Combination Of Ensemble Research Articles

Deep Learning-Based Prediction of Stress and Strain Maps in Arterial Walls for Improved Cardiovascular Risk Assessment

Conducting computational stress-strain analysis using finite element methods (FEM) is a common approach when dealing with the complex geometries of atherosclerosis, which is a leading cause of global mortality and complex cardiovascular disease. The considerable expense linked to FEM analysis encourages the substitution of FEM with a considerably faster data-driven machine learning (ML) approach. This study investigated the potential of end-to-end deep learning tools as a more effective substitute for FEM in predicting stress-strain fields within 2D cross sections of arterial walls. We first proposed a U-Net-based fully convolutional neural network (CNN) to predict the von Mises stress and strain distribution based on the spatial arrangement of calcification within arterial wall cross-sections. Further, we developed a conditional generative adversarial network (cGAN) to enhance, particularly from the perceptual perspective, the prediction accuracy of stress and strain field maps for arterial walls with various calcification quantities and spatial configurations. On top of U-Net and cGAN, we also proposed their ensemble approaches to improve the prediction accuracy of field maps further. Our dataset, consisting of input and output images, was generated by implementing boundary conditions and extracting stress-strain field maps. The trained U-Net models can accurately predict von Mises stress and strain fields, with structural similarity index scores (SSIM) of 0.854 and 0.830 and mean squared errors of 0.017 and 0.018 for stress and strain, respectively, on a reserved test set. Meanwhile, the cGAN models in a combination of ensemble and transfer learning techniques demonstrate high accuracy in predicting von Mises stress and strain fields, as evidenced by SSIM scores of 0.890 for stress and 0.803 for strain. Additionally, mean squared errors of 0.008 for stress and 0.017 for strain further support the model’s performance on a designated test set. Overall, this study developed a surrogate model for finite element analysis, which can accurately and efficiently predict stress-strain fields of arterial walls regardless of complex geometries and boundary conditions.

Two-Speed Deep-Learning Ensemble for Classification of Incremental Land-Cover Satellite Image Patches

High-velocity data streams present a challenge to deep learning-based computer vision models due to the resources needed to retrain for new incremental data. This study presents a novel staggered training approach using an ensemble model comprising the following: (i) a resource-intensive high-accuracy vision transformer; and (ii) a fast training, but less accurate, low parameter-count convolutional neural network. The vision transformer provides a scalable and accurate base model. A convolutional neural network (CNN) quickly incorporates new data into the ensemble model. Incremental data are simulated by dividing the very large So2Sat LCZ42 satellite image dataset into four intervals. The CNN is trained every interval and the vision transformer trained every half interval. We call this combination of a complementary ensemble with staggered training a “two-speed” network. The novelty of this approach is in the use of a staggered training schedule that allows the ensemble model to efficiently incorporate new data by retraining the high-speed CNN in advance of the resource-intensive vision transformer, thereby allowing for stable continuous improvement of the ensemble. Additionally, the ensemble models for each data increment out-perform each of the component models, with best accuracy of 65% against a holdout test partition of the RGB version of the So2Sat dataset.

Multiple Imputation Ensembles for Time Series (MIE-TS)

Time series classification has become an interesting field of research, thanks to the extensive studies conducted in the past two decades. Time series may have missing data, which may affect both the representation and also modeling of time series. Thus, recovering missing data using appropriate time series-based imputation methods is an essential step. Multiple imputation is a data recovery method where it produced multiple imputed data. The method proves its usefulness in terms of reflecting the uncertainty inherit in missing data; however, it is under-researched in time series problems. In this article, we propose two multiple imputation approaches for time series. The first is a multiple imputation method based on interpolation. The second is a multiple imputation and ensemble method. First, we simulate missing consecutive sub-sequences under a Missing Completely at Random mechanism; then, we use single/multiple imputation methods. The imputed data are used to build bagging and stacking ensembles. We build ensembles using standard classification algorithms as well as time series classifiers. The standard classifiers involve Random Forest, Support Vector Machines, K-Nearest Neighbour, C4.5, and PART while TSCHIEF, Proximity Forest, Time Series Forest, RISE, and BOSS are chosen as time series classifiers. Our findings show that the combination of multiple imputation and ensemble improves the performance of the majority of classifiers tested in this study, often above the performance obtained from the complete data, even under increasing missing data scenarios. This may be because the diversity injected by multiple imputation has a very favourable and stabilising effect on the classifier performance, which is a very important finding.

Conjunction of cluster ensemble-model ensemble techniques for spatiotemporal assessment of groundwater depletion in semi-arid plains

• The cluster ensemble technique improved the final cluster structures. • The spatiotemporal ensemble clustering revealed four different patterns of GWL. • LSTM was not appropriate for monthly GWL forecasting compared to ensemble methods. • The neural ensemble has the lowest disorder and a large reduction in uncertainty. • The simultaneous clustering-forecasting ensemble techniques enhanced the outcomes. In this study, first to identify the patterns of groundwater level (GWL) over the Ghorveh-Dehgolan plain (GDP) located in western Iran, as a data pre-processing scheme, three different types of clustering algorithms were applied to monthly GWL data sets of the piezometers. Then, the best structures of all clustering methods were integrated by Combining Multiple Clusterings via Similarity Graph (COMUSA) method to obtain the most homogenous patterns of GWL. The final results of the clustering step indicated that applying COMUSA could enhance the homogeneity of the clusters up to 25%. After dividing the GWL of GDP into four patterns, three single artificial intelligence (AI)-based models were applied to forecast multi-step-ahead GWL of centroid piezometer of each cluster. To benefit from the advantages of the single models, the outcomes were then combined with a neural averaging ensemble (NAE) technique as a post-processing step. Additionally, the assessment of the deep learning (DL) -based long-short-term memory (LSTM) application in multi-step ahead forecasting of GWL showed this method is not an appropriate choice compared to ensemble techniques for modeling the process with limited observed data. The comparison of the proposed GWL forecasting models of this study revealed the superiority of the NAE technique that enhanced the accuracy of the single models up to 23% in the testing phase. It could be concluded that the combination of cluster ensemble and model ensemble techniques could improve the performance of the individual method in reliable forecasting of the future GWL condition and the methodology of this study can be applied to the GWL of other plains.

A new CAM6 + DART reanalysis with surface forcing from CAM6 to other CESM models

An ensemble Kalman filter reanalysis has been archived in the Research Data Archive at the National Center for Atmospheric Research. It used a CAM6 configuration of the Community Earth System Model (CESM), several million observations per day, and the Data Assimilation Research Testbed (DART). The data saved from this global, sim 1^circ resolution, 80 member ensemble span 2011–2019. They include ensembles of: sub-daily, real world, atmospheric forcing for use by all of the nonatmospheric models of CESM; weekly, CAM6, restart file sets; 6 hourly, prior hindcast estimates of the assimilated observations; 6 hourly, land model, plant growth variables, and 6 hourly, ensemble mean, gridded, atmospheric analyses. This data can be used for hindcast studies and data assimilation using component models of CESM; CAM6, CLM5, CICE5, POP2. MOM6, MOSART, and CISM; and non-CESM Earth system models. This large dataset (~ 120 Tb) has a unique combination of a large ensemble, high frequency, and multiyear time span, which provides opportunities for robust statistical analysis and use as a machine learning training dataset.

Carillon: History and Performing Capabilities of the Musical Instrument, its “Philosophy” and Prospects of Development in Ukrainian Culture

The article focuses on the history of origin, the specifics of sound production, the philosophy of performance of the carillon – a European musical instrument, which in recent decades has become active in Ukraine. The purpose of the research is to analyze the history of the carillon origin, its design, technical and expressive means of sound extraction and musical representation of composition and performance. Research methodology. The article uses historical, axiological, musicological and culturological approaches and corresponding methods. The historical-chronological method is used to consider the history of origin, stages of carillon distribution in Europe and the world, axiological – to determine the artistic and psychotherapeutic value of bell (carillon) sound, musicological and culturological methods contribute to various analysis’ aspects of the place and meaning of Ukrainian carillon. Scientific novelty of research. A historical retrospective of the carillon’s spread has been carried out; generalizations regarding its constructive, melodic-intonation, technical, figurative-artistic properties, in particular, in the Ukrainian sound space, have been introduced into the scientific circulation of Ukrainian musicology. Conclusions. The history of the creation of stationary and mobile carillons in Ukraine, stages of entering the specified musical instrument into the Ukrainian sound space, the character of the repertoire performed on it, pedagogical and performing aspects of its popularization have been analyzed. The dynamics of the International Festival of Bell and Carillon Art ‘Bells of Yasna Hora Unite Everyone’ is noted. The performance possibilities of the timbre combination of the carillon with other instruments and singing voices are noted. The technical and expressive significance of synthesizing carillon and the ensemble of bandurists has been highlighted and analyzed. Textural variation of ensemble combination of carillon and ensemble of bandurists and possibilities of thematic and genre extension of repertoire has been proposed.

Covid19 Identification from Chest X-ray Images using Machine Learning Classifiers with GLCM Features

From staying quarantined at home, practicing work from home to moving outside wearing masks and carrying sanitizers, every individual has now become so adaptive to so called ‘New Normal’ post series of lockdowns across the countries. The situation triggered by novel Coronavirus has changed the behaviour of every individual towards every other living as well as non-living entity. In the Wuhan city of China, multiple cases were reported of pneumonia caused due to unknown reasons. The concerned medical authorities confirmed the cause to be Coronavirus. The symptoms seen in these cases were not much different than those seen in case of pneumonia. Earlier the research has been carried out in the field of pneumonia identification and classification through X-ray images of chest. The difficulty in identifying Covid19 infection at initial stage is due to high resemblance of its symptoms with the infection caused due to pneumonia. Hence it is trivial to well distinguish cases of coronavirus from pneumonia that may help in saving life of patients. The paper uses chest X-ray images to identify Covid19 infection in lungs using machine learning classifiers and ensembles with Gray-Level Cooccurrence Matrix (GLCM) features. The advocated methodology extracts statistical texture features from X-ray images by computing a GLCM for each image. The matrix is computed by considering various stride combinations. These GLCM features are used to train the machine learning classifiers and ensembles. The paper explores both the multiclass classification (X-ray images are classified into one of the three classes namely Covid19 affected, Pneumonia affected and normal lungs) and binary classification (Covid19 affected and other). The dataset used for evaluating performance of the method is open sourced and can be accessed easily. Proposed method being simple and computationally effective achieves noteworthy performance in terms of Accuracy, F-Measure, MCC, PPV and Sensitivity. In sum, the best stride combination of GLCM and ensemble of machine learning classifiers is suggested as vital outcome of the proposed method for effective Covid19 identification from chest X-ray images

Viability of temperate fruit tree varieties in Spain under climate change according to chilling accumulation

Fruit trees stop their growth over the coldest period of the year to avoid damage. To resume their growth and for successful fruit production, they need to accumulate winter chill. It is expected that global warming will diminish winter chill availability with potentially negative impacts on the viability and yield of these crops. The objective of this study was to assess the viability of seven tree crops among the most relevant in peninsular Spain and the Balearic Islands. For this purpose, chilling requirements were gathered from the literature to define a requirement range for each tree crop encompassing most of the varieties used in Spain. Then the bias-adjusted outputs of an ensemble of 10 regional climate models under the two representative concentration pathways (RCPs), 4.5 and 8.5, were used to feed a chilling model for calculating chill accumulation. The ensemble's outcome agreement index was applied to each combination of chill requirement and climate ensemble to assess crop viability. This was done by testing the hypothesis that the winter chill accumulation will be greater than the safe winter chill for the 2021–2050 (near future) and 2071–2100 (far future) periods and for both RCPs. A future reduction in the safe winter chill areas is projected with high agreement in climate projections across peninsular Spain and the Balearic Islands independently of the RCP or future period. The crops studied would be viable in the near future period as long as varieties with low chilling requirements are used. These varieties, however, would not be adequate in the far future in some currently highly productive regions, where the situation would become more severe, especially under the RCP8.5 scenario. In these cases, adaptation would be possible by shifting the crop to adjacent areas together with careful variety selection in terms of chilling requirements. The results indicate that the RCP8.5 scenario in the far future has an especially negative impact on the crops analysed, calling for resolute mitigation measures to guarantee tree crop production and food security. Recommendations for adaptation, with low uncertainty regarding climate projections, were included here, using actual tree varieties, thereby facilitating interpretation and on-field application for farmers and agricultural technicians.

Unsupervised local cluster-weighted bootstrap aggregating the output from multiple stochastic simulators

In the present work, we consider the problem of combining the output from multiple stochastic computer simulators to make inference on a quantity of interest, as a means of reducing the inherent model-form uncertainty in the absence of any measurements. In most real-world situations, judging an individual stochastic simulator to be the “best” for any given point in the input space is highly doubtful. Thus, making inference by relying on the so-deemed best simulator may not be adequate, especially when the sampled data is limited. To this end, we propose an ensemble learning method based on local Clustering and bootstrap aggregation (Bagging), which rather than treating the stochastic predictions of the simulators as competing individual information sources, treats those as part of an ensemble, thus diversifying the hypothesis space. We call the proposed method: unsupervised local cluster-weighted bootstrap aggregation. Variational Bayesian Gaussian mixture clustering is the first step in this ensemble learning approach for discriminating the outputs, and deriving the probability map (weights) of the clustered simulators output. Clustering is performed on the stochastic output corresponding to the binned input space. Performing the clustering independently and deriving the probability map for each local region of the binned input space is a novelty that guarantees an adaptive solution, whereby certain simulators are potentially more fitting than others in corresponding regions of the input space. The second step consists in a local cluster-weighted Bootstrap Aggregation, which serves the purpose of weighted combination of the clustered ensemble of outputs from the individual simulators. Based on simulations, we demonstrate how the input bin size, sample size, output dispersion and level of agreement amongst the simulators affect the performance of the proposed method. We compare the unsupervised local cluster-weighted bootstrap aggregation method to classical Bagging, Bayesian Model Averaging and Stacking of predictive distributions. Finally, we demonstrate the method by evaluating the fatigue damage equivalent load on a wind turbine blade, using 10 finite element based simulators. The results point to the need for practitioners to consider this as a useful method, when model-form uncertainty is of concern and when output from multiple stochastic simulators are available.

Ultralow Pt loading nanoporous Au-Cu-Pt thin film as highly active and durable catalyst for formic acid oxidation

Ultralow Pt loading tri-metallic nanoporous (np) Au-Cu-Pt catalysts for formic acid oxidation (FAO) have been synthesized by electrodeposition of Au0.27Cu0.70Pt0.03 alloy thin film, followed by dealloying for oxidative Cu removal. Underpotential deposition and CO stripping are used for measuring the electrochemical active surface (overall and Pt fraction) of synthesized catalysts and SEM is performed for characterization of their morphology. Also, EDS and inductive coupled plasma are used for compositional analysis. Potential cycling FAO tests show excellent performance of these catalysts manifested by up to 250(mgPt) and 17(mgPt+Au) times higher mass activity (at 0.0V/MSE) than commercial Pt/C(40%) nanoparticulate counterparts. This is accompanied by remarkable durability in both cycling and constant potential tests. A no-hysteresis voltammetry for 55 cycles of synthesized np Au-Cu-Pt catalysts demonstrates unseen to date resistance to anodic passivation along with outstanding CO-poisoning tolerance. A combination of electronic, ensemble, and synergistic effects is believed to enable catalysts’ outstanding performance.

Spatial interpolation of two‐metre temperature over Norway based on the combination of numerical weather prediction ensembles and in situ observations

AbstractAccurate hourly two‐metre temperature gridded fields available in near real‐time are valuable products for numerous applications, such as civil protection and energy production planning. An analysis ensemble of temperature is obtained from the combination of a numerical weather prediction ensemble (background) and in situ observations. At the core of the flow‐dependent spatial interpolation method lies the analysis step of the local ensemble transform Kalman filter (LETKF). A scaling factor and a localization procedure have been added to correct for deficiencies of the background. Each observation is characterized by its own representativeness, which is allowed to vary in time. We call the method described here an Ensemble‐based Statistical Interpolation (EnSI) scheme for spatial analysis and it has been integrated into the operational post‐processing systems in use at the Norwegian Meteorological Institute (MET Norway). The benefits of the analysis are assessed over a 1‐year time period (July 2017–July 2018) and a case‐study is presented for a challenging situation over complex terrain. EnSI gives more accurate results than an interpolation method based exclusively on observations. The analysis ensemble provides a more informative representation of the uncertainty than a spatial analysis based on a single‐field background. EnSI reduces the number of large prediction errors in the analysis compared to the background by almost 50%, reduces the ensemble spread and increases its accuracy.

Ensemble-based uncertainty estimation in Full Waveform Inversion

SUMMARY Uncertainty estimation and quality control are critically missing in most geophysical tomographic applications. The few solutions to cope with that issue are often left out in practical applications when these ones grow in scale and involve complex modeling. We present a joint full waveform inversion and ensemble data assimilation scheme, allowing local Bayesian estimation of the solution that brings uncertainty estimation to the tomographic problem. This original methodology relies on a deterministic square root ensemble Kalman filter commonly used in the data assimilation community: the ensemble transform Kalman filter. Combined with a 2D visco-acoustic frequency domain full waveform inversion scheme, the resulting method allows to access a low-rank approximation of the posterior covariance matrix of the solution. It yields uncertainty information through an ensemble-representation, that can conveniently be mapped across the physical domain for visualization and interpretation. The combination of ensemble transform Kalman filter with full waveform inversion is discussed along with the scheme design and algorithmic details that lead to our mixed application. Both synthetic and field-data results are presented, along with the biases that are associated with the limited rank ensemble representation. Finally, we review the open questions and developments perspectives linked with data assimilation applications to the tomographic problem.

“Case Carmen” by Maxim Shalygin: heroine’s characteristic Features in a retrospective ballet composition

Topicality of research. In contemporary culture artists take a remarkable interest in remakes due to the desire to actualize the classical subjects and show new semantic layers familiar to the person of the 21st century. Dance performance “Case Carmen” (2017, The Hague) by the Dutch choreographer Lonneke van Leth opens up a new modern look at one of the brightest heroines in the history of world literature and art.Scientific novelty. The musical series of ballet, written by the Ukrainian-Dutch composer of the new generation Maxim Shalygin, for the first time serve as an object of scientific consideration which is a new step in the process of studying the composer’s multifaceted creativity.The main objective of the study is to consider the features of Carmen’s figurative characteristic in retrospective composition of “Case Carmen” dance performance.Methodology. The article uses methods of historical and theoretical musicology. Ballet electro acoustic score has determined the peculiarity of the analytical method which combines genre-intonation and auditory descriptive analysis. A comparative method is used to reveal similar and distinctive features of the modern ballet version and primary sources (P. Merime’s novels, J. Bizet’s opera); 2) in Carmen’s image interpretation in the ballets of the 20–21 centuries. An appeal to the retrospective category led to the use of a linguistic approach.Relevance of the study. “Case Carmen” (Dance Performance About the Murder Case) is a «dance thriller» whose characters (Carmen, Jose, Mikaela) are borrowed from J. Bizet’s opera but are inscribed in the picture of modern life with its oppressive routine and routine relationships. M. Shalygin’s music as well as L. van Leet’s choreography emphasizes the acute-modern orientation of the play (this is evidenced by the performer composition itself – the combination of instrumental ensemble, female choir and electronics). There were found connections of the ballet with P. Merimee’s novella, which are traced at the level of constructing the whole (retrospective composition stemming from the specific organization of artistic time) as well as the level of Carmen’s characteristics (formation of the heroine’s image from individual memories and representations about her). The value of the Detective’s image (dramatic role) is revealed, it is a new character that lives through Carmen’s case in retrospective reflection. Being obsessed with the fatal beauty, he reconstructs the history of her relationship with Jose, speaking as a guide between past and present. The importance of a wide range of noise in the musical fabric of the ballet was emphasized, which “cemented” the whole structure, while enhancing the feeling of nonlinear time, the habitat of events in partially erased, blurry or distorted by the consciousness memories. On the basis of the dance performance musical series analysis the complexity and ambiguity of the image of Carmen with the inherent unpredictability, mystery, demonism, aspiration of mortal danger, a combination of incompatible features such as purity and promiscuity are described. Accordingly, the musical characterization of the heroine is colorful and contradictory, based on a variety of genre-tone sources (elevated chorale and club dubstep, inspired flexible cantilena and soulless mechanistic motility).Results and conclusions. The dance performance “Case Carmen” is marked by the discontinuity, multidimensionality and artistic time reversal. In this complex fluctuation of the present and past images of the main characters appear among which the most complex and ambiguous is Carmen’s characteristic. An exceptionally original and multi-valued interpretation of the eternal story in the modern format of the “dance thriller” presented by the choreographer L. van Leet and composer M. Shalygin contains numerous directions for new research. First of all, this relates to the ballet musical number which is one of the most important components of the performance that gives the key to solving the mystery of the fatal beauty.

Telomere DNA G-quadruplex folding within actively extending human telomerase

Telomerase reverse transcribes short guanine (G)-rich DNA repeat sequences from its internal RNA template to maintain telomere length. G-rich telomere DNA repeats readily fold into G-quadruplex (GQ) structures in vitro, and the presence of GQ-prone sequences throughout the genome introduces challenges to replication in vivo. Using a combination of ensemble and single-molecule telomerase assays, we discovered that GQ folding of the nascent DNA product during processive addition of multiple telomere repeats modulates the kinetics of telomerase catalysis and dissociation. Telomerase reactions performed with telomere DNA primers of varying sequence or using GQ-stabilizing K+ versus GQ-destabilizing Li+ salts yielded changes in DNA product profiles consistent with formation of GQ structures within the telomerase-DNA complex. Addition of the telomerase processivity factor POT1-TPP1 altered the DNA product profile, but was not sufficient to recover full activity in the presence of Li+ cations. This result suggests GQ folding synergizes with POT1-TPP1 to support telomerase function. Single-molecule Förster resonance energy transfer experiments reveal complex DNA structural dynamics during real-time catalysis in the presence of K+ but not Li+, supporting the notion of nascent product folding within the active telomerase complex. To explain the observed distributions of telomere products, we globally fit telomerase time-series data to a kinetic model that converges to a set of rate constants describing each successive telomere repeat addition cycle. Our results highlight the potential influence of the intrinsic folding properties of telomere DNA during telomerase catalysis, and provide a detailed characterization of GQ modulation of polymerase function.

CENDroid—A cluster-ensemble classifier for detecting malicious Android applications

Given the use of mobile phones in our day-to-day activities—from basic applications (such as alarm clocks) to sensitive applications (such as banking)—these devices perform a vital function in today's world. Because of the sensitive and critical information they contain, these devices are among the prime targets of hackers. The phone market is dominated by Android-based phones. The open source nature of Android has also spawned various security concerns because of the extensive spread of malware. Accordingly, various classification algorithms (individual as well as ensemble) have been employed for Android malware detection. In this paper, we propose a novel Android malware detection system—CENDroid, which uses static features (API tags and permissions) along with a combination of clustering and ensemble of classifiers for classifying Android apps as benign or malicious. With the proposed cluster-ensemble method, a comparative assessment is implemented on the performance of popular individual classifiers and their ensembles; experiments on three datasets of malicious and benign Android applications are conducted as well. Relevant statistical tests validate that our proposed system outperforms individual classifiers as well as their ensemble in delivering high threshold and rank metrics for malware detection.

Free and Chaperone Bound Unfolded States of Outer Membrane Proteins

Structural and dynamic investigations of unfolded proteins are important for understanding protein-folding mechanisms as well as the interactions of unfolded polypeptide chains with chaperones. In the case of bacterial outer-membrane proteins (OMPs), unfolded-state properties are of particular physiological relevance, because they remain unfolded for extended periods of time during their biogenesis and rely on interactions with periplasmic chaperones to prevent aggregation and support correct folding. Here, we study the unfolded-state properties of outer-membrane proteins. Using a combination of ensemble and single-molecule spectroscopy techniques including single-molecule FRET, we find that under strongly denaturing conditions and in the absence of chaperones, OMPs adopt conformationally heterogeneous unfolded states that lack the fast chain reconfiguration motions expected for an unstructured, fully unfolded chain. Interestingly, when complexed with periplasmic chaperones seventeen kilodalton protein (Skp) and survival factor A (SurA), we observe for the protein OmpLA adopts an expanded, yet squished conformational ensemble, in which the overall broad distribution of unfolded states is preserved. Differently sized OMPs like OmpX (8 β-strands) and OmpF (16 β-strands) allow a broader perspective on chaperone enabled conformations of OMPs. These findings indicate that periplasmic chaperones Skp and SurA prestructure OMPs for their next binding partner (e.g., the β-barrel assembly machinery (BAM)) to facilitate insertion and proper folding into the outer membrane.

A deep Boltzmann machine and multi-grained scanning forest ensemble collaborative method and its application to industrial fault diagnosis

The essence of big data-based intelligent industrial fault diagnosis lies in the process of machine learning and feature engineering. Deep learning methods can discover the complex relationship between data and potential faults, and outperform the traditional machine learning methods. The gcForest is able to generate a deep forest ensemble, which allows gcForest to do representation learning and fault classification. At each node of the random tree, gcForest selects the one with the best gini value from candidates for splitting. However, most of the data acquired from industrial scene is with continuous and unstructured attributes, accordingly the node-splitting procedure will be generally intractable. We present a novel approach with the combination of deep Boltzmann machine and multi-grained scanning forest ensemble, to effectively deal with industrial fault diagnosis based on big data. At first, we use deep Boltzmann machine to turn all features of data to be processed by forests into binary, and then utilize multi-grained scanning forest ensemble to process them in every layer of deep Boltzmann machine. By means of the collaborative method, we can address the aforementioned issues. The experimental results and analysis on industrial fault diagnosis under different experimental conditions, show that the fault classification accuracy of the proposed approach is competitive to other popular deep learning algorithms, but also takes much less time than gcForest.

Slow Interconversion in a Conformationally Heterogeneous Unfolded-State Ensemble of Outer Membrane Phospholipase A

Structural and dynamic investigations of unfolded proteins are important for understanding protein-folding mechanisms as well as the interactions of unfolded polypeptide chains with other cell components. In the case of outer membrane proteins (OMPs), unfolded-state properties are of particular physiological relevance because these proteins remain unfolded for extended periods of time during their biogenesis and rely on interactions with binding partners to support proper folding. Using a combination of ensemble and single-molecule spectroscopy, we have scrutinized the unfolded state of Outer Membrane Phospholipase A (OmpLA) to provide a detailed view of its structural dynamics from nanoseconds to milliseconds. We find that, even under strongly denaturing conditions and in the absence of residual secondary structure, OmpLA populates an ensemble of slowly (>100 ms) interconverting and conformationally heterogeneous unfolded states that lack the fast chain reconfiguration motions expected for an unstructured, fully unfolded chain. The drastically slowed sampling of potentially folding-competent states as compared with a random-coil polypeptide may contribute to the slow in vitro folding kinetics observed for many OMPs. In vivo, however, slow intramolecular long-range dynamics might be advantageous for entropically favored binding of unfolded OMPs to chaperones and, by facilitating conformational selection after release from chaperones, for preserving binding-competent conformations before insertion into the outer membrane.

A simple plug-in bagging ensemble based on threshold-moving for classifying binary and multiclass imbalanced data.

Class imbalance presents a major hurdle in the application of classification methods. A commonly taken approach is to learn ensembles of classifiers using rebalanced data. Examples include bootstrap averaging (bagging) combined with either undersampling or oversampling of the minority class examples. However, rebalancing methods entail asymmetric changes to the examples of different classes, which in turn can introduce their own biases. Furthermore, these methods often require specifying the performance measure of interest a priori, i.e., before learning. An alternative is to employ the threshold moving technique, which applies a threshold to the continuous output of a model, offering the possibility to adapt to a performance measure a posteriori, i.e., a plug-in method. Surprisingly, little attention has been paid to this combination of a bagging ensemble and threshold-moving. In this paper, we study this combination and demonstrate its competitiveness. Contrary to the other resampling methods, we preserve the natural class distribution of the data resulting in well-calibrated posterior probabilities. Additionally, we extend the proposed method to handle multiclass data. We validated our method on binary and multiclass benchmark data sets by using both, decision trees and neural networks as base classifiers. We perform analyses that provide insights into the proposed method.

Ensemble and Single Particle Fluorescence Characterization of Dye-Labeled Cellulose Nanocrystals.

Cellulose nanocrystals (CNCs) have been covalently labeled with both fluorescein and rhodamine and studied by a combination of UV-vis absorption spectroscopy and ensemble and single molecule fluorescence spectroscopy. For all samples, the fluorescence anisotropy and lifetimes were consistent with effects expected for covalently bound dye molecules. Low dye loading levels (∼0.1 dye/particle) were estimated for the fluorescein-labeled CNC which coupled with the strong pH dependence make this a less suitable fluorophore for most applications. Rhodamine-labeled CNCs were prepared from both sulfated and carboxylated CNCs and had loading levels that varied from 0.25 to ∼15 dye molecules/CNC. For the sulfated samples, the absorption due to (nonfluorescent) dimeric dye increased with dye loading; in contrast, the carboxylated sample, which had the highest rhodamine content, had a low dimer yield. Single particle fluorescence studies for two of the rhodamine-labeled CNCs demonstrated that individual particles are readily detected by their stepwise blinking/bleaching behavior and by polarization effects. Overall, the results indicate the importance of understanding the effects of loading on dye photophysics to select an optimal dye concentration to maximize sensitivity while minimizing the effect of the dye on the CNC behavior. The results also demonstrate that CNCs with relatively low dye loadings (e.g., ∼1 dye/particle) are readily detectable by fluorescence and should be adequate for use in fluorescence-based biological assays or to probe the distribution of CNCs in composite materials.