Kernel-based Machine Learning Research Articles

Discrimination of Explosive Residues by Standoff Sensing Using Anodic Aluminum Oxide Microcantilever Laser Absorption Spectroscopy with Kernel-Based Machine Learning.

Standoff laser absorption spectroscopy (LAS) has attracted considerable interest across many applications for environmental safety. Herein, we propose an anodic aluminum oxide (AAO) microcantilever LAS combined with machine learning (ML) for sensitive and selective standoff discrimination of explosive residues. A nanoporous AAO microcantilever with a thickness of <1 μm was fabricated using a micromachining process; its spring constant (18.95 mN/m) was approximately one-third of that of a typical Si microcantilever (53.41 mN/m) with the same dimensions. The standoff infrared (IR) spectra of pentaerythritol tetranitrate, cyclotrimethylene trinitramine, and trinitrotoluene were measured using our AAO microcantilever LAS over a wide range of wavelengths, and they closely matched the spectra obtained using standard Fourier transform infrared spectroscopy. The standoff IR spectra were fed into ML models, such as kernel extreme learning machines (KELMs), support vector machines (SVMs), random forest (RF), and backpropagation neural networks (BPNNs). Among these four ML models, the kernel-based ML models (KELM and SVM) were found to be efficient learning models able to satisfy both a high prediction accuracy (KELM: 94.4%, SVM: 95.8%) and short hyperparameter optimization time (KELM: 5.9 s, SVM: 7.6 s). Thus, the AAO microcantilever LAS with kernel-based learners could emerge as an efficient sensing method for safety monitoring.

A hyperparameter study for quantum kernel methods

Quantum kernel methods are a promising method in quantum machine learning thanks to the guarantees connected to them. Their accessibility for analytic considerations also opens up the possibility of prescreening datasets based on their potential for a quantum advantage. To do so, earlier works developed the geometric difference, which can be understood as a closeness measure between two kernel-based machine learning approaches, most importantly between a quantum kernel and a classical kernel. This metric links the quantum and classical model complexities, and it was developed to bound generalization error. Therefore, it raises the question of how this metric behaves in an empirical setting. In this work, we investigate the effects of hyperparameter choice on the model performance and the generalization gap between classical and quantum kernels. The importance of hyperparameters is well known also for classical machine learning. Of special interest are hyperparameters associated with the quantum Hamiltonian evolution feature map, as well as the number of qubits to trace out before computing a projected quantum kernel. We conduct a thorough investigation of the hyperparameters across 11 datasets, and we identify certain aspects that can be exploited. Analyzing the effects of certain hyperparameter settings on the empirical performance, as measured by cross validation accuracy, and generalization ability, as measured by geometric difference described above, brings us one step closer to understanding the potential of quantum kernel methods on classical datasets.

Life-cycle production optimization with nonlinear constraints using a least-squares support-vector regression proxy

When nonlinear constraints such as field liquid or water production rate, injection pressures, etc., as functions of time need to be honored in addition to linear ones, the life-cycle production optimization problem, a component of a closed loop reservoir management, becomes challenging and computationally expensive to perform using a high-fidelity reservoir simulator with the existing gradient-based methods using the adjoint or stochastic approximate gradient methods. Therefore, the objective of this study is to present computationally efficient methods for deterministic production optimization under nonlinear constraints using a kernel-based machine learning method where the cost function is the net present value (NPV). We use the least-squares support-vector regression (LSSVR) to approximate the NPV function. To achieve computational efficiency, we generate a set of output values of the NPV and nonlinear constraint functions, which are field liquid production rate (FLPR) and water production rate (FWPR) in this study, by running the high-fidelity simulator for a broad set of input design variables (well controls) and then using the collection of input/output data to train LSSVR proxy models to replace the high-fidelity simulator to compute NPV and nonlinear state constraint functions during iterations of sequential quadratic programming (SQP). To obtain improved (higher) estimated optimal NPV values, we use the existing so-called iterative sampling refinement (ISR) method to update the LSSVR proxy so that the updated proxy remains predictive toward promising regions of search space during the optimization. Direct and indirect ways of constructing LSSVR-based NPVs as well as different combinations of input data, including nonlinear state constraints and/or the bottomhole pressures (BHPs) and water injection rates, are tested as alternative feature spaces. The results obtained from our proposed LSSVR-based optimization methods are compared with those obtained from our in-house stochastic simplex approximate gradient (StoSAG)-based line-search SQP programming (LS-SQP-StoSAG) algorithm that uses directly a high-fidelity simulator to compute the gradients of the objective function and the nonlinear state functions with StoSAG for the Brugge reservoir model. The results show that nonlinear constrained optimization with the LSSVR ISR with SQP is computationally 3.25 fold more efficient than LS-SQP-StoSAG. In addition, the results show that constructing NPV indirectly using the field liquid and water rates for a waterflooding problem where inputs come from LSSVR proxies of the nonlinear state constraints requires significantly fewer training samples than the method constructing NPV directly from the NPVs computed from a high-fidelity simulator. LSSVR has advantages for its computational efficiency which is the main goal in our research, and robustness against overfitting especially for the cases where we have limited data. However, DNNs and random forest require large training size and require much more computational resources and longer training times. Also, Random Forests and Gradient Boosting Machines can be prone to overfitting and become computationally intensive.

Do energy, inflation, and financial development stimulate economic welfare in India? Empirical insights from novel dynamic ARDL and KRLS simulations

The present study focuses on investigating the effect of renewable and non-renewable energy consumption (REC and NREC), financial development (FD), and inflation (INF) on economic welfare (EW) in India during 1990–2019. In this research, we employ a novel dynamic autoregressive distributed lag model and a kernel-based machine learning algorithm and regularise least squares to detect causal linkages among the variables. To measure the economic welfare index, this research uses the PCA analysis with GDP, human capital, environmental damage, number of employed persons, life expectancy at birth, and real consumption in households and by the government. According to the findings, the NREC and FD have been favourably contributing to EW. Conversely, the EW is adversely affected by REC. However, INF also has an adverse impact on EW, although it may be insignificant in the long run and significant in the short run.

Predicting the Occurrence of Substituted and Unsubstituted, Polycyclic Aromatic Compounds in Coking Wastewater Treatment Plant Effluent using Machine Learning Regression

Organic contaminants such as polycyclic aromatic compounds (PACs) occurring in industrial effluents can not only persist in wastewater but transform into more toxic and mobile, substituted heterocyclic products during treatment. Thus, predicting the occurrence of PACs and their heterocyclic derivatives (HPACs) in coking wastewater is of utmost importance to reduce the environmental risks in water bodies that receive industrial effluents. Although HPACs can be monitored through sampling and analysis, the characterisation techniques used in their analyses are costly and time-consuming. In this study, we propose 3 distinct kernel-based machine learning (ML) models for predicting PACs including substituted HPACs and alkylated PACs occurring in coking wastewater. By using routinely measured wastewater quality data as input for our models, we predicted the occurrence of 14 HPACs in the final effluent of a coking wastewater treatment plant. Support Vector Machine based regression model (SVR) used for HPAC prediction showed the highest R2 of 0.83. Performance assessment of SVR model showed a mean absolute logarithmic error (MALE) of 0.46 and root mean square error (RMSE) of 0.073 ng/L. Comparatively, K-Nearest Neighbor and Random Forest models showed lower R2 of 0.75 and 0.76 respectively for HPAC prediction. Feature analysis attributed the superior predictability of SVR model likely to its higher weightage (81%) towards dissolved organic carbon and total ammonia as input variables. Both these variables could capture the underlying secondary PAC transformations likely occurring in the treatment plant. Partial dependence plots predicted that ammonia levels higher than 120 mg/L and DOC levels of 50–60 mg/L were likely linked to higher HPACs occurring in the final effluent. This work highlights the capability of kernel-based ML models in capturing nonlinear wastewater chemistry and offers a tool for monitoring trace organic contaminants released in coking effluents.

Fast, accurate, and interpretable decoding of electrocorticographic signals using dynamic mode decomposition

Dynamic mode (DM) decomposition decomposes spatiotemporal signals into basic oscillatory components (DMs). DMs can improve the accuracy of neural decoding when used with the nonlinear Grassmann kernel, compared to conventional power features. However, such kernel-based machine learning algorithms have three limitations: large computational time preventing real-time application, incompatibility with non-kernel algorithms, and low interpretability. Here, we propose a mapping function corresponding to the Grassmann kernel that explicitly transforms DMs into spatial DM (sDM) features, which can be used in any machine learning algorithm. Using electrocorticographic signals recorded during various movement and visual perception tasks, the sDM features were shown to improve the decoding accuracy and computational time compared to conventional methods. Furthermore, the components of the sDM features informative for decoding showed similar characteristics to the high-γ power of the signals, but with higher trial-to-trial reproducibility. The proposed sDM features enable fast, accurate, and interpretable neural decoding.

A kernel-based machine learning potential and quantum vibrational state analysis of the cationic Ar hydride (Ar2H+).

One of the most fascinating discoveries in recent years, in the cold and low pressure regions of the universe, was the detection of ArH+ and HeH+ species. The identification of such noble gas-containing molecules in space is the key to understanding noble gas chemistry. In the present work, we discuss the possibility of [Ar2H]+ existence as a potentially detectable molecule in the interstellar medium, providing new data on possible astronomical pathways and energetics of this compound. As a first step, a data-driven approach is proposed to construct a full 3D machine-learning potential energy surface (ML-PES) via the reproducing kernel Hilbert space (RKHS) method. The training and testing data sets are generated from CCSD(T)/CBS[56] computations, while a validation protocol is introduced to ensure the quality of the potential. In turn, the resulting ML-PES is employed to compute vibrational levels and molecular spectroscopic constants for the cation. In this way, the most common isotopologue in ISM, [36Ar2H]+, was characterized for the first time, while simultaneously, comparisons with previously reported values available for [40Ar2H]+ are discussed. Our present data could serve as a benchmark for future studies on this system, as well as on higher-order cationic Ar-hydrides of astrophysical interest.

Identification of reservoir types in deep carbonates based on mixed-kernel machine learning using geophysical logging data

Identification of reservoir types in deep carbonates has always been a great challenge due to complex logging responses caused by the heterogeneous scale and distribution of storage spaces. Traditional cross-plot analysis and empirical formula methods for identifying reservoir types using geophysical logging data have high uncertainty and low efficiency, which cannot accurately reflect the nonlinear relationship between reservoir types and logging data. Recently, the kernel Fisher discriminant analysis (KFD), a kernel-based machine learning technique, attracts attention in many fields because of its strong nonlinear processing ability. However, the overall performance of KFD model may be limited as a single kernel function cannot simultaneously extrapolate and interpolate well, especially for highly complex data cases. To address this issue, in this study, a mixed kernel Fisher discriminant analysis (MKFD) model was established and applied to identify reservoir types of the deep Sinian carbonates in central Sichuan Basin, China. The MKFD model was trained and tested with 453 datasets from 7 coring wells, utilizing GR, CAL, DEN, AC, CNL and RT logs as input variables. The particle swarm optimization (PSO) was adopted for hyper-parameter optimization of MKFD model. To evaluate the model performance, prediction results of MKFD were compared with those of basic-kernel based KFD, RF and SVM models. Subsequently, the built MKFD model was applied in a blind well test, and a variable importance analysis was conducted. The comparison and blind test results demonstrated that MKFD outperformed traditional KFD, RF and SVM in the identification of reservoir types, which provided higher accuracy and stronger generalization. The MKFD can therefore be a reliable method for identifying reservoir types of deep carbonates.

Biophysical Variable Retrieval of Silage Maize with Gaussian Process Regression and Hyperparameter Optimization Algorithms

Quantification of vegetation biophysical variables such as leaf area index (LAI), fractional vegetation cover (fCover), and biomass are among the key factors across hydrological, agricultural, and irrigation management studies. The present study proposes a kernel-based machine learning algorithm capable of performing adaptive and nonlinear data fitting so as to generate a suitable, accurate, and robust algorithm for spatio-temporal estimation of the three mentioned variables using Sentinel-2 images. To this aim, Gaussian process regression (GPR)–particle swarm optimization (PSO), GPR–genetic algorithm (GA), GPR–tabu search (TS), and GPR–simulated annealing (SA) hyperparameter-optimized algorithms were developed and compared against kernel-based machine learning regression algorithms and artificial neural network (ANN) and random forest (RF) algorithms. The accuracy of the proposed algorithms was assessed using digital hemispherical photography (DHP) data and destructive measurements performed during the growing season of silage maize in agricultural fields of Ghale-Nou, southern Tehran, Iran, in the summer of 2019. The results on biophysical variables against validation data showed that the developed GPR-PSO algorithm outperformed other algorithms under study in terms of robustness and accuracy (0.917, 0.931, 0.882 using R2 and 0.627, 0.078, and 1.99 using RMSE in LAI, fCover, and biomass of Sentinel-2 20 m, respectively). GPR-PSO also possesses the unique ability to generate pixel-based uncertainty maps (confidence level) for prediction purposes (i.e., estimated uncertainty level &lt;0.7 in LAI, fCover, and biomass, for 96%, 98%, and 71% of the total study area, respectively). Altogether, GPR-PSO appears to be the most suitable option for mapping biophysical variables at the local scale using Sentinel-2 images.

Reconstructing Kernel-Based Machine Learning Force Fields with Superlinear Convergence.

Kernel machines have sustained continuous progress in the field of quantum chemistry. In particular, they have proven to be successful in the low-data regime of force field reconstruction. This is because many equivariances and invariances due to physical symmetries can be incorporated into the kernel function to compensate for much larger data sets. So far, the scalability of kernel machines has however been hindered by its quadratic memory and cubical runtime complexity in the number of training points. While it is known that iterative Krylov subspace solvers can overcome these burdens, their convergence crucially relies on effective preconditioners, which are elusive in practice. Effective preconditioners need to partially presolve the learning problem in a computationally cheap and numerically robust manner. Here, we consider the broad class of Nyström-type methods to construct preconditioners based on successively more sophisticated low-rank approximations of the original kernel matrix, each of which provides a different set of computational trade-offs. All considered methods aim to identify a representative subset of inducing (kernel) columns to approximate the dominant kernel spectrum.

Application of machine learning techniques to build digital twins for long train dynamics simulations

The paper shows the feasibility of building closed-form and fast-to-evaluate surrogate models via supervised kernel-based machine learning regressions behaving as digital twins for computationally expensive multibody simulations. The aforementioned surrogate models are adopted to predict the railway vehicle dynamics safety indexes defined in the international standards, depending on the wheel-rail forces, directly from the results of longitudinal train dynamics simulations. The digital twin models are trained with the outputs of Simpack multibody simulations of a reference freight wagon, to which the in-train forces calculated by an in-house MATLAB longitudinal train dynamics simulator are applied. Two machine learning techniques are considered: the support vector machine and the least-squares support vector machine regressions. Both techniques ensure a good accuracy even with a limited number of training samples. The derivation of the surrogate models can strongly enhance the modelling capabilities of common longitudinal train dynamics simulators, that cannot evaluate the wheel-rail contact forces. At the same time, the method shown in the paper allows to strongly reduce the total computational times, as the evaluation of the closed-form surrogate models allows to effectively estimate the safety indexes with no need for computationally demanding multibody simulations.

Machine learning aided catalyst activity modelling and design for direct conversion of CO2 to lower olefins

In the recent years, the demand for utilisation of CO2 into different chemicals has gathered interest due to increased concerns for global warming. The current work focuses on development of machine learning (ML) framework for catalyst modelling and design for direct conversion of CO2 to lower olefins (LO) based on the structural-composition-operating parameters. Comprehensive review, and data mining exercise was carried out and data base was formed from -55 relevant reports, including 18 input parameters and catalyst activity (i.e., CO2 conversion (%) & LO selectivity (%)) as output parameter. Artificial neural network (ANN) models were developed using Bayesian-Regularisation (BR) and Levenberg-Marquardt (LM) backpropagation learning algorithms for prediction of catalyst activity. Performance of the developed ANN models are compared with linear, tree-based, and kernel-based ML models and has been evaluated based on statistical measures. Out of these ML models, ANN-BR is able to predict CO2 conversion & LO selectivity with less deviation from experimental data (R = 0.90 & 0.8, RMSE = 8.43 & 16.73, AAD = 5.8 & 9.5 for test data respectively), compared to other ML models. Input contribution on post analysis of modelling is considered to understand the significance of predominate feature affecting the target variables. Further, integrated catalyst and process design carried out using inverse design based on multi-objective optimization (NSGA-II) with ANN-BR as objective function. Results indicate two-to-three-folds increase in yields with optimal catalyst composition, operating conditions, and novel combination of catalysts for efficient conversion of CO2 to lower olefins compared to reported experimental results.

Brain-age prediction: A systematic comparison of machine learning workflows

The difference between age predicted using anatomical brain scans and chronological age, i.e., the brain-age delta, provides a proxy for atypical aging. Various data representations and machine learning (ML) algorithms have been used for brain-age estimation. However, how these choices compare on performance criteria important for real-world applications, such as; (1) within-dataset accuracy, (2) cross-dataset generalization, (3) test-retest reliability, and (4) longitudinal consistency, remains uncharacterized. We evaluated 128 workflows consisting of 16 feature representations derived from gray matter (GM) images and eight ML algorithms with diverse inductive biases. Using four large neuroimaging databases covering the adult lifespan (total N = 2953, 18–88 years), we followed a systematic model selection procedure by sequentially applying stringent criteria. The 128 workflows showed a within-dataset mean absolute error (MAE) between 4.73–8.38 years, from which 32 broadly sampled workflows showed a cross-dataset MAE between 5.23–8.98 years. The test-retest reliability and longitudinal consistency of the top 10 workflows were comparable. The choice of feature representation and the ML algorithm both affected the performance. Specifically, voxel-wise feature spaces (smoothed and resampled), with and without principal components analysis, with non-linear and kernel-based ML algorithms performed well. Strikingly, the correlation of brain-age delta with behavioral measures disagreed between within-dataset and cross-dataset predictions. Application of the best-performing workflow on the ADNI sample showed a significantly higher brain-age delta in Alzheimer's and mild cognitive impairment patients compared to healthy controls. However, in the presence of age bias, the delta estimates in the patients varied depending on the sample used for bias correction. Taken together, brain-age shows promise, but further evaluation and improvements are needed for its real-world application.

Urban natural gas consumption forecasting by novel wavelet-kernelized grey system model

Natural gas is playing a key role in the Carbon Neutral path, which is clean and abundant. However it is difficult to collect sufficient data of urban natural gas consumption in China, and such data sets often present high nonlinearity and complex features, making it difficult to make accurate forecasts for the mid-small cities based on small samples. In this work, a novel wavelet kernel-based grey system model is proposed by using the wavelet kernel-based machine learning and the grey system modelling, taking advantage of the features of nonlinearity and periodicity of the wavelet kernel. A complete computational algorithm is presented by utilizing a hold-out cross validation-based grid-search scheme for selecting the optimal hyperparameters. Three case studies are carried out based on the real-world data sets of urban natural gas consumption in Kunming China, in which the proposed model outperforms other 15 time series forecasting models (including kernel-based models, grey system models and deep learning models), illustrating its priority in such forecasting tasks and high potential in similar applications.

Synergy of dual – polarimetric radar vegetation descriptor and Gaussian processes regression algorithm for estimation of leaf area index

ABSTRACT Gaussian Process Regression (GPR) emerged as a powerful algorithm since last decade in many applications. However, it is not fully explored in remote sensing applications, especially for predicting plant biophysical variables in the heterogeneous environment of India. This kernel-based machine learning technique has effectively replaced conventional approaches for estimating vegetation characteristics from remotely sensed data. In this work, an attempt has been made to test the ability of GPR to estimate the leaf area index (LAI) of wheat crops using the Sentinel-1 (S1) derived Dual-Polarized Radar Vegetation Index (DpRVI) and Sentinel-2 (S2) Top of Atmosphere (TOA) products. Further, the ability of the atmospheric correction procedure was tested by S2 Bottom of Atmosphere (BOA) images. To accomplish this, the field measurements of LAI were carried out from January to March 2020. Further comparisons of the GPR’s performance were made with the Artificial Neural Network (ANN) coupled with the PROSAIL (PROSPECT+SAIL) radiative transfer model, available through the Sentinel Application Platform (SNAP) Biophysical processor. The accuracy of the estimated LAI was evaluated using the statistical indicators, for example, coefficient of determination (R 2 ), Root Mean Square Error (RMSE), and Nash Sutcliffe Efficiency (NSE). The results showed that the synergy of the GPR and DpRVI provided the most accurate result (R 2 = 0.822, RMSE = 0.503 m2 m−2, NSE = 0.831) as compared to the GPR and TOA (R 2 = 0.816, RMSE = 0.596 m2 m−2, NSE = 0.803) and SNAP biophysical processor (based on ANN) (R 2 = 0.696, RMSE = 0.760 m2 m−2, NSE = 0.578). Therefore, the study demonstrated the importance of S1 SAR images and GPR as an alternative tool among the other well-established machine learning algorithms to estimate the crop biophysical parameters.

Quantum machine learning corrects classical forcefields: Stretching DNA base pairs in explicit solvent.

In order to improve the accuracy of molecular dynamics simulations, classical forcefields are supplemented with a kernel-based machine learning method trained on quantum-mechanical fragment energies. As an example application, a potential-energy surface is generalized for a small DNA duplex, taking into account explicit solvation and long-range electron exchange-correlation effects. A long-standing problem in molecular science is that experimental studies of the structural and thermodynamic behavior of DNA under tension are not well confirmed by simulation; study of the potential energy vs extension taking into account a novel correction shows that leading classical DNA models have excessive stiffness with respect to stretching. This discrepancy is found to be common across multiple forcefields. The quantum correction is in qualitative agreement with the experimental thermodynamics for larger DNA double helices, providing a candidate explanation for the general and long-standing discrepancy between single molecule stretching experiments and classical calculations of DNA stretching. The new dataset of quantum calculations should facilitate multiple types of nucleic acid simulation, and the associated Kernel Modified Molecular Dynamics method (KMMD) is applicable to biomolecular simulations in general. KMMD is made available as part of the AMBER22 simulation software.

Predicting the yield of stepped corrugated solar distiller using kernel-based machine learning models

• Stepped solar still with corrugated basin was designed and constructed. • Maximum distillate improvement was 255% for stepped still over conventional one. • A two machine learning models were developed for predicting both still performances. • Five kernel functions were used to find the optimum structure of each model. • Radial basis function (RBF) was the most appropriate kernel function. • Support vector machine combined RBF had a superior accuracy in predicting the still yield. Recently, accurate prediction of the distilled productivity is a decisive issue to assess and compare the ability of these designs of solar stills to provide distilled water to a certain application without carrying out time-consuming and more costly field experiments. Machine learning techniques have become powerful alternatives to conventional modeling approaches in different engineering disciplines. In this study, two improved machine learning models, namely relevance vector machine (RVM) and least squares support vector machine (LSSVM), are proposed to predict the productivity of two solar distillers. The two solar distillers are a conventional single slope single basin solar still (SSSS) and a modified stepped corrugated solar still (SCSS). Initially, the experimental energetic and exergetic performance of the SCSS is compared with that of SSSS. The experimental results showed that the water yield, energy efficiency, and exergy efficiency of the SCSS are improved by about 255 %, 239%, and 379%, respectively, compared with that of the SSSS. Then, the proposed RVM and LSSVM models are used to predict the productivity of two solar stills. During the simulation, different kernel functions, such as Laplace, linear, spline, Gaussian (G), and radial basis function (RBF), are embedded in both models to obtain the optimal function that maximizes the model accuracy. The simulated results showed that Gaussian and radial basis functions are the most appropriate kernel functions for RVM and LSSVM algorithms, respectively. Particularly, the LSSVM-RBF and RVM-G models predict the distilled production of the SCSS with determination coefficient values of 0.999 and 0.982, and root mean square error values of about 6.646 e-06, and 0.002, respectively Consequently, the LSSVM-RBF model is recommended as a promising prediction tool in predicting the production of both solar distillers.

Prediction of Maternal Hemorrhage Using Machine Learning: Retrospective Cohort Study

BackgroundPostpartum hemorrhage remains one of the largest causes of maternal morbidity and mortality in the United States.ObjectiveThe aim of this paper is to use machine learning techniques to identify patients at risk for postpartum hemorrhage at obstetric delivery.MethodsWomen aged 18 to 55 years delivering at a major academic center from July 2013 to October 2018 were included for analysis (N=30,867). A total of 497 variables were collected from the electronic medical record including the following: demographic information; obstetric, medical, surgical, and family history; vital signs; laboratory results; labor medication exposures; and delivery outcomes. Postpartum hemorrhage was defined as a blood loss of ≥1000 mL at the time of delivery, regardless of delivery method, with 2179 (7.1%) positive cases observed. Supervised learning with regression-, tree-, and kernel-based machine learning methods was used to create classification models based upon training (21,606/30,867, 70%) and validation (4630/30,867, 15%) cohorts. Models were tuned using feature selection algorithms and domain knowledge. An independent test cohort (4631/30,867, 15%) determined final performance by assessing for accuracy, area under the receiver operating curve (AUROC), and sensitivity for proper classification of postpartum hemorrhage. Separate models were created using all collected data versus models limited to data available prior to the second stage of labor or at the time of decision to proceed with cesarean delivery. Additional models examined patients by mode of delivery.ResultsGradient boosted decision trees achieved the best discrimination in the overall model. The model including all data mildly outperformed the second stage model (AUROC 0.979, 95% CI 0.971-0.986 vs AUROC 0.955, 95% CI 0.939-0.970). Optimal model accuracy was 98.1% with a sensitivity of 0.763 for positive prediction of postpartum hemorrhage. The second stage model achieved an accuracy of 98.0% with a sensitivity of 0.737. Other selected algorithms returned models that performed with decreased discrimination. Models stratified by mode of delivery achieved good to excellent discrimination but lacked the sensitivity necessary for clinical applicability.ConclusionsMachine learning methods can be used to identify women at risk for postpartum hemorrhage who may benefit from individualized preventative measures. Models limited to data available prior to delivery perform nearly as well as those with more complete data sets, supporting their potential utility in the clinical setting. Further work is necessary to create successful models based upon mode of delivery and to validate the findings of this study. An unbiased approach to hemorrhage risk prediction may be superior to human risk assessment and represents an area for future research.

Compact quantum kernel-based binary classifier

Quantum computing opens exciting opportunities for kernel-based machine learning methods, which have broad applications in data analysis. Recent works show that quantum computers can efficiently construct a model of a classifier by engineering the quantum interference effect to carry out the kernel evaluation in parallel. For practical applications of these quantum machine learning methods, an important issue is to minimize the size of quantum circuits. We present the simplest quantum circuit for constructing a kernel-based binary classifier. This is achieved by generalizing the interference circuit to encode data labels in the relative phases of the quantum state and by introducing compact amplitude encoding, which encodes two training data vectors into one quantum register. When compared to the simplest known quantum binary classifier, the number of qubits is reduced by two and the number of steps is reduced linearly with respect to the number of training data. The two-qubit measurement with post-selection required in the previous method is simplified to single-qubit measurement. Furthermore, the final quantum state has a smaller amount of entanglement than that of the previous method, which advocates the cost-effectiveness of our method. Our design also provides a straightforward way to handle an imbalanced data set, which is often encountered in many machine learning problems.

Fast and interpretable genomic data analysis using multiple approximate kernel learning.

MotivationDataset sizes in computational biology have been increased drastically with the help of improved data collection tools and increasing size of patient cohorts. Previous kernel-based machine learning algorithms proposed for increased interpretability started to fail with large sample sizes, owing to their lack of scalability. To overcome this problem, we proposed a fast and efficient multiple kernel learning (MKL) algorithm to be particularly used with large-scale data that integrates kernel approximation and group Lasso formulations into a conjoint model. Our method extracts significant and meaningful information from the genomic data while conjointly learning a model for out-of-sample prediction. It is scalable with increasing sample size by approximating instead of calculating distinct kernel matrices.ResultsTo test our computational framework, namely, Multiple Approximate Kernel Learning (MAKL), we demonstrated our experiments on three cancer datasets and showed that MAKL is capable to outperform the baseline algorithm while using only a small fraction of the input features. We also reported selection frequencies of approximated kernel matrices associated with feature subsets (i.e. gene sets/pathways), which helps to see their relevance for the given classification task. Our fast and interpretable MKL algorithm producing sparse solutions is promising for computational biology applications considering its scalability and highly correlated structure of genomic datasets, and it can be used to discover new biomarkers and new therapeutic guidelines.Availability and implementationMAKL is available at https://github.com/begumbektas/makl together with the scripts that replicate the reported experiments. MAKL is also available as an R package at https://cran.r-project.org/web/packages/MAKL.Supplementary information Supplementary data are available at Bioinformatics online.