Abstract: Similarity matrices are 2D representations of the degree of similarity between points of a given dataset which are employed in different fields such as data mining, genetics or machine learning. However, their calculation presents quadratic complexity and, thus, it is specially expensive for large datasets. MPICorMat is able to accelerate the construction of these matrices through the use of a hybrid parallelization strategy based on MPI and OpenMP. The previous version of this tool achieved high performance and scalability, but it only implemented one single similarity metric, the Pearson’s correlation. Therefore, it was suitable only for those problems where data are normally distributed and there is a linear relationship between variables. In this work, we present an extension to MPICorMat that incorporates eight additional metrics for similarity so that the users can choose the one that best adapts to their problem. The performance and energy consumption of each metric is measured in two platforms: a multi-core platform with two Intel Xeon Sandy-Bridge processors and a many-core Intel Xeon Phi KNL. Results show that MPICorMat executes faster and consumes less energy on the many-core architecture. The new version of MPICorMat is publicly available to download from its website: https://sourceforge.net/projects/mpicormat/

Abstract: The problem of counting collisions or interactions is common in areas as computer graphics and scientific simulations. Since it is a major bottleneck in applications of these areas, a lot of research has been done on such subject, mainly focused on techniques that allow calculations to be performed within pruned sets of objects. This paper focuses on how interaction calculation (such as collisions) within these sets can be done more efficiently than existing approaches. Two algorithms are proposed: a sequential algorithm that has linear complexity at the cost of high memory usage; and a parallel algorithm, mathematically proved to be correct, that manages to use GPU resources more efficiently than existing approaches. The proposed and existing algorithms were implemented, and experiments show a speedup of 21.7 for the sequential algorithm (on small problem size), and 1.12 for the parallel proposal (large problem size). By improving interaction calculation, this work contributes to research areas that promote interconnection in the modern world, such as computer graphics and robotics.

Abstract: Current research challenges in hydrology require models with a high resolution on a global scale. These requirements stand in great contrast to the current capabilities of distributed land surface mod- els. Hardly any literature noting efficient scalability past approximately 64 processors could be found. Porting these models to supercomputers is no simple task, because the greater part of the computational load stems from the evaluation of highly parametrized equations. Further- more, the load is heterogeneous in both spatial and temporal dimension, and considerable load-imbalances are triggered by input data. We inves- tigate different domain-decomposition methods for distributed land sur- face models and focus on their properties concerning load balancing and communication minimizing partitionings. Artificial strong scaling exper- iments from a single core to 8, 192 cores show that graph-based methods can distribute the computational load of the application almost as ef- ficiently as coordinate-based methods, while the partitionings found by the graph-based methods significantly reduce communication overhead.

Abstract: Wildcard record is a type of resource records (RRs) in DNS, which can allow any domain name in the same zone to map to a single record value. Former works have made use of DNS zone file data and domain name blacklists to understand the usage of wildcard domain names. In this paper, we analyze wildcard domain names in real network DNS logs, and present some novel findings. By analyzing web contents, we found that the proportion of domain names related to pornography and online gambling contents (referred as abused domain names in this work) in wildcard domain names is much higher than that in non-wildcard domain names. By analyzing behaviors of registration, resolution and maliciousness, we found that abused wildcard domain names have remarkably higher risks in security than normal wildcard domain names. Then, based on the analysis, we proposed GSCS algorithm to detect abused wildcard domain names. GSCS is based on a domain graph, which can give insights on the similarities of abused wildcard domain names’ resolution behaviors. By applying spectral clustering algorithm and seed domains, GSCS can distinguish abused wildcard domain names from normal ones effectively. Experiments on real datasets indicate that GSCS can achieve about 86% detection rates with 5% seed domains, performing much better than BP algorithm.

Abstract: Many scientific communities have adopted community-based models that integrate multiple components to simulate whole system dynamics. The community-based models' software complex, stems from the integration of multiple individual software components that were developed under different application requirements and various machine architectures, has become a challenge for effective software system understanding and continuous software development. The paper presents an integrated software toolkit called X-ray Software Scanner (in abbreviation, XScan) for a better understanding of large-scale community-based scientific codes. Our software tool provides support to quickly summarize the overall information of scientific codes, including the number of lines of code, programming languages, external library dependencies, as well as architecture-dependent parallel software features. The XScan toolkit also realizes a static software analysis component to collect detailed structural information and provides an interactive visualization and analysis of the functions. We use a large-scale community-based Earth System Model to demonstrate the workflow, functions, and visualization of the toolkit. We also discuss using advanced graph analytics techniques to assist software modular design and component refactoring.

Abstract: In this work, we mainly study the mechanism of learning the steganographic algorithm as well as combining the learning process with adversarial learning to learn a good steganographic algorithm. To handle the problem of embedding secret messages into the specific medium, we design a novel adversarial modules to learn the steganographic algorithm, and simultaneously train three modules called generator, discriminator and steganalyzer. Different from existing methods, the three modules are formalized as a game to communicate with each other. In the game, the generator and discriminator attempt to communicate with each other using secret messages hidden in an image. While the steganalyzer attempts to analyze whether there is a transmission of confidential information. We show that through unsupervised adversarial training, the adversarial model can produce robust steganographic solutions, which act like an encryption. Furthermore, we propose to utilize supervised adversarial training method to train a robust steganalyzer, which is utilized to discriminate whether an image contains secret information. Numerous experiments are conducted on publicly available dataset to demonstrate the effectiveness of the proposed method.

Abstract: Outlier detection is an important task in data mining, with many practical applications ranging from fraud detection to public health. However, with the emergence of more and more multi-source data in many real-world scenarios, the task of outlier detection becomes even more challenging as traditional mono-source outlier detection techniques can no longer be suitable for multi-source heterogeneous data. In this paper, a general framework based the consistent representations is proposed to identify multi-source heterogeneous outlier. According to the information compatibility among different sources, Manifold learning are combined in the proposed method to obtain a shared representation space, in which the information-correlated representations are close along manifold while the semantic-complementary instances are close in Euclidean distance. Furthermore, the multi-source outliers can be effectively identified in the affine subspace which is learned through affine combination of shared representations from different sources in the feature-homogeneous space. Comprehensive empirical investigations are presented that confirm the promise of our proposed framework.

Abstract: Anomaly detection is an important data mining method aiming to discover outliers that show significant diversion from their expected behavior. A widely used criteria for determining outliers is based on the number of their neighboring elements, which are referred to as Nearest Neighbors (NN). Existing NN-based Anomaly Detection (NN-AD) algorithms cannot detect streaming outliers, which present time sensitive abnormal behavior characteristics in different time intervals. In this paper, we propose a fast NN-based approach for Time Sensitive Anomaly Detection (NN-TSAD), which can find outliers that present different behavior characteristics, including normal and abnormal characteristics, within different time intervals. The core idea of our proposal is that we combine the model of sliding window with Locality Sensitive Hashing (LSH) to monitor streaming elements distribution as well as the number of their Nearest Neighbors as time progresses. We use an ϵ-approximation scheme to implement the model of sliding window to compute Nearest Neighbors on the fly. We conduct widely experiments to examine our approach for time sensitive anomaly detection using three real-world data sets. The results show that our approach can achieve significant improvement on recall and precision for anomaly detection within different time intervals. Especially, our approach achieves two orders of magnitude improvement on time consumption for streaming anomaly detection, when compared with traditional NN-based anomaly detection algorithms, such as exact-Storm, approx-Storm, MCOD etc, while it only uses 10 percent of memory consumption.

Abstract: Oil and gas reservoirs are distributed across the globe at different depth and geological ages. Although some petroleum deposits are situated spatially far from each other, they may share similar distributions of continuous attributes describing formation and fluid properties as well as categorical attributes describing tectonic regimes and depositional environments. In that case they are called reservoir analogues. Information about thousands of reservoirs from around the world forms a solid basis for uncertainty evaluation and missing data imputation. Besides these routine tasks in the industry, such dataset allows to obtain probabilistic reasoning through frequency analysis. This work presents graphical representation of causal links between geological attributes of reservoir analogues.

Abstract: Statistical extraction of relevant n-grams in natural language corpora is important for text indexing and classication since it can be language independent. We show how a theoretical model identies the distribution properties of the distinct n-grams and singletons appearing in large corpora and how this knowledge contributes to understanding the performance of an n-gram cache system used for extraction of rel- evant terms. We show how this approach allowed us to evaluate the benets from using Bloom lters for excluding singletons and from using static prefetching of nonsingletons in an n-gram cache. In the context of the distributed and parallel implementation of the LocalMaxs extraction method, we analyze the performance of the cache miss ratio and size, and the eciency of n-gram cohesion calculation with LocalMaxs.

Abstract: Blue Waters Workforce Development: Delivering National Scale HPC Workforce Development The National Science Foundation funds the Blue Waters project, which supports an Education, Outreach and Training (EOT) program focused on preparing an HPC-capable workforce with an emphasis on petascale computing competencies. The Blue Waters EOT team engages undergraduate students in internships, graduate students in fellowships, researchers as participants in training sessions, trainers and educators as PIs of education allocations, and underrepresented communities as PIs of broadening participation allocations. All of these communities benefit from access to on one of the most advanced computing environments available to the open science research community. Educators, researchers and students are asked to present their research via conference presentations (e.g. at the annual Blue Waters Symposium) and publications (e.g. the Journal of Computational Science Education). HPC experts offer a variety of training sessions to assist researchers and educators with incorporating state-of-the-art resources, tools, and methods within their research and education endeavors. The Virtual School of Computational Science and Engineering (VSCSE) delivered graduate level computational science and HPC courses to students at colleges and universities across the country, and to students at international locations. The VSCSE courses were led by experts in the field and were conducted in collaboration with their faculty at the recipient institutions in order to provide the students with access to course content and mentoring that would otherwise not have been available to them at their home institution. The Blue Waters Graduate Fellowships have provided a year of funded support and access to the Blue Waters system for the research being conducted by 44 graduate students in a variety of science and engineering fields of study. The Blue Waters Student Internship Program is designed to motivate undergraduate students to pursue advanced science and engineering studies and careers. The undergraduate interns begin their involvement with an intensive two-week parallel programming institute held at NCSA to prepare the students for a full-year of funded research. To date, the internship program has benefitted 120 undergraduate students and resulted in numerous papers being published by the students in the Journal of Computational Science Education (JOCSE) . Blue Waters has added substantial content to a repository of education and training materials, including 30 undergraduate course modules that are applicable to multiple disciplines. The repository facilitates broad dissemination of these and other related materials, which have been used to support workforce development at the high school, undergraduate and graduate levels. Education Allocations The initial proposal for the Blue Waters project requested that 1% of the available computing resources be devoted to educational activities to prepare a larger and more diverse computationally literate workforce. At that time, 1% of the system was a substantial commitment - providing more computational resources than were available to researchers via all of the other NSF funded HPC systems. Education allocations are available to faculty and staff at any US institution to support undergraduate and graduate courses, training sessions, workshops, webinars, institutes, and Research Experiences for Undergraduates (REU) programs. We encourage innovative approaches to educating the community. Requests range from one day training events to a full year program of structured learning, such as through internships and fellowships. Allocation requests typically range from 5,000 to 25,000 node hours, although allocations of larger amounts have been granted for programs serving large numbers of participants or for conducting more complex semester course requirements. Broadening Participation Allocations In order to engage a more diverse community of researchers, the Blue Waters project created a new allocations category for Broadening Participation. The purpose was to encourage principal investigators who were women, minorities or individuals at NSF designated EPSCoR institutions to apply for an allocation of time on the Blue Waters system. These were intended as start-up allocations of up to 200,000 node-hours to allow the research team to scale-up their codes to Blue Waters. Twenty-one teams from around the country were selected in the first year. Included among the Principal Investigators (PIs) are ten females and two underrepresented minorities. In addition, there are four female and eight underrepresented minority colleagues listed as co-PIs. Among the lead institutions, five are Minority Serving Institutions and ten are within EPSCoR jurisdictions. After working on the Blue Waters system for nearly a year, one of these teams received a PRAC allocation from NSF, which will allow the team to significantly advance their computational research. Graduate Fellowships The Blue Waters project offers a very unique federally supported program that provides PhD students with a full year of computational science and engineering research support. Each fellow receives an allocation of 50,000 node-hours to pursue their computational and/or data-enabled research on the Blue Waters system. The fellows also receive a $38,000 stipend. Each fellow is able to request up to $12,000 in tuition allowance to help offset their educational expenses. Each fellow is invited to attend the annual Blue Waters Symposium to make a formal presentation and display a poster to share their research progress with the other attendees. They are also encouraged to give a presentation of their Blue Waters supported research at a domain conference of their choosing. There are between four and 10 fellows selected each year through a very competitive application process that is open to students in all US academic institutions. After their fellowship year ends, the fellows are welcome to continue using the Blue Waters system to pursue their research while completing their PhD. They are also welcome to continue using the system during a subsequent postdoctoral appointment. Many of the fellows have gone on to faculty positions, postdoctoral positions, and professional positions in academia, government agencies and academic institutions. We continue to track the progress of each fellow to facilitate a longitudinal analysis of the impact of the fellowship program. Internships The Blue Waters Student Internship Program is designed to motivate and prepare the next generation of computational researchers by engaging them in year-long research projects. The Internship Program supports about 20 students each year, with a $5,000 stipend spread out over the full year of their appointment. The program welcomes applications from undergraduates at all degree granting US institutions. Each year, the program kicks-off by offering a two-week parallel programming intensive institute at NCSA that also engages the interns in learning to make effective use of the Blue Waters system. Following the institute, the students are matched with mentors to guide them through their year-long research project. The students are matched with a faculty member either on their own campus, or at another campus. Towards the end of their year-long research endeavors, the students apply to present a poster on their research project at the annual Blue Waters Symposium in the May/June timeframe. The faculty report that the combination of a two-week institute and support for a full year have proven to be very effective for the student and the project the students pursue. Journal of Computational Science Education All of the interns are encouraged to publish their research in the peer-reviewed Journal of Computational Science Education (JOCSE). The students are also encouraged to describe their experiences and the impacts of the program on their academic pursuits, and career goals. JOCSE accepts articles from the international community. Articles are encouraged that address the teaching and learning of computational science and engineering, the development and applications of instructional materials, projects, as well as innovative approaches for conducting workforce development. The editors welcome articles that address the assessment of materials or programs, methods for achieving improved learning outcomes, and innovative computational science programs. The journal articles and instructions for submissions are available at http://jocse.org/. Training The Blue Waters project conducts a variety of training events throughout the year to assist participants in learning computational and data enabled science and engineering methods, tools, and resources. The training is designed to prepare participants to make effective use of computing resources, with an emphasis on petascale computing. The training events include webinars, workshops, symposia, tutorials, presentations, hackathons, and other related activities. They are delivered as in-person events, as webcasts and as self-paced tutorials. HPC University Repository The HPC University portal was established to provide a mechanism for disseminating HPC related training and education material. It is built on the foundation and principles established by the Computational Science Education Reference Desk (CSERD) , which is among the collections funded by the National Science Digital Library (NSDL) funded by NSF. The Blue Waters team developed 30 “Undergraduate petascale modules” appropriate for teaching parallel computational modeling to undergraduate or graduate students in science and other STEM disciplines. These modules are among the vast collection of HPC related training and education materials available from this repository. The Blue Waters Symposium The Blue Waters Symposium is an annual gathering of Blue Waters staff, researchers, students, and professionals from among the computational science and engineering community. The Symposia participants share successes and challenges in utilizing large-scale heterogeneous computing systems. Each of the scientific teams using the Blue Waters system are asked to provide updates on their research, and to highlight how the petascale system has helped to advance their research. Nnationally and internationally recognized leaders are invited as keynote speakers to present innovative, impactful and at times controversial ideas that advance knowledge and provoke interactions among the attendees. There are numerous opportunities for the participants to discuss challenges, opportunities, and the future of scientific computing. The discussions often times result in new collaborations and cooperative ventures. Summary The Blue Waters project actively recruits students, faculty, professionals, and mentors in these activities from across the United States, with an emphasis on engaging women, minorities and people with disabilities. Since going into full-service operations in 2013, over 200 education and training allocations have been utilized for activities ranging from one-day workshops to two-week institutes. The Blue Waters project has engaged more than 3,700 people in learning to make effective use of computational and data-enabled science and engineering tools, resources, and methods. The participants in the activities came from 219 academic institutions, of which 65 are within EPSCoR jurisdictions. The impact and benefits have been widespread, including directly reaching people located in many foreign countries, as well as freely disseminating materials that have been downloaded and used by thousands of people world-wide. The Blue Waters project places a high importance on sharing what we have learned to help others to be even more successful in their own endeavors. We look forward to sharing our experiences at the ICCS 2019 Conference and fostering an exchange of lessons learned among the attendees.

Abstract: Combinatorial scientific computing refers to the field of using combinatorial algorithms to solve problems in computational science and data science. Teaching even elementary topics from this area is difficult because it involves bridging the gap between scientific computing and graph theory. Furthermore, it is often necessary to understand not only the methodologies from mathematics and computer science, but also from different scientific domains from which the underlying problems arise. To enrich the learning process in combinatorial scientific computing, we designed and implemented a set of interactive educational modules called EXPLAIN. The central idea behind EXPLAIN is its focus on describing the equivalence of a problem in terms of scientific computing and graph theory. That is, in EXPLAIN, the scientific computing problem and its graph theoretical representation are treated as two sides of the same coin. The process of solving a problem is interactively explored by visualizing transformations on an object from scientific computing, simultaneously, with the corresponding transformations on a suitably defined graph. We describe the redesign of the EXPLAIN software with an emphasis on integrating a domain-specific scripting language and a hierarchical visualization for recursively defined problems.

Abstract: One of the core missions of Michigan State University's new Department of Computational Mathematics, Science, and Engineering is to provide education in computational modeling and data science to MSU's undergraduate and graduate students. In this paper, we describe our creation of CMSE 201, "Introduction to Computational Modeling and Data Analysis," which is intended to be a standalone course teaching students core concepts in data analysis, data visualization, and computational modeling. More broadly, we discuss the education-research-based rationale behind the "flipped classroom" instructional model that we have chosen to use in CMSE 201, which has also informed the design of other courses taught in the department. We also explain the course'€™s design principles and implementation.

Abstract: Today’s world is full of data. Data scientists are needed everywhere to design and implement the processes that analyze the data and turn them into meaningful information. Consequently, it is not surprising that students from all disciplines increasingly feel the need of having to learn how to build software for their solving their data analysis problems. The course "Computational Thinking and Programming in Python" at Utrecht University has been designed for accommodating the needs of these students. Computational thinking is about expressing problems and their solutions in ways that a computer could execute. It is considered one of the fundamental skills of the 21st century. To develop student’s computational thinking skills for data analysis problems, the course covers ways for systematically approaching such problems (CRISP-DM model, reference processes), abstract program description techniques (UML diagrams) and elementary software design principles (reuse, modularization). Programming is the process of designing and building an executable computer program for accomplishing a specific computing task. The course introduces students to programming with Python, which is currently one of the most popular programming languages in data science. After familiarization with the basics, the course addresses more advanced topics, such as access to web services, statistical analyses with the pandas package and data visualization with the matplotlib package. Furthermore, there are some lectures on additional practical topics like the FAIR principles and workflow management systems. Every lecture is accompanied by a practical BYOD lab session where students can work on the weekly homework assignments with support of the teaching assistants. To practice the work with more complex, realistic data analysis problems, students furthermore work on small group projects during the course, and present their results at the end. The presentation will discuss the specific learning goals and design of the course, also in light of practical conditions such as class size and teaching staff available. Furthermore, it will elaborate on the specific challenges involved, experiences and lessons learned that can be beneficial for computational science educators teaching similar courses.

Abstract: Resistivity well logging characterizes the geological formation around a borehole by measuring the electrical resistivity. On logging while drilling techniques, real-time imaging of the well surroundings is decisive to correct the well trajectory in real time for geosteering purposes. Thus, we require numerical methods that rapidly solve Maxwell's equations. In this work, we study the main features and limitations of rIGA to solve borehole resistivity problems. We apply rIGA method to approximate 3D electromagnetic fields that result from solving Maxwell's equations through the 2.5D formulation. We use a spline-based generalization of a H(curl) x H^1 functional space. In particular, we use the H(curl) spaces previously introduced by Buffa et al. to set the high-continuity curl-conforming space discretization.

Abstract: Despite the existence of several hp-adaptive algorithms in the literature (e.g. [1]), very few are used in industrial context due to their high implementational complexity, computational cost, or both. This occurs mainly because of two limitations associated with hp-adaptive methods: (1) The data structures needed to support hp-refined meshes are often complex, and (2) the design of a robust automatic hp-adaptive strategy is challenging. To overcome limitation (1), we adopt the multi-level approach of D’Angela et al. [2]. This method handles hanging nodes via a multilevel technique with massive use of Dirichlet nodes. Our main contribution in this work is intended to overcome limitation (2) by introducing a novel automatic hp-adaptive strategy. For that, we have developed a simple energy-based coarsening approach that takes advantage of the hierarchical structure of the basis functions. Given any grid, the main idea consists in detecting those unknowns that contribute least to the energy norm, and remove them. Once a sufficient level of unrefinement is achieved, a global h, p, or any other type of refinements can be performed. We tested and analyzed our algorithm on one-dimensional (1D) and two- dimensional (2D) benchmark cases. In this presentation, we shall illustrate the main advantages and limitations of the proposed hp-adapted method. References: 1. L. Demkowicz. Computing with hp-adaptive finite elements. Vol. 1. One and two dimensional elliptic and Maxwell problems. Applied Mathematics and Nonlinear Science Series. Chapman & Hall/CRC, Boca Raton, FL, 2007. ISBN 978-1-58488- 671-6; 1-58488-671-4. 2. D. D’Angella, N. Zander, S. Kollmannsberger, F. Frischmann, E. Rank, A. Schröder, and A. Reali. Multi-level hp-adaptivity and explicit error estimation. Advanced Modeling and Simulation in Engineering Sciences, 3(1):33, 2016. ISSN 2213-7467.

Abstract: We present a framework to run Cahn-Hilliard simulations with a web interface. We use a popular Continous Integration tool Jenkins. This setup allows launching computations from any machine and without the need to sustain a connection to the computational environment. Moreover, the results of the computations are automatically post-processed and stored upon job completion for future retrieval in the form of a sequence of bitmaps, and the video illustrating the simulation. We extract the mobility and chemical potential functions from the Cahn-Hilliard equation to the interface, allowing for several numerical applications. The discretization is performed with isogeometric analysis, and it is parameterized with the number of time steps, time step size, mesh dimensions, and the order of the B-splines. The interface is linked with the fast alternating direction semi-implicit solver [1], resulting in a linear computational cost of the simulation.

Abstract: We want to estimate the effective (homogenized) compressional velocity of a highly heterogeneous porous rock at low frequencies. To achieve this goal is necessary to repeat virtually the rock domain several times until it becomes at least two-wavelengths long. Otherwise, boundary conditions (e.g., a PML) pollute the estimated effective velocity. Due to this requirement on the computational domain size, traditional conforming fitting element grids result in a humongous number of elements that cannot be simulated with today's computers. To overcome this problem, we consider non-fitting meshes, in which each finite element includes highly-discontinuous material properties. To maintain accuracy under this scenario, we show it is sufficient to perform exact integration. Being this operation also computationally expensive for such large domains, we precompute the element stiffness matrices. The presence of a PML makes the implementation of this precomputation step more challenging. In this presentation, we illustrate the main challenges for solving this upscaling/homogenization problem, which is of great interest to the oil & gas industry, and we detail the computational techniques employed to overcome them. The performance of the proposed method is also showcased with different numerical experiments.

Abstract: We present a Distributed Memory Parallel (DMP) implementation of agent based economic models, which facilitates large scale simulations with millions of agents. A major obstacle in scalable DMP implementation is balancely distributing the agents among MPI processes, while making all the topological graphs, over which the agents interact, available at a minimum communication cost. We balancely distributed the computational workload among MPI processes by partitioning a representative employer-employee interaction graph, and all the other interaction graphs are made available at negligible communication costs by mimicking the organizations of the real-world's economic entities. Cache friendly and low memory intensive algorithms and data structures are proposed to improve runtime and parallel scalability, and their effectivenesses are demonstrated. It is demonstrated that the current implementation is capable of simulating 1:1 scale models of medium size countries.

Abstract: Concurrent multiscale methods are essential for the understanding and prediction of behavior of engineering systems when a small-scale event will eventually determine the performance of the entire system. Here, we describe the recently-proposed [1] domain-decomposition-based Schwarz alternating method as a means for concurrent multiscale coupling in finite deformation quasistatic and dynamic solid mechanics. The approach is based on the simple idea that if the solution to a partial differential equation is known in two or more regularly shaped domains comprising a more complex domain, these local solutions can be used to iteratively build a solution for the more complex domain. The proposed approach has a number of advantages over competing multiscale coupling methods, most notably its concurrent nature, its ability to couple non-conformal meshes with different element topologies, and its non-intrusive implementation into existing codes. In this talk, we will first overview our original formulation of the Schwarz alternating method for multiscale coupling in the context of quasistatic solid mechanics problems [1]. We will discuss the method's proven convergence properties, and demonstrate its accuracy, convergence and scalability of the proposed Schwarz variants on several quasistatic solid mechanics examples simulated using the Albany/LCM code. The bulk of the talk will present some recent extensions of the Schwarz alternating formulation to dynamic solid mechanics problems [2]. Our dynamic Schwarz formulation is not based on a space-time discretization like other dynamic Schwarz-like methods; instead, it uses a governing time-stepping algorithm that controls time-integrators within each subdomain. As a result, the method is straight-forward to implement into existing codes (e.g, Albany/LCM), and allows the analyst to use different time-integrators with different time steps within each domain. We demonstrate on several test cases (including bolted-joint problems of interest to production) that coupling using the proposed method introduces no dynamic artifacts that are pervasive in other coupling methods (e.g., spurious wave reflections near domain boundaries), regardless of whether the coupling is done with different mesh resolutions, different element types like hexahedral or tetrahedral elements, or even different time integration schemes, like implicit and explicit. Furthermore, on dynamic problems where energy is conserved, we show that the method is able to preserve the property of energy conservation. REFERENCES [1] A. Mota, I. Tezaur, C. Alleman. “The alternating Schwarz method for concurrent multiscale coupling”, Comput. Meth. Appl. Mech. Engng. 319 (2017) 19-51. [2] A. Mota, I. Tezaur, G. Phlipot. "The Schwarz alternating method for dynamic solid mechanics", in preparation for submission to Comput. Meth. Appl. Mech. Engng.

Abstract: The Exascale Computing Project (ECP, https://exascaleproject.org/) is a U.S. Dept. of Energy effort developing hardware, software infrastructure, and applications for computational platforms capable of performing 10^18 floating point operations per second (one “exaop”). The Exascale Additive Manufacturing Project (ExaAM) is one of the applications selected for development of models that would not be possible on even the largest of today’s computational systems. In addition to ORNL, partners include Lawrence Livermore National Laboratory (LLNL), Los Alamos National Laboratory (LANL), the National Institute for Standards and Technology (NIST), as well as key universities such as Purdue Univ., UCLA, and Penn. State Univ. Since we are both leveraging existing simulation software and also developing new capabilities, we will describe the physics components that comprise our simulation environment and report on progress to date using highly-resolved melt pool simulations to inform part-scale finite element thermomechanics simulations, drive microstructure evolution, and determine constitutive mechanical property relationships based on those microstructures using polycrystal plasticity. The coupling of melt pool dynamics and thermal behavior, microstructure evolution, and microscale mechanical properties provides a unique, high-fidelity model of the process-structure-property relationship for additively manufactured parts. We will report on the numerics, implementation, and performance of the nonlinearly consistent coupling strategy, including convergence behavior, sensitivity to fluid flow fidelity, and challenges in timestepping. The ExaAM team includes James Belak, co-PI (LLNL), Nathan Barton (LLNL), Matt Bement (LANL), Curt Bronkhorst (Univ. of Wisc.), Neil Carlson (LANL), Robert Carson (LLNL), Jean-Luc Fattebert (ORNL), Neil Hodge (LLNL), Zach Jibben (LANL), Brandon Lane (NIST), Lyle Levine (NIST), Chris Newman (LANL), Balasubramaniam Radhakrishnan (ORNL), Matt Rolchigo (LLNL), Stuart Slattery (ORNL), and Steve Wopschall (LLNL). This work was supported by the Exascale Computing Project (17-SC-20-SC), a collaborative effort of the U.S. Department of Energy Office of Science and the National Nuclear Security Administration.

Abstract: We introduce a new parallelizable numerical multiscale method for advection-dominated problems as they often occur in engineering and geosciences. State of the art multiscale simulation methods work well in situations in which stationary and elliptic scenarios prevail but are prone to fail when the model involves dominant lower order terms which is common in applications. We suggest to overcome the associated difficulties through a reconstruction of subgrid variations into a modified basis by solving many independent (local) inverse problems that are constructed in a semi-Lagrangian step. Globally the method looks like a Eulerian method with multiscale stabilized basis. The method is extensible to other types of Galerkin methods, higher dimensions, nonlinear problems and can potentially work with real data. We provide examples inspired by tracer transport in climate systems in one and two dimensions and numerically compare our method to standard methods.

Abstract: Galerkin projection provides a formal means to project a differential equation onto a set of preselected basis functions. This may be done for the purpose of formulating a numerical method, as in the case of spectral methods, or formulation of a reduced-order model (ROM) for a complex system. Here, a new method is proposed in which the basis functions used in the projection process are determined from an asymptotic (perturbation) analysis. These asymptotic basis functions (ABF) are obtained from the governing equation itself; therefore, they contain physical information about the system and its dependence on parameters contained within the mathematical formulation. We refer to this as reduced-physics modeling (RPM) as the basis functions are obtained from a physical, i.e.\ model-driven, rather than data-driven, approach. Therefore, the ABF hold the potential to provide an accurate RPM of a system that captures the physical dependence on model parameters and is accurate over a wider range of parameters than possible for traditional ROM methods. This new approach is tailor-made for modeling multiscale problems as the various scales, whether overlapping or distinct in time or space, are formally accounted for in the ABF. A regular-perturbation problem is used to illustrate that projection of the governing equations onto the ABF allows for determination of accurate approximate solutions for values of the ``small'' parameter that are much larger than possible with the asymptotic expansion alone.

Abstract: A mathematical model of the flow of a polyatomic gas containing a combination of the Navier-Stokes-Fourier model (NSF) and the model kinetic equation of polyatomic gases is presented. At the heart of the hybrid components is a unified physical model, as a result of which the NSF model is a strict first approximation of the model kinetic equation. The model allows calculations of flow fields in a wide range of Knudsen numbers (Kn), as well as fields containing regions of high dynamic nonequilibrium. The boundary conditions on a solid surface are set at the kinetic level, which allows, in particular, to formulate the boundary conditions on the surfaces absorbing or emitting gas. The hybrid model was tested. The example of the problem of the shock wave profile shows that up to Mach numbers near 2 the combined model gives smooth solutions even in those cases where the sewing point is in a high gradient region. For the Couette flow, smooth solutions are obtained at M=5, Kn=0.2. As a result of research, a weak and insignificant difference between the kinetic region of the hybrid model and the “pure” kinetic model was established. A model effect was discovered: in the region of high nonequilibrium, there is an almost complete coincidence of the solutions of the kinetic region of the combined model and the “pure” kinetic solution. This work was conducted with the financial support of the Ministry of Education and Science of the Russian Federation, project №9.7170.2017/8.9.

Abstract: With the rapid advancement of intelligent agriculture, the application of IoT technology in aquaculture is becoming more and more widespread. In this process, there's a lot of structured, semi-structured, unstructured data. On the one hand, traditional relational database management systems cannot store this data flexibly and scalably. On the other hand, the stream data generated by the sensor usually requires a flow processing operation with high throughput, low latency and high performance.Therefore, based on Flink, MongoDB and Kafka, this paper proposes and implements an aquaculture monitoring system. Among them, Flink platform provides high throughput, low latency and high performance stream processing as a stream data processing platform. Kafka, as a distributed publish-subscribe message system, acquires different sensor data and builds reliable pipelines for transmitting real-time data between application programs. MongoDB stores sensor data in different formats. Finally,as a highly reliable and high-performance column database, HBase is often used in sensor data storage schemes. So, this paper presents a performance evaluation on how efficiently MongoDB and HBase insertions and queries perform. The experimental results show that the efficiency of MongoDB was much higher than that of HBase, which provided a feasible solution for the sensor data storage and processing of aquaculture.

Abstract: Hydroponic cultivation is an agricultural method where nutrients are efficiently provided as mineral nutrient solutions. This modern agriculture sector provides numerous advantages such as efficient location and space requirements, adequate climate control, water-saving and controlled nutri-ents usage. The Internet of things (IoT) concept assumes that various “things,” which include not only communication devices but also every other physical object on the planet, are going to be connected and will be controlled across the Internet. Mobile computing technologies in general and mobile applications, in particular, can be assumed as significant meth-odologies to handle data analytics and data visualisation. Using IoT and mobile computing is possible to develop automatic systems for enhanced hydroponic agriculture environmental monitoring. Therefore, this paper presents an IoT monitoring system for hydroponics named iHydroIoT. The solution is composed of a prototype for data collection and an iOS mobile application for data consulting and real-time analytics. The collected data is stored using Plotly, a data analytics and visualisation library. The proposed system provides not only temporal changes monitoring of light, tempera-ture, humidity, CO2, pH and electroconductivity but also water level for enhanced hydroponic supervision solutions. The iHydroIoT offers real-time notifications to alert the hydroponic farm manager when the condi-tions are not favourable. Therefore, the system is a valuable tool for hydro-ponics condition analytics and to support decision making on possible in-tervention to increase productivity. The results reveal that the system can generate a viable hydroponics appraisal, allowing to anticipate technical interventions that improve agricultural productivity.

Abstract: Environmental noise pollution has a significant impact on health. The noise effects on health are related to annoyance, sleep and cognitive performance for both adults and children are reported in the literature. The smart city concept can be assumed as a strategy to mitigate the problems generated by the urban population growth and rapid urbanisation. Noise mapping is an important step for noise pollution reduction. Although, noise maps are particularly time-consuming and costly to create as they are produced with standard methodologies and are based on specific sources such as road traffic, railway traffic, aircraft and industrial. Therefore, the actual noise maps are significantly imperfect because the noise emission models and sources are extremely limited. Smartphones have incredible processing capabilities as well as several powerful sensors such as microphone and GPS. Using the resources present in a smartphone as long with participatory sensing, a crowdsourcing noise mobile application can be used to provide environmental noise supervision for enhanced living environments. Crowdsourcing techniques applied to environmental noise monitoring allow creating reliable noise maps at low-cost. This paper presents a mobile crowdsourcing solution for environmental noise monitoring named iNoiseMapping. The environmental noise data is collected through participatory sensing and stored for further analysis. The results obtained can ensure that mobile crowdsourcing offers several enhanced features for environmental noise supervision and analytics. Consequently, this mobile application is a significant decision-making tool to plan interventions for noise pollution reduction.

Abstract: Indoor environment quality (IEQ) has a significant impact on all human activities. Temperature and humidity are extremely important not only for enhanced living environments but particularly for supervising laboratory activities. On the one hand, laboratories are spaces characterised by numerous pollution sources that can lead to relevant unhealthy indoor environments. The laboratory activities such as the case of thermography experiments require real-time monitoring supervision. On the other hand, buildings are responsible for about 40% of the global energy consumption, and over 30% of the CO2 emissions; also, a considerable proportion of this energy is used for thermal comfort. The IBM Watson IoT Platform is a fully managed, cloud-hosted service for the Internet of Things (IoT) that allows data to be sent securely to the cloud using MQTT messaging protocol. This paper aims to present an IoT solution for indoor temperature and humidity real-time supervision named iTemp+. The solution is composed by a hardware prototype for ambient data collection and use IBM Watson IoT for data storing and consulting. The IBM Watson IoT Platform provides data integration, security methods, data collection, visualization, analytics, device management, artificial intelligence and blockchain functionalities which are not implemented in the concurrent IoT platforms. The results obtained reveal that IBM Watson IoT platform offers several enhanced features for device management and analytics and can be used as a powerful approach to provide IEQ supervision.

Abstract: With rapid advancement in exploration and production of shale gas reservoirs over the world, the fast simulation of shale gas flows that is required especially in engineering applications has attracted extensive attentions from the engineering and academic communities. In this study, we apply a popular global model reduction method, proper orthogonal decomposition (POD), to speed up the simulation of shale gas reservoirs. However, different from incompressible fluid flows, the compressibility of shale gas induces additional challenges to construct accurate and efficient POD reduced-order model (ROM). First, the compressibility of shale gas increases the nonlinearity of the flow system, the dimension of the projected Darcy-type pressure equation in low-dimensional space still depends on the dimension of the original system, which complicates the computation and worsens the acceleration of the POD-ROM substantially. Second, due to the POD projection term containing the compressibility of shale gas, the additional computational cost is needed to solve the equation of state (EOS) of shale gas to obtain the compressibility factor. To handle these problems, we adopt another model reduction approach, discrete empirical interpolation method (DEIM), to approximate the nonlinearity in pressure equation by only using few selected representative interpolation points over the domain, thus the nonlinearity of the variables and the computation of EOS can be greatly reduced. Combined the POD-ROM with DEIM, a POD-DEIM-ROM for fast simulation of shale gas flows with different gas states in single-continuum porous media is developed. The performances of the proposed model are validated by comparing to the traditional method without acceleration techniques through different numerical cases, the computational speed and numerical accuracy are analyzed in detail. Results show that the proposed model can achieve great acceleration of the simulation (two orders of magnitude) without sacrificing the numerical accuracy obviously. Especially, the influence of the type of EOS (such as ideal gas EOS, Van der Waals EOS, and Peng–Robinson EOS, etc.), the number of POD modes and interpolation points, as well as the permeability filed distribution on the overall performance of the proposed model are investigated.

Abstract: A direct numerical simulation (DNS) code was developed for solving incompressible homogeneous isotropic turbulence with high Reynolds numbers in a periodic box using the Fourier spectral method. The code was parallelized using the Message Passing Interface and OpenMP with two-directional domain decomposition and optimized on the K computer. High resolution DNSs with up to $12288^3$ grid points were performed on the K computer using the code. Efficiencies of 3.84\%, 3.14\%, and 2.24\% peak performance were obtained in double precision DNSs with $6144^3$, $8192^3$, and $12288^3$ grid points, respectively. In addition, a two-path alias-free procedure is proposed and clarified its effectiveness for some number of parallel processes.

Abstract: This work presents Direct Numerical Simulation of mass transfer from buoyancy-driven bubbles rising in a wall-confined vertical channel, by means of a multiple marker level-set method. The Navier-Stokes equations and mass transfer equations are discretized using a finite-volume method on a collocated unstructured mesh, whereas a multiple marker approach is used to avoid the numerical coalescence of bubbles. This approach is implemented in the framework of a mass conservative level-set method, whereas unstructured flux-limiter schemes are used to discretize the convective term of momentum equation, level-set equations, and mass transfer equation, in order to improve the stability of the solver in bubbly flows with high Reynolds number and high-density ratio. The capabilities of this model are proved in the buoyancy-driven motion of single bubbles and bubble swarms in a vertical channel of circular cross-section.

Abstract: A hybrid particle/mesh Vortex Method, called remeshed vortex method, is proposed in this work to simulate three-dimensional incompressible flows. After a validation study of the present method in the context of Direct Numerical Simulations, an anisotropic artificial viscosity model is proposed in this paper in order to handle bi-level simulations. The bi-level approach implies two different mesh sizes for the discretization of the two coupled variables of the problem, namely the vorticity field and the velocity field: the vorticity field is computed on a fine grid while the velocity field is solved on a coarse grid. This approch is proposed with the objective of performing efficient computations based on hybrid GPU-CPU architectures.

Abstract: Aspect Based Sentiment Analysis is a type of fine-grained sentiment analysis. It is popular in both industry and academic communities, since it provides more detailed information on the user generated text in product reviews or social network. We propose a novel framework based on neural network to determine the polarity of a review given a specific target. Not only the words close to the target but also the words far from the target determine the polarity of the review given a certain target, so we use self attention to solve the problem of long distance dependence. Briefly, we do multiple linear mapping on the review, do multiple attention and combine them to attend to the information from different representation sub-spaces. Besides, we use domain embedding to get close to the real word embedding in a certain domain, since the meaning of the same word may be different in different situation. Moreover, we use position embedding to underline the target and pay more attention to the words that are close to the target to get better performance on the task. We validate our model on four benchmarks, they are SemEval 2014 restaurant dataset, SemEval 2014 laptop dataset, SemEval 2015 restaurant dataset and SemEval 2016 restaurant dataset. The final results show that our model is effective and strong, which brings a 0.74% boost averagely based on the previous state-of-the-art work.

Abstract: Law enforcement agencies and private security companies work to prevent, detect and counteract any threat with the resources they have, including alarms and video surveillance. Even so, there are still terrorist attacks or shootings in schools in which armed people move around a venue exercising violence and generating victims, showing the limitations of current systems. For example, they force security agents to monitor continuously all the images coming from the installed cameras, and potential victims nearby are not aware of the danger until someone triggers a general alarm, which also does not give them information on what to do to protect themselves. In this article we present a project that is being developed to apply the latest technologies in early threat detection and optimal response. The system is based on the automatic processing of video surveillance images to detect weapons and a mobile app that serves both for detection through the analysis of mobile device sensors, and to send users personalised and dynamic indications. The objective is to react in the shortest possible time and minimise the damage suffered.

Abstract: This paper is devoted to new aspects of Natural Language Interface to Relational Database (NLIDB) integration into third party corporate environments related to control data access. Because there is no schema information in the input NL-query and the different relational database management system (RDBMS) requires different meta-data types and rules to control data access, developers meet a problem addressed to automatic data access control in the case of NL-interface implementation to relational databases. In the paper we suggest a comprehensive approach which takes into account permissions throughout the pipeline of transforming NL-query into SQL-query with an intermediate SPARQL representation. Our integration solutions based on well-known Ontology Based Data Access (OBDA) approach, which gives us the opportunity to adapt the proposed solutions to the specifics of data access control in different RDBMS. Suggested approach has been implemented within intelligent service named Reply and tested in the real-world projects.

Abstract: Cross-modal retrieval has attracted increasing attention in recent years. Deep supervised hashing methods have been widely used for cross-modal similarity retrieval on large-scale datasets, because the deep architectures can generate more discriminative feature representations. Traditional hash methods adopt a symmetric way to learn the hash function for both query points and database points. However, those methods take an immense amount of work and time for model training, which is inefficient with the explosive growth of data volume. To solve this issue, we propose an Asymmetric Deep Cross-modal Hashing (ADCH) method to perform more effective hash learning by simultaneously preserving the semantic similarity and the underlying data structures. More specifically, ADCH treats the query points and database points in an asymmetric way. Furthermore, to provide more similarity information, a detailed definition for cross-modal similarity matrix is also proposed. The training of ADCH takes less time than traditional symmetric deep supervised hashing methods. Extensive experiments on two widely used datasets show that the proposed approach achieves the state-of-the-art performance in cross-modal retrieval.

Abstract: In Dial-a-Ride Problem (DARP) customers request from an operator a transportation service from a pick-up to a drop-off place. Depending on the formulation, the problem can address several constraints, being associated with problems such as door-to-door transportation for elderly / disabled people or occasional private drivers. This paper addresses the latter case where a private drivers company transports passengers in a heterogeneous fleet of saloons, estates, people carriers and minibuses. The problem is formulated as a multiple objective DARP which tries to minimize the total distances, the number of empty seats, and the wage differential between the drivers. To solve the problem a Non-dominated Sorting Genetic Algorithm-II is hybridized with a local search. Results for daily scheduling are shown.