry research uses quantitative methods; this implies the dominance of the positivist paradigm, which only observes the surface of a case. Tourism science is complicated because it has a multiplier effect involving social, economic, political, technological, cultural, religious, and other aspects with different complexities, so to get an in-depth study, it must be analyzed with a different approach. This study claims to obtain relevant and unbiased findings in formulating development strategies in the tourism industry; the analysis must be carried out through testing (quantitative) and observation (qualitative). This study describes how to apply mixed methods and the process of mixing data. The researchers used tourism data in Tanjungpinang City for matching practice as a case study. Using the concurrent triangulation model and the IFAS EFAS approach, which is a balanced analysis of quantitative and qualitative mixtures, then reanalyzed by sorting out internal and external problems to produce an IE matrix. The data collection method used is questionnaires, interviews, and observations. The quantitative data were analyzed using the percentage technique, and the qualitative using the Miles and Huberman technique. The results describe the tourism potential that can be developed to become a tourism icon of Tanjujgpinan City, namely art, culture, and natural attractions, but the supporting infrastructure still needs to be improved. The IE analysis produces a matrix in cell II, namely the grow and built strategy; this strategy can help tourism planners determine the right direction of development and strategy.

Multiscale topology optimization with stress, buckling and dynamic constraints using adaptive geometric components

This research focuses on the multiscale topology optimization (TO) of structures to meet stress, stability, and vibration requirements. The compliance minimization problems of multiscale structures under the global stress constraints, critical buckling load factor, and natural frequencies are established and solved by employing a method using adaptive geometric components (AGCs). During the TO process, the dynamic and stability frequency gap constraints are added to the optimal model to prevent mode switches. The maximum stress is measured using the global p-norm stress aggregation. The concurrent material modeling technique, which involves AGCs, illustrates the macrostructure and microstructure of material under multiple constraints simultaneously, which are optimized by focusing on the AGC parameters under multiple constraints. Benchmark numerical tests are presented and investigated to validate the applicability of the proposed approach.

Efficient computational framework for image-based micromechanical analysis of additively manufactured Ti-6Al-4V alloy

The increase in additively manufactured (AM) Ti-6Al-4V alloys in high-performance industrial applications has necessitated the development of robust computational models that can aid in their qualification and certification. Physics-based micromechanical models, relating the AM-processed material microstructure and defect state with the overall material response and life, can play an important role in reducing uncertainty in component behavior and increasing acceptance. Motivated by this need, the present paper develops a novel image-based crystal plasticity finite element model (CPFEM) for efficient micromechanical simulation of the additively manufactured Ti-6Al-4V alloy, whose Widmanstätten microstructure is characterized by 12 HCP α lath variants in the parent β grain. A unique feature of this work is the creation of an efficient crystal plasticity framework for the parent β grain polycrystalline ensembles with parametric representation of the α lath statistics of size, shape, orientation, and crystallography. This statistical representation is expected to significantly enhance its efficiency over models that represent each α lath explicitly in the microstructure. Defects in the form of voids are represented at two scales. The smaller voids in the microstructure are manifested as porosity or void volume fraction distribution in the crystal plasticity model. Larger voids are represented explicitly in the statistically equivalent microstructural volume element (SEMVE) model. The models are built from experimentally acquired electron back scatter diffraction (EBSD) and micro-focus X-ray computed tomography (XCT) images and calibrated and validated with mechanical testing data. This paper extends the developments in Pinz et al. (2022) through the development of a special self-consistent boundary condition in the context of a concurrent model to overcome limitations of periodicity boundary conditions. The concurrent model embeds the SEMVE in a homogenized exterior domain represented by a rate-dependent isotropic plasticity model. Parametric studies are conducted to comprehend the effect of void size, shape and orientation on the overall material response.

Multi-scale modelling of deformation heterogeneities in explosively welded layered sheets

A multi-scale numerical investigation of local heterogeneities in the strain and stress fields occurring during forming of explosively welded layered metallic sheets is the main goal of the research. Explosive welding is a complex process involving various phenomena occurring in materials during an impact at high velocities and pressures, especially at the interfaces of colliding metals. As a result, the interface of the layered metallic sheets is often highly heterogeneous at the microscale level, what directly affects the sheet behaviour under subsequent forming conditions. To investigate this issue, the mesh-free numerical model of the explosive welding process is used first to recreate the characteristic features of the interface morphology. Various detonation velocities are used to provide diversified morphological features at the interface. Obtained results are then used as input data to develop the concurrent multi-scale finite element model of material behaviour under deformation conditions. The multi-scale modelling concept with explicit representation of the interface region is used. The highly refined heterogeneous FE mesh was generated in the interface region to capture local heterogeneities occurring at the microscale. Particular attention is put on numerical investigation of an influence of interface morphology in the welding zone on the development of the stress localisation that may directly lead to fracture initiation during forming.

Concurrent Optimization of Subway Vertical Alignments and Station Elevations With Improved Particle Swarm Optimization Algorithm

Station elevations and vertical alignment between adjacent stations often influence each other. Therefore, they should be optimized concurrently. However, existing studies mainly focus on optimizing station elevations or links separately. In this paper, a concurrent optimization model of stations and links is developed, in which the vertical points of intersection in stations and links are design variables. The comprehensive cost including construction, energy consumption, and travelling time, is the optimized objective function. Particle swarm optimization (PSO) is a method for searching in continuous space and is widely used for alignment optimization. Considering the characteristics of concurrent optimization for stations and links, the PSO algorithm is improved by modifying the updating formulas for particles and designing two strategies for particle updating, namely “Stations before Links” and “Links before Stations” strategy. A dynamic adaptive feasible region is proposed to handle the complex constraints during optimization. This method is applied here to a real-world case. The applications demonstrate that this method can automatically generate vertical alignments which jointly optimize the locations of stations and links and satisfy all the complex constrains.

Real-time Associations between Functional Behaviors and Depressed Mood in Post-stroke Individuals: Smartphone-based Ecological Momentary Assessment

<h3>Research Objectives</h3> To assess real-time associations between daily functional behaviors and depressed mood (DM) in indivdiuals after stroke using smartphone-based ecological momentary assessment (EMA). <h3>Design</h3> EMA involved repeated sampling of real-time daily functional activities and DM in stroke survivors' natural environments. We used multilevel models to examine concurrent and time-lagged associations between daily functional behaviors and DM. <h3>Setting</h3> Home and community. <h3>Participants</h3> Two hundred and two participants with mild-to-moderate chronic stroke participants (90% ischemic stroke; 45% female; 44% black). <h3>Interventions</h3> Not applicable. <h3>Main Outcome Measures</h3> Participants completed EMA five times per day for two weeks by reporting their daily functional behaviors regarding where, with whom, and what activity was spent and their DM level. <h3>Results</h3> Participants spent 28% of the wake-up time in passive leisure activities but spent less time in physical and vocational activities (4% and 9%, respectively). Lower DM was concurrently associated with time spent in social activities (β=-0.023 ± 0.011) and instrumental activities of daily living (β=-0.061 ± 0.013); spending time with family members (β=-0.061 ± 0.014), spouses (β=-0.043, ± 0.016), friends (β=-0.094, ± 0.021), and coworkers (β=-0.050 ± 0.021); locating in restaurants (β=-0.068 ± 0.029), and homes of family members (β=-0.039 ± 0.020) or friends (β=-0.069 ± 0.031). DM was reported higher when patients rated lower in perceived satisfaction, performance, and engagement of activities in concurrent (βs=-0.036 ± 0.003, -0.053 ± 0.003, -0.044 ± 0.003) and time-lagged models (βs=-0.011 ± 0.004, -0.012 ± 0.004, -0.013 ± 0.004). <h3>Conclusions</h3> EMA can delineate real-time associations between post-stroke functional behaviors and DM. Findings suggest promoting participation in social activities and instrumental activities of daily living as important intervention targets to reduce DM in stroke survivors. <h3>Author(s) Disclosures</h3> None disclosed.

Concurrent modeling of porosity and microstructure in multilayer three-dimensional simulations of powder-bed fusion additive manufacturing of INCONEL 718

Laser powder-bed fusion (L-PBF) additive manufacturing is a complex process whereby a laser selectively scans over layers of powder, inducing local melting and consolidation of solid metal. A key bottleneck issue which prevents L-PBF from further industrial adoption is the low repeatability of the process across different printers and powder batches. To explore and address the challenges of process repeatability, high-fidelity comprehensive physics-based modeling is desirable because it allows fast exploration of the process parameters space by simulations rather than by trial-and-error experiments. In this work, we present a complete framework which is capable of modeling defects (porosity) as well as microstructure explicitly in three dimensions and for multiple layers, effectively demonstrating explicit L-PBF of a digital cube of material at millimeter scale. The phase field method is coupled to the Lattice Boltzmann method, and the framework is used to model both the solid/liquid phase transitions as well as grain nucleation and growth. Explicit parallel scheme allows to run multiple simulations in a reasonable time and explore the role of process parameters on porosity and microstructure concurrently. After model calibration using experimentally-printed and characterized samples, the model was deployed across multiple parameter combinations. The predicted porosity and microstructure agreed well with the experiments. Our framework provides a practical demonstration of digitalization of the L-PBF process and paves the way for assessing process quality and repeatability by computer modeling and simulations.