Machine Learning-Assisted Determination of C6H14 Mole Fraction From Molecular Emissions of Laser-Induced Hexane-Air Plasmas

Laser-induced plasmas of materials containing hydrocarbons present strong carbon molecular emission features. Using these emissions to build models relating changes in spectral features to a physical parameter of the system, such as hydrocarbon content, can be difficult because of the dynamic complexity of the spectral features and temperature disequilibrium between molecular species. This study presents machine learning models trained to quantify the mole fraction of hexane in hexane-air plasmas from CN Violet and C2 Swan spectral features. Ensemble regression methods provide the most accurate predictions with root mean squared error on the order 10-2. Artificial neural network regressions produce predictions with superlative sensitivity, exhibiting detection limits as low as 0.008. These foundational models can be further refined with more advanced data to quantify the presence of carbon species in complex plasma environments, such as high-speed reacting flows.

Optimization of breeding program design through stochastic simulation with kernel regression

In recent years, breeding programs have increased significantly in size and complexity, with various highly interdependent parameters and many contrasting breeding goals. As a result, resource allocation in these programs has become more complex, and deriving an optimal breeding strategy has become increasingly challenging. To address this, a common practice is to reduce the optimization problem to a set of scenarios that differ only in a few parameters and can therefore be analyzed in detail. The goal of this article is to provide a framework for the numerical optimization of breeding programs that goes beyond the simple comparison of scenarios. For this, we first determine the space of potential breeding programs only limited by basic constraints like the budget and housing capacities. Subsequently, the goal is to identify the optimal breeding program by finding the parametrization that maximizes the target function by combining different breeding goals. To assess the value of the target function for a parametrization, we propose using stochastic simulations and the subsequent use of a kernel regression method to cope with the stochasticity of simulation outcomes. This procedure is performed iteratively to narrow down the most promising areas of the search space and perform more and more simulations in these areas of interest. In a simplified example applied to a dairy cattle program, our proposed framework has shown its ability to identify an optimal breeding strategy that aligns with a target function aiming at genetic gain and genetic diversity conservation limited by budget constraints.

Flexible joint modeling of mean and dispersion for the directional tuning of neuronal spike counts

The study of how the number of spikes in a middle temporal visual area (MT/V5) neuron is tuned to the direction of a visual stimulus has attracted considerable attention over the years, but recent studies suggest that the variability of the number of spikes might also be influenced by the directional stimulus. This entails that Poisson regression models are not adequate for this type of data, as the observations usually present over/underdispersion (or both) with respect to the Poisson distribution. This paper makes use of the double exponential family and presents a flexible model to estimate, jointly, the mean and dispersion functions, accounting for the effect of a circular covariate. The empirical performance of the proposal is explored via simulations and an application to a neurological data set is shown.

Distance-Metric Learning for Personalized Survival Analysis

Personalized time-to-event or survival prediction with right-censored outcomes is a pervasive challenge in healthcare research. Although various supervised machine learning methods, such as random survival forests or neural networks, have been adapted to handle such outcomes effectively, they do not provide explanations for their predictions, lacking interpretability. In this paper, an alternative method for survival prediction by weighted nearest neighbors is proposed. Fitting this model to data entails optimizing the weights by learning a metric. An individual prediction of this method can be explained by providing the user with the most influential data points for this prediction, i.e., the closest data points and their weights. The strengths and weaknesses in terms of predictive performance are highlighted on simulated data and an application of the method on two different real-world datasets of breast cancer patients shows its competitiveness with established methods.

Associating somatic mutation with clinical outcomes through kernel regression and optimal transport

Somatic mutations in cancer patients are inherently sparse and potentially high dimensional. Cancer patients may share the same set of deregulated biological processes perturbed by different sets of somatically mutated genes. Therefore, when assessing the associations between somatic mutations and clinical outcomes, gene-by-gene analysis is often under-powered because it does not capture the complex disease mechanisms shared across cancer patients. Rather than testing genes one by one, an intuitive approach is to aggregate somatic mutation data of multiple genes to assess their joint association with clinical outcomes. The challenge is how to aggregate such information. Building on the optimal transport method, we propose a principled approach to estimate the similarity of somatic mutation profiles of multiple genes between tumor samples, while accounting for gene-gene similarities defined by gene annotations or empirical mutational patterns. Using such similarities, we can assess the associations between somatic mutations and clinical outcomes by kernel regression. We have applied our method to analyze somatic mutation data of 17 cancer types and identified at least five cancer types, where somatic mutations are associated with overall survival, progression-free interval, or cytolytic activity.

In silico and in vitro assays reveal potential inhibitors against 3CLpro main protease of SARS-CoV-2

The COVID-19 pandemic, caused by the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is not showing any sign of slowing down even after the ongoing efforts of vaccination. The threats of new strains are concerning, as some of them are more infectious than the original one. A therapeutic against the disease is, therefore, of urgent need. Here, we use the DrugBank database to screen for potential inhibitors against the 3CL^pro main protease of SARS-CoV-2. Instead of using the traditional approach of computational screening by docking, we developed a kernel ridge regressor (using a part of the docking data) to predict the binding energy of ligands. We used this model to screen the DrugBank database and shortlist two lead candidates (bromocriptine and avoralstat) for in vitro enzymatic study. Our results show that the 3CL^pro enzyme activity in presence of 100 μM concentration of bromocriptine and avoralstat is 9.9% and 15.9%, respectively. Remarkably, bromocriptine exhibited submicromolar IC₅₀ of 130 nM (0.13 μM). Avoralstat showed an IC₅₀ of 2.16 μM. Further, the interactions of both drugs with 3CL^pro were analyzed using molecular dynamics simulations of 100 ns. Results indicate that both ligands are stable in the binding pocket of the 3CL^pro receptor. In addition, the MM-PBSA analysis revealed that bromocriptine (-29.37 kcal/mol) has a lower binding free energy compared to avoralstat (-6.91 kcal/mol). Further, hydrogen bond analysis also showed that bromocriptine interacts with the two catalytic residues, His⁴¹ and Cys¹⁴⁵, more frequently than avoralstat.Communicated by Ramaswamy H. Sarma.

Robust neural decoding by kernel regression with Siamese representation learning

Objective. Brain-machine interfaces (BMIs) provide a direct pathway between the brain and external devices such as computer cursors and prosthetics, which have great potential in motor function restoration. One critical limitation of current BMI systems is the unstable performance, partly due to the variability of neural signals. Studies showed that neural activities exhibit trial-to-trial variability, and the preferred direction of neurons frequently changes under different conditions. Therefore, a fixed decoding function does not work well.Approach. To deal with the problems, we propose a novel kernel regression framework. The nonparametric kernel regression is used to fit diverse decoding functions by finding similar neural patterns to handle neural variations caused by varying tuning functions. Further, the representations of raw neural signals are learned by Siamese networks and constrained by kinematic parameters, which can alleviate neural variations caused by intrinsic noises and task-irrelevant information. The representations are jointly learned with the kernel regression framework in an end-to-end manner so that neural variations can be tackled effectively.Main results. Experiments on two datasets demonstrate that our approach outperforms most existing methods and significantly improves the robustness in challenging situations such as limited samples and missing channels.Significance. The proposed approach demonstrates robust performance with different conditions and provides a new and inspiring perspective toward robust BMI control.

Sub-Graph Regularization on Kernel Regression for Robust Semi-Supervised Dimensionality Reduction

Dimensionality reduction has always been a major problem for handling huge dimensionality datasets. Due to the utilization of labeled data, supervised dimensionality reduction methods such as Linear Discriminant Analysis tend achieve better classification performance compared with unsupervised methods. However, supervised methods need sufficient labeled data in order to achieve satisfying results. Therefore, semi-supervised learning (SSL) methods can be a practical selection rather than utilizing labeled data. In this paper, we develop a novel SSL method by extending anchor graph regularization (AGR) for dimensionality reduction. In detail, the AGR is an accelerating semi-supervised learning method to propagate the class labels to unlabeled data. However, it cannot handle new incoming samples. We thereby improve AGR by adding kernel regression on the basic objective function of AGR. Therefore, the proposed method can not only estimate the class labels of unlabeled data but also achieve dimensionality reduction. Extensive simulations on several benchmark datasets are conducted, and the simulation results verify the effectiveness for the proposed work.

Microscopic evaluation of traffic safety at signal coordinated intersections: A before-after study

OBJECTIVE

This research aims to evaluate the safety impacts of signal coordination on signalized intersections and provide a scientific basis to design and improve signal control and management from a traffic safety perspective.

METHODS

A kernel regression model is adopted to evaluate the safety performance of intersections before and after implementing the signal coordination strategy. By using this statistical method, the authors identify the nonlinear relationship between crash frequency and the crash's spatial location and examine the discrepancy of crash spatial distributions between the coordination and noncoordination conditions at disaggregated levels, such as time of day and crash type. A case study is presented with the use of Michigan crash data (2003-2007).

RESULTS

The study finds that the (1) crash distribution on arterials tends to be spatially disperse when the signal coordination is in operation and (2) crash frequency at the approaches of intersections is increased with the use of signal coordination under the following conditions: Nonpeak hours, rear-end and sideswipe crashes, intersections with low speed limits, and both injury and property damage-only crashes.

CONCLUSION

Signal coordination poses safety concerns in addition to its operational benefits for intersections.

Machine learning of accurate energy-conserving molecular force fields

Using conservation of energy-a fundamental property of closed classical and quantum mechanical systems-we develop an efficient gradient-domain machine learning (GDML) approach to construct accurate molecular force fields using a restricted number of samples from ab initio molecular dynamics (AIMD) trajectories. The GDML implementation is able to reproduce global potential energy surfaces of intermediate-sized molecules with an accuracy of 0.3 kcal mol-1 for energies and 1 kcal mol-1 Å̊-1 for atomic forces using only 1000 conformational geometries for training. We demonstrate this accuracy for AIMD trajectories of molecules, including benzene, toluene, naphthalene, ethanol, uracil, and aspirin. The challenge of constructing conservative force fields is accomplished in our work by learning in a Hilbert space of vector-valued functions that obey the law of energy conservation. The GDML approach enables quantitative molecular dynamics simulations for molecules at a fraction of cost of explicit AIMD calculations, thereby allowing the construction of efficient force fields with the accuracy and transferability of high-level ab initio methods.

Abundant Inverse Regression using Sufficient Reduction and its Applications

Statistical models such as linear regression drive numerous applications in computer vision and machine learning. The landscape of practical deployments of these formulations is dominated by forward regression models that estimate the parameters of a function mapping a set of p covariates, x , to a response variable, y. The less known alternative, Inverse Regression, offers various benefits that are much less explored in vision problems. The goal of this paper is to show how Inverse Regression in the "abundant" feature setting (i.e., many subsets of features are associated with the target label or response, as is the case for images), together with a statistical construction called Sufficient Reduction, yields highly flexible models that are a natural fit for model estimation tasks in vision. Specifically, we obtain formulations that provide relevance of individual covariates used in prediction, at the level of specific examples/samples - in a sense, explaining why a particular prediction was made. With no compromise in performance relative to other methods, an ability to interpret why a learning algorithm is behaving in a specific way for each prediction, adds significant value in numerous applications. We illustrate these properties and the benefits of Abundant Inverse Regression (AIR) on three distinct applications.

KERNEL PARTIAL LEAST SQUARES REGRESSION FOR RELATING FUNCTIONAL BRAIN NETWORK TOPOLOGY TO CLINICAL MEASURES OF BEHAVIOR

In this paper we present a novel method for analyzing the relationship between functional brain networks and behavioral phenotypes. Drawing from topological data analysis, we first extract topological features using persistent homology from functional brain networks that are derived from correlations in resting-state fMRI. Rather than fixing a discrete network topology by thresholding the connectivity matrix, these topological features capture the network organization across all continuous threshold values. We then propose to use a kernel partial least squares (kPLS) regression to statistically quantify the relationship between these topological features and behavior measures. The kPLS also provides an elegant way to combine multiple image features by using linear combinations of multiple kernels. In our experiments we test the ability of our proposed brain network analysis to predict autism severity from rs-fMRI. We show that combining correlations with topological features gives better prediction of autism severity than using correlations alone.

Nonparametric Estimates of Gene × Environment Interaction Using Local Structural Equation Modeling

Gene × environment (G × E) interaction studies test the hypothesis that the strength of genetic influence varies across environmental contexts. Existing latent variable methods for estimating G × E interactions in twin and family data specify parametric (typically linear) functions for the interaction effect. An improper functional form may obscure the underlying shape of the interaction effect and may lead to failures to detect a significant interaction. In this article, we introduce a novel approach to the behavior genetic toolkit, local structural equation modeling (LOSEM). LOSEM is a highly flexible nonparametric approach for estimating latent interaction effects across the range of a measured moderator. This approach opens up the ability to detect and visualize new forms of G × E interaction. We illustrate the approach by using LOSEM to estimate gene × socioeconomic status interactions for six cognitive phenotypes. Rather than continuously and monotonically varying effects as has been assumed in conventional parametric approaches, LOSEM indicated substantial nonlinear shifts in genetic variance for several phenotypes. The operating characteristics of LOSEM were interrogated through simulation studies where the functional form of the interaction effect was known. LOSEM provides a conservative estimate of G × E interaction with sufficient power to detect statistically significant G × E signal with moderate sample size. We offer recommendations for the application of LOSEM and provide scripts for implementing these biometric models in Mplus and in OpenMx under R.

Empirical Likelihood for Estimating Equations with Nonignorably Missing Data

We develop an empirical likelihood (EL) inference on parameters in generalized estimating equations with nonignorably missing response data. We consider an exponential tilting model for the nonignorably missing mechanism, and propose modified estimating equations by imputing missing data through a kernel regression method. We establish some asymptotic properties of the EL estimators of the unknown parameters under different scenarios. With the use of auxiliary information, the EL estimators are statistically more efficient. Simulation studies are used to assess the finite sample performance of our proposed EL estimators. We apply our EL estimators to investigate a data set on earnings obtained from the New York Social Indicators Survey.

NONPARAMETRIC COVARIANCE MODEL

There has been considerable attention on estimation of conditional variance function in the literature. We propose here a nonparametric model for conditional covariance matrix. A kernel estimator is developed accordingly, its asymptotic bias and variance are derived, and its asymptotic normality is established. A real data example is used to illustrate the proposed estimation procedure.