Benchmark Experimental Data Sets Research Articles

Cyberbullying Detection on Twitter Using Deep Learning-Based Attention Mechanisms and Continuous Bag of Words Feature Extraction

Since social media platforms are widely used and popular, they have given us more opportunities than we can even imagine. Despite all of the known benefits, some users may abuse these opportunities to humiliate, insult, bully, and harass other people. This issue explains why there is a need to reduce such negative activities and create a safe cyberspace for innocent people by detecting cyberbullying activity. This study provides a comparative analysis of deep learning methods used to test and evaluate their effectiveness regarding a well-known global Twitter dataset. To recognize abusive tweets and overcome existing challenges, attention-based deep learning methods are introduced. The word2vec with CBOW concatenated formed the weights included in the embedding layer and was used to extract the features. The feature vector was input into a convolution and pooling mechanism, reducing the feature dimensionality while learning the position-invariant of the offensive words. A SoftMax function predicts feature classification. Using benchmark experimental datasets and well-known evaluation measures, the convolutional neural network model with attention-based long- and short-term memory was found to outperform other DL methods. The proposed cyberbullying detection methods were evaluated using benchmark experimental datasets and well-known evaluation measures. Finally, the results demonstrated the superiority of the attention-based 1D convolutional long short-term memory (Conv1DLSTM) classifier over the other implemented methods.

Simplicial Complex-Enhanced Manifold Embedding of Spatiotemporal Data for Structural Health Monitoring

Structural Health Monitoring requires the continuous assessment of a structure’s operational conditions, which involves the collection and analysis of a large amount of data in both spatial and temporal domains. Conventionally, both data-driven and physics-based models for structural damage detection have relied on handcrafted features, which are susceptible to the practitioner’s expertise and experience in feature selection. The limitations of handcrafted features stem from the potential for information loss during the extraction of high-dimensional spatiotemporal data collected from the sensing system. To address this challenge, this paper proposes a novel, automated structural damage detection technique called Simplicial Complex Enhanced Manifold Embedding (SCEME). The key innovation of SCEME is the reduction of dimensions in both the temporal and spatial domains for efficient and information-preserving feature extraction. This is achieved by constructing a simplicial complex for each signal and using the resulting topological invariants as key features in the temporal domain. Subsequently, curvature-enhanced topological manifold embedding is performed for spatial dimension reduction. The proposed methodology effectively represents both intra-series and inter-series correlations in the low-dimensional embeddings, making it useful for classification and visualization. Numerical simulations and two benchmark experimental datasets validate the high accuracy of the proposed method in classifying different damage scenarios and preserving useful information for structural identification. It is especially beneficial for structural damage detection using complex data with high spatial and temporal dimensions and large uncertainties in reality.

Post-Authorship Attribution Using Regularized Deep Neural Network

Post-authorship attribution is a scientific process of using stylometric features to identify the genuine writer of an online text snippet such as an email, blog, forum post, or chat log. It has useful applications in manifold domains, for instance, in a verification process to proactively detect misogynistic, misandrist, xenophobic, and abusive posts on the internet or social networks. The process assumes that texts can be characterized by sequences of words that agglutinate the functional and content lyrics of a writer. However, defining an appropriate characterization of text to capture the unique writing style of an author is a complex endeavor in the discipline of computational linguistics. Moreover, posts are typically short texts with obfuscating vocabularies that might impact the accuracy of authorship attribution. The vocabularies include idioms, onomatopoeias, homophones, phonemes, synonyms, acronyms, anaphora, and polysemy. The method of the regularized deep neural network (RDNN) is introduced in this paper to circumvent the intrinsic challenges of post-authorship attribution. It is based on a convolutional neural network, bidirectional long short-term memory encoder, and distributed highway network. The neural network was used to extract lexical stylometric features that are fed into the bidirectional encoder to extract a syntactic feature-vector representation. The feature vector was then supplied as input to the distributed high networks for regularization to minimize the network-generalization error. The regularized feature vector was ultimately passed to the bidirectional decoder to learn the writing style of an author. The feature-classification layer consists of a fully connected network and a SoftMax function to make the prediction. The RDNN method was tested against thirteen state-of-the-art methods using four benchmark experimental datasets to validate its performance. Experimental results have demonstrated the effectiveness of the method when compared to the existing state-of-the-art methods on three datasets while producing comparable results on one dataset.

Multi-harmonic phase demodulation method for instantaneous angular speed estimation using harmonic weighting

Phase demodulation is arguably the most used technique for the estimation of the instantaneous angular speed from vibration signals measured on rotating machinery. Although phase demodulation offers a straightforward approach to determine accurately the rotation speed of a particular shaft in a rotating machine, it does have strict limitations that can hinder the estimation accuracy and its overall reliability. In general, phase demodulation relies on the presence of a single harmonic with a high signal-to-noise ratio for the full duration of the measured vibration signal. Such a precondition hinders its applicability for a generic speed estimation approach for rotating machinery. There are copious real-world scenarios where this requirement is not met, especially given the increasing complexity and dynamic operating range of rotating machines nowadays. In such complex rotating systems, the deterministic signals are not necessarily all harmonically related to a fundamental harmonic. The presence of crossing harmonic orders from non-synchronous rotating components or of lightly damped structural resonances can significantly skew the instantaneous phase demodulation. This paper proposes a novel approach towards phase demodulation by both incorporating multiple harmonics in the demodulation process and also giving the individual harmonic phases time-dependent weighting. This combination of incorporating multiple harmonics and weighting them allows for fully exploiting the information contained within the vibration signal while also promising to be more robust to changing operating regimes. The proposed multi-harmonic demodulation method is investigated thoroughly by assessing its performance on simulated data, on two benchmark experimental data sets, and on a large wind turbine vibration data set consisting of thousands of vibration measurements.

Constitutive Model for Brittle Granular Materials Considering Competition between Breakage and Dilation

A constitutive model is presented for brittle granular materials based on a recent reformulation of the breakage mechanics theory. Compared with previous breakage mechanics-based models, the proposed model is improved to capture strain softening towards the critical state following the peak stress observed in dense specimens under shearing, and simultaneous evolution of breakage and dilation. Considering the competition between dilation and particle breakage allows the model to capture breakage-induced reduction in dilatancy and peak strength as confining pressure increases. The influence of the model parameters on the overall material response is described through a detailed calibration procedure based on a benchmark experimental dataset. Comparison of the results of drained triaxial compression experiments on two sands with the predictions of the model indicates that the enriched model successfully captures strain softening in dilatant specimens, the shearing-driven evolution of stress-strain behavior towards the critical state at different confinement levels, the transition from dilatant to compactive volumetric response, and the evolution of particle grading due to distributed breakage events. The proposed framework is capable of qualitatively reproducing the experimentally observed stress-dilatancy-breakage relationship in brittle granular materials in the low pressure regime.

Refined Empirical Force Field to Model Protein-Self-Assembled Monolayer Interactions Based on AMBER14 and GAFF.

Understanding protein interaction with material surfaces is important for the development of nanotechnological devices. The structures and dynamics of proteins can be studied via molecular dynamics (MD) if the protein-surface interactions can be accurately modeled. To answer this question, we computed the adsorption free energies of peptides (representing eleven different amino acids) on a hydrophobic self-assembled monolayer (CH3-SAM) and compared them to the benchmark experimental data set. Our result revealed that existing biomolecular force fields, GAFF and AMBER ff14sb, cannot reproduce the experimental peptide adsorption free energies by Wei and Latour (Langmuir, 2009, 25, 5637-5646). To obtain the improved force fields, we systematically tuned the Lennard-Jones parameters of selected amino acid sidechains and the functional group of SAM with repeated metadynamics and umbrella sampling simulations. The final parameter set has yielded a significant improvement in the free energy values with R = 0.83 and MSE = 0.65 kcal/mol. We applied the refined force field to predict the initial adsorption orientation of lysozyme on CH3-SAM. Two major orientations-face-down and face-up-were predicted. Our analysis on the protein structure, solvent accessible surface area, and binding of native ligand NAG3 suggested that lysozyme in the face-up orientation can remain active after initial adsorption. However, because of its weaker affinity (ΔΔG = 7.86 kcal/mol) for the ligand, the bioactivity of the protein is expected to reduce. Our work facilitates the use of MD for the study of protein-SAM systems. The refined force field compatible with GROMACS is available at https://cbbio.cis.um.edu.mo/software/SAMFF .

Inter-Laboratory Characterization of the Velocity Field in the FDA Blood Pump Model Using Particle Image Velocimetry (PIV).

A credible computational fluid dynamics (CFD) model can play a meaningful role in evaluating the safety and performance of medical devices. A key step towards establishing model credibility is to first validate CFD models with benchmark experimental datasets to minimize model-form errors before applying the credibility assessment process to more complex medical devices. However, validation studies to establish benchmark datasets can be cost prohibitive and difficult to perform. The goal of this initiative sponsored by the U.S. Food and Drug Administration is to generate validation data for a simplified centrifugal pump that mimics blood flow characteristics commonly observed in ventricular assist devices. The centrifugal blood pump model was made from clear acrylic and included an impeller, with four equally spaced, straight blades, supported by mechanical bearings. Particle Image Velocimetry (PIV) measurements were performed at several locations throughout the pump by three independent laboratories. A standard protocol was developed for the experiments to ensure that the flow conditions were comparable and to minimize systematic errors during PIV image acquisition and processing. Velocity fields were extracted at the pump entrance, blade passage area, back gap region, and at the outlet diffuser regions. A Newtonian blood analog fluid composed of sodium iodide, glycerin, and water was used as the working fluid. Velocity measurements were made for six different pump flow conditions, with the blood-equivalent flow rate ranging between 2.5 and 7 L/min for pump speeds of 2500 and 3500rpm. Mean intra- and inter-laboratory variabilities in velocity were ~ 10% at the majority of the measurement locations inside the pump. However, the inter-laboratory variability increased to more than ~ 30% in the exit diffuser region. The variability between the three laboratories for the peak velocity magnitude in the diffuser region ranged from 5 to 25%. The bulk velocity field near the impeller changed proportionally with the rotational speed but was relatively unaffected by the pump flow rate. In contrast, flow in the exit diffuser region was sensitive to both the flow rate and the rotational speed. Specifically, at 3500rpm, the exit jet tilted toward the inner wall of the diffuser at a flow rate of 2.5 L/min, but the jet tilted towards the outer wall when the flow rate was 7 L/min. Inter-laboratory experimental mean velocity data (and the corresponding variance) were obtained for the FDA pump model and are available for download at https://nciphub.org/wiki/FDA_CFD . Experimental datasets from the inter-laboratory characterization of benchmark flow models, including the blood pump model presented herein and our previous nozzle model, can be used for validating future CFD studies and to collaboratively develop guidelines on best practices for verification, validation, uncertainty quantification, and credibility assessment of CFD simulations in the evaluation of medical devices (e.g. ASME V&V 40 standards working group).

Trans-Allelic Model for Prediction of Peptide:MHC-II Interactions.

Major histocompatibility complex class two (MHC-II) molecules are trans-membrane proteins and key components of the cellular immune system. Upon recognition of foreign peptides expressed on the MHC-II binding groove, CD4+ T cells mount an immune response against invading pathogens. Therefore, mechanistic identification and knowledge of physicochemical features that govern interactions between peptides and MHC-II molecules is useful for the design of effective epitope-based vaccines, as well as for understanding of immune responses. In this article, we present a comprehensive trans-allelic prediction model, a generalized version of our previous biophysical model, that can predict peptide interactions for all three human MHC-II loci (HLA-DR, HLA-DP, and HLA-DQ), using both peptide sequence data and structural information of MHC-II molecules. The advantage of this approach over other machine learning models is that it offers a simple and plausible physical explanation for peptide–MHC-II interactions. We train the model using a benchmark experimental dataset and measure its predictive performance using novel data. Despite its relative simplicity, we find that the model has comparable performance to the state-of-the-art method, the NetMHCIIpan method. Focusing on the physical basis of peptide–MHC binding, we find support for previous theoretical predictions about the contributions of certain binding pockets to the binding energy. In addition, we find that binding pocket P5 of HLA-DP, which was not previously considered as a primary anchor, does make strong contribution to the binding energy. Together, the results indicate that our model can serve as a useful complement to alternative approaches to predicting peptide–MHC interactions.

A Benchmarking Study of Peptide–Biomineral Interactions

A long-standing goal in the field of biomineralization has been to achieve a molecular-level mechanistic understanding of how proteins participate in the nucleation and growth of inorganic crystals (both in vitro and in vivo). Computational methods offer an approach to explore these interactions and propose mechanisms at the atomic scale; however, to have confidence in the predictions of a computational method, the method must first be validated against a benchmark experimental data set of protein–mineral interactions. Relatively little work has been done to test the ability of computation to reproduce experimental results on mineral systems with biologically relevant additives present. The goal of this work is to develop a standard and varied benchmark to test whether a computational method is able to match experimental results at the length and time scales of biomineral–peptide interactions. We compare the results of the RosettaSurface algorithm to an experimental benchmark of kinetic and thermodynamic ...

Modelling the role of self-weight consolidation on the morphodynamics of accretional mudflats

We develop a consolidation module and merge it into a morphodynamic model to assess the role of consolidation on estuarine morphodynamics. We test the model using two settings: point models without hydrodynamic forcing to validate against two benchmark experimental datasets; and a profile model to simulate a mudflat restoration. The modelled self-weight consolidation influences the simulations by gradually reducing the bed level and decreasing the bed erodibility (i.e., increasing the critical bed shear stress). Both effects modify sediment transport processes on mudflats, leading to long-term morphodynamic effects. Depending on the initial bathymetry, the hydrodynamic forcing and the soil properties, the simulated morphological change of the restored mudflat may differ considerably with and without considering consolidation. The consolidation model developed can be utilised to assess the medium to long term effects related to estuarine development (e.g., wetland restoration) and aims to be a publicly available tool.

A Multi‐Ionic Continuum‐Based Model for C3S Hydration

A computationally efficient strategy for modeling tricalcium silicate hydration based on through‐solution‐phase kinetics is demonstrated. This study extends a recently introduced advanced continuum‐based single particle model by including rigorous multi‐ionic transport, nonlinear reversible reaction kinetics and portlandite precipitation. Model parameters were either fixed based on known values, estimated using experimental measures or extracted by model fitting to benchmark experimental datasets. The model is now able to generate calorimetric hydration and evolution of pore solution chemistry responses that are in good agreement with available experimental results and predictions of other multiphysical modeling platforms. Oncecalibrated, the model was tested to see if it could predict the effect of water to cement ratio (w/c) and particle size on hydration outcomes. The findings support the need for a mechanism that limits the volume into which product can form.

Particle migration analysis in iterative classification of cryo-EM single-particle data.

Recently developed classification methods have enabled resolving multiple biological structures from cryo-EM data collected on heterogeneous biological samples. However, there remains the problem of how to base the decisions in the classification on the statistics of the cryo-EM data, to reduce the subjectivity in the process. Here, we propose a quantitative analysis to determine the iteration of convergence and the number of distinguishable classes, based on the statistics of the single particles in an iterative classification scheme. We start the classification with more number of classes than anticipated based on prior knowledge, and then combine the classes that yield similar reconstructions. The classes yielding similar reconstructions can be identified from the migrating particles (jumpers) during consecutive iterations after the iteration of convergence. We therefore termed the method "jumper analysis", and applied it to the output of RELION 3D classification of a benchmark experimental dataset. This work is a step forward toward fully automated single-particle reconstruction and classification of cryo-EM data.

Thermal Nanofluid Property Model With Application to Nanofluid Flow in a Parallel-Disk System—Part I: A New Thermal Conductivity Model for Nanofluid Flow

This is a two-part paper, which proposes a new theory explaining the experimentally observed enhancement of the thermal conductivity, knf, of nanofluids (Part I) and discusses simulation results of nanofluid flow in a radial parallel-plate channel using different knf-models (Part II). Specifically, Part I provides the derivation of the new model as well as comparisons with benchmark experimental data sets and other theories, focusing mainly on aluminum and copper oxide nanoparticles in water. The new thermal conductivity expression consists of a base-fluid static part, kbf, and a new “micromixing” part, kmm, i.e., knf = kbf + kmm. While kbf relies on Maxwell’s theory, kmm encapsulates nanoparticle characteristics and liquid properties as well as Brownian-motion induced nanoparticle fluctuations, nanoparticle volume fractions, mixture-temperature changes, particle–particle interactions, and random temperature fluctuations causing liquid-particle interactions. Thus, fundamental physics principles include the Brownian-motion effect, an extended Langevin equation with scaled interaction forces, and a turbulence-inspired heat transfer equation. The new model predicts experimental data for several types of metal-oxide nanoparticles (20 < dp < 50 nm) in water with volume fractions up to 5% and mixture temperatures below 350 K. While the three competitive theories considered match selectively experimental data, their needs for curve-fitted functions and arbitrary parameters make these models not generally applicable. The new theory can be readily extended to accommodate other types of nanoparticle-liquid pairings and to include nonspherical nanomaterial.

Laminar‐to‐turbulent fluid–nanoparticle dynamics simulations: Model comparisons and nanoparticle‐deposition applications

SUMMARYRelying on benchmark experimental data sets for flow in conduits with local constrictions, LES and three widely used RANS turbulence models, i.e. the low Reynolds number (LRN) k‐ω model, standard k‐ω model and shear stress transport (SST) transition model, were compared and evaluated to gain new physical insight and provide useful turbulence modeling information. These two geometric test cases may represent stenosed arteries and a segment of the human upper airways where the velocity fields undergo all flow regimes, i.e. from laminar, via transitional, to fully turbulent.The comparison study revealed that the standard k‐ω models amplify the flow instabilities after the constrictions, and hence fail to capture the laminar flow behavior at relatively LRNs. The overall performances of LES, the LRN k‐ω model and SST transition model do not have measurable differences in predicting laminar flows and transition to turbulent flow, while the SST transition model may give a better prediction of turbulence kinetic energy profiles in some cases. Clearly, LES can provide instantaneous velocity fluctuations, which may be significant for turbulent micron particle transport/deposition in the respiratory tract. However, it requires 100‐fold more computational time than RANS turbulence models. The use of different turbulence models has a minor effect on nanoparticle deposition in human upper airways when the inspiratory flow rate is low, say, Q = 10L/min. The relative difference for deposition fraction (DF) of nanoparticles with dp>10nm is measurable at a medium inhalation flow rate (say, Q = 30L/min) when employing different turbulence models. However, the absolute difference in DFs is within 0.5% for all‐sized nanoparticles (i.e. 1nm⩽dp⩽50nm) because the DF in the oral airway is very low (say, <1.5%) when 10 nm and 10 L/min.The modeling and simulation information provided are most useful for computational fluid–particle dynamics practitioners to obtain accurate lung deposition concentrations of inhaled toxic or therapeutic nanoparticles. The physical insight provided sheds additional light on the laminar‐to‐turbulent airflow and nanoparticle transport/deposition in locally constricted conduits. Copyright © 2011 John Wiley & Sons, Ltd.

Experimental and theoretical studies of nanofluid thermal conductivity enhancement: a review

Nanofluids, i.e., well-dispersed (metallic) nanoparticles at low- volume fractions in liquids, may enhance the mixture's thermal conductivity, knf, over the base-fluid values. Thus, they are potentially useful for advanced cooling of micro-systems. Focusing mainly on dilute suspensions of well-dispersed spherical nanoparticles in water or ethylene glycol, recent experimental observations, associated measurement techniques, and new theories as well as useful correlations have been reviewed.It is evident that key questions still linger concerning the best nanoparticle-and-liquid pairing and conditioning, reliable measurements of achievable knf values, and easy-to-use, physically sound computer models which fully describe the particle dynamics and heat transfer of nanofluids. At present, experimental data and measurement methods are lacking consistency. In fact, debates on whether the anomalous enhancement is real or not endure, as well as discussions on what are repeatable correlations between knf and temperature, nanoparticle size/shape, and aggregation state. Clearly, benchmark experiments are needed, using the same nanofluids subject to different measurement methods. Such outcomes would validate new, minimally intrusive techniques and verify the reproducibility of experimental results. Dynamic knf models, assuming non-interacting metallic nano-spheres, postulate an enhancement above the classical Maxwell theory and thereby provide potentially additional physical insight. Clearly, it will be necessary to consider not only one possible mechanism but combine several mechanisms and compare predictive results to new benchmark experimental data sets.

Benchmark Experimental Data Set and Assessment of Adsorption Free Energy for Peptide−Surface Interactions

With the increasing interest in protein adsorption in fields ranging from bionanotechnology to biomedical engineering, there is a growing need to understand protein-surface interactions at a fundamental level, such as the interaction between individual amino acid residues of a protein and functional groups presented by a surface. However, relatively little data are available that experimentally provide a quantitative, comparative measure of these types of interactions. To address this deficiency, the objective of this study was to generate a database of experimentally measured standard state adsorption free energy (DeltaGoads) values for a wide variety of amino acid residue-surface interactions using a host-guest peptide and alkanethiol self-assembled monolayers (SAMs) with polymer-like functionality as the model system. The host-guest amino acid sequence was synthesized in the form of TGTG-X-GTGT, where G and T are glycine and threonine amino acid residues and X represents a variable residue. In this paper, we report DeltaGoads values for the adsorption of 12 different types of the host-guest peptides on a set of nine different SAM surfaces, for a total of 108 peptide-surface systems. The DeltaGoads values for these 108 peptide-surface combinations show clear trends in adsorption behavior that are dependent on both peptide composition and surface chemistry. These data provide a benchmark experimental data set from which fundamental interactions that govern peptide and protein adsorption behavior can be better understood and compared.

Microfluidics of nano-drug delivery

After a brief review of microfluidics, a bio-MEMS application in terms of nanofluid flow in microchannels is presented. Specifically, the transient 3-D problem of controlled nano-drug delivery in a heated microchannel has been numerically solved to gain new physical insight and to determine suitable geometric and operational system parameters. Computer model accuracy was verified via numerical tests and comparisons with benchmark experimental data sets. The overall design goals of near-uniform nano-drug concentration at the microchannel exit plane and desired mixture fluid temperature were achieved with computer experiments considering different microchannel lengths, nanoparticle diameters, channel flow rates, wall heat flux areas, and nanofluid supply rates. Such micro-systems, featuring controlled transport processes for optimal nano-drug delivery, are important in laboratory-testing of predecessors of implantable smart devices as well as for analyzing pharmaceuticals and performing biomedical precision tasks.