Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Feature importance”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

A comparison of model validation approaches for echo state networks using climate model replicates

As global temperatures continue to rise, climate mitigation strategies such as stratospheric aerosol injections (SAI) are increasingly discussed, but the downstream effects of these strategies are not well understood. As such, there is interest in developing statistical methods to quantify the evolution of climate variable relationships during the time period surrounding an SAI. Feature importance applied to echo state network (ESN) models has been proposed as a way to understand the effects of SAI using a data-driven model. This approach depends on the ESN fitting the data well. If not, the feature importance may place importance on features that are not representative of the underlying relationships. Typically, time series prediction models such as ESNs are assessed using out-of-sample performance metrics that divide the times series into separate training and testing sets. However, this model assessment approach is geared towards forecasting applications and not scenarios such as the motivating SAI example where the objective is using a data driven model to capture variable relationships. Here, in this paper, we demonstrate a novel use of climate model replicates to investigate the applicability of the commonly used repeated hold-out model assessment approach for the SAI application. Simulations of an SAI are generated using a simplified climate model, and different initialization conditions are used to provide independent training and testing sets containing the same SAI event. The climate model replicates enable out-of-sample measures of model performance, which are compared to the single time series hold-out validation approach. For our case study, it is found that the repeated hold-out sample performance is comparable, but conservative, to the replicate out-of-sample performance when the training set contains enough time after the aerosol injection.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Semi-supervised Bayesian Low-shot Learning

Deep neural networks (NNs) typically outperform traditional machine learning (ML) approaches for complicated, non-linear tasks. It is expected that deep learning (DL) should offer superior performance for the important non-proliferation task of predicting explosive device configuration based upon observed optical signature, a task which human experts struggle with. However, supervised machine learning is difficult to apply in this mission space because most recorded signatures are not associated with the corresponding device description, or “truth labels.” This is challenging for NNs, which traditionally require many samples for strong performance. Semi-supervised learning (SSL), low-shot learning (LSL), and uncertainty quantification (UQ) for NNs are emerging approaches that could bridge the mission gaps of few labels and rare samples of importance. NN explainability techniques are important in gaining insight into the inferential feature importance of such a complex model. In this work, SSL, LSL, and UQ are merged into a single framework, a significant technical hurdle not previously demonstrated. Exponential Average Adversarial Training (EAAT) and Pairwise Neural Networks (PNNs) are chosen as the SSL and LSL methods of choice. Permutation feature importance (PFI) for functional data is used to provide explainability via the Variable importance Explainable Elastic Shape Analysis (VEESA) pipeline. A variety of uncertainty quantification approaches are explored: Bayesian Neural Networks (BNNs), ensemble methods, concrete dropout, and evidential deep learning. Two final approaches, one utilizing ensemble methods and one utilizing evidential learning, are constructed and compared using a well-quantified synthetic 2D dataset along with the DIRSIG Megascene.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Using diverse potentials and scoring functions for the development of improved machine-learned models for protein–ligand affinity and docking pose prediction

The advent of computational drug discovery holds the promise of significantly reducing the effort of experimentalists, along with monetary cost. More generally, predicting the binding of small organic molecules to biological macromolecules has far-reaching implications for a range of problems, including metabolomics. However, problems such as predicting the bound structure of a protein–ligand complex along with its affinity have proven to be an enormous challenge. In recent years, machine learning-based methods have proven to be more accurate than older methods, many based on simple linear regression. Nonetheless, there remains room for improvement, as these methods are often trained on a small set of features, with a single functional form for any given physical effect, and often with little mention of the rationale behind choosing one functional form over another. Moreover, it is not entirely clear why one machine learning method is favored over another. Here, we endeavor to undertake a comprehensive effort towards developing high-accuracy, machine-learned scoring functions, systematically investigating the effects of machine learning method and choice of features, and, when possible, providing insights into the relevant physics using methods that assess feature importance. Here, we show synergism among disparate features, yielding adjusted R 2 with experimental binding affinities of up to 0.871 on an independent test set and enrichment for native bound structures of up to 0.913. When purely physical terms that model enthalpic and entropic effects are used in the training, we use feature importance assessments to probe the relevant physics and hopefully guide future investigators working on this and other computational chemistry problems.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Recombination smooths the time-signal disrupted by latency in within-host HIV phylogenies

Within-host HIV evolution involves several features that may disrupt standard phylogenetic reconstruction. One important feature is re-activation of latently integrated provirus, which has the potential to disrupt the temporal signal, leading to variation in the branch lengths and apparent evolutionary rates in a tree. Yet, real within-host HIV phylogenies tend to show clear, ladder-like trees structured by the time of sampling. Another important feature is recombination, which violates the fundamental assumption that evolutionary history can be represented by a single bifurcating tree. Thus, recombination complicates the within-host HIV dynamic by mixing genomes and creating evolutionary loop structures that cannot be represented in bifurcating trees. In this paper, we develop a coalescent-based simulator of within-host HIV evolution that includes latency, recombination, and effective population size dynamics that allows us to study the relationship between the true, complex genealogy of within-host HIV evolution, encoded as an Ancestral Recombination Graph (ARG), and the observed phylogenetic tree. To compare our ARG results to the familiar phylogeny format, we calculate the expected bifurcating tree after decomposing the ARG into all unique site trees, their combined distance matrix, and the overall corresponding bifurcating tree. While latency and recombination separately disrupt the phylogenetic signal, remarkably, we find that recombination recovers the temporal signal of within-host HIV evolution caused by latency by mixing fragments of old, latent genomes into the contemporary population. In effect, recombination averages over extant heterogeneity, whether it stems from mixed time-signals or population bottlenecks. Further, we establish that the signals of latency and recombination can be observed in phylogenetic trees despite being an incorrect representation of the true evolutionary history. Using an Approximate Bayesian Computation method, we develop a set of statistical probes to tune our simulation model to nine longitudinally-sampled within-host HIV phylogenies. Because ARGs are exceedingly difficult to infer from real HIV data, our simulation system allows investigating effects of latency, recombination, and population size bottlenecks by matching decomposed ARGs to real data as observed in standard phylogenies.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

A Comparison of Infectious Disease Forecasting Methods across Locations, Diseases, and Time

Accurate infectious disease forecasting can inform efforts to prevent outbreaks and mitigate adverse impacts. This study compares the performance of statistical, machine learning (ML), and deep learning (DL) approaches in forecasting infectious disease incidences across different countries and time intervals. We forecasted three diverse diseases: campylobacteriosis, typhoid, and Q-fever, using a wide variety of features (n = 46) from public datasets, e.g., landscape, climate, and socioeconomic factors. We compared autoregressive statistical models to two tree-based ML models (extreme gradient boosted trees [XGB] and random forest [RF]) and two DL models (multi-layer perceptron and encoder–decoder model). The disease models were trained on data from seven different countries at the region-level between 2009–2017. Forecasting performance of all models was assessed using mean absolute error, root mean square error, and Poisson deviance across Australia, Israel, and the United States for the months of January through August of 2018. The overall model results were compared across diseases as well as various data splits, including country, regions with highest and lowest cases, and the forecasted months out (i.e., nowcasting, short-term, and long-term forecasting). Overall, the XGB models performed the best for all diseases and, in general, tree-based ML models performed the best when looking at data splits. There were a few instances where the statistical or DL models had minutely smaller error metrics for specific subsets of typhoid, which is a disease with very low case counts. Feature importance per disease was measured by using four tree-based ML models (i.e., XGB and RF with and without region name as a feature). The most important feature groups included previous case counts, region name, population counts and density, mortality causes of neonatal to under 5 years of age, sanitation factors, and elevation. This study demonstrates the power of ML approaches to incorporate a wide range of factors to forecast various diseases, regardless of location, more accurately than traditional statistical approaches.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Machine Learning Augmented Predictive and Generative Model for Rupture Life in Ferritic and Austenitic Steels

The Larson-Miller parameter (LMP) offers an efficient and fast scheme to estimate the creep rupture life of alloy materials for high temperature applications. However, owing to poor generalizability and dependence on the constant C, which is typically not known a-priori, estimations using the Larson-Miller parameter often result in suboptimal performance for a wide range of materials. At best it is useful in comparing alloys of similar composition. In this work, three machine learning (ML) schemes were developed for rupture life prediction for 9-12% Cr ferritic-martensitic steels and austenitic stainless steels, i.e., a hierarchical model to parameterize LMP using the LMP constant C to compute rupture life, a hierarchical model to parameterize both C and LMP to compute rupture life, and a direct prediction of rupture life. Specifically, we show that the third scheme, using a gradient boosting algorithm, can be used to train ML models for very accurate prediction of rupture life in a variety of alloys (Pear-son Correlation Coefficient > 0.9 for 9-12% Cr and > 0.8 for austenitic stainless steels). In addition, the Shapley value was used to quantify feature importance, making the model interpretable by identifying the effect of various features on the model performance. Furthermore, a variational autoencoder-based generative model was built by conditioning on the experimental dataset to sample hypothetical synthetic candidate alloys from the learnt joint distribution not existing in both 9-12 % Cr ferritic-martensitic steel and austenitic stainless steel datasets. Finally, based on the predictive and generative model, a reinforcement learning strategy has been proposed to guide experimentalists into designing better heat resistant alloys.

Mamun, Md Osman G.↗
DOE OSTI2026

Mitigating Algorithmic Bias in Cancer Site Classification Models

Purpose Integrating artificial intelligence in cancer diagnostics has improved tumor classification beyond rule-based systems. Despite these advancements, these models may still encode demographic biases. We conducted a large-scale, applied bias-probing study of a deep learning–based cancer site classifier to quantify race information encoded in document embeddings. We then evaluated how performance changes when race-correlated embedding dimensions are removed in a post-training sensitivity analysis. Methods The cancer site classifier was trained using 3.5 million electronic cancer pathology reports from six of the National Cancer Institute's SEER registries. We trained a hierarchical self-attention network to generate 400-dimensional document embeddings. These embeddings were used to train two downstream, gradient-boosted decision tree classifiers: one to classify the cancer sites and another to predict racial categories. We identified overlapping features by intersecting the top 50 feature-importance rankings from the site and race models and computed their cumulative feature importance in each model. As a post hoc sensitivity analysis, we progressively pruned these overlapping dimensions, retrained the site model, and compared overall macro-F1 and accuracy, race-stratified macro-F1, and group fairness metrics on the basis of demographic parity and equalized odds before and after pruning. Results The analysis revealed minimal feature overlap between the cancer site and race prediction models, and the cumulative importance scores indicated a negligible influence of racial information on clinical predictions. Post-training pruning of overlapping features did not compromise the models' diagnostic accuracy, with a 0.07% loss in accuracy. Conclusion Our findings demonstrate that HiSAN-generated embeddings from SEER data can be used effectively in cancer site classification without significant demographic bias influencing the outcomes. Post-training pruning therefore functions as a practical audit and sensitivity check.

Shivanna, Abhishek [ORNL] (ORCID:0009000665228593)↗
DOE OSTI2025

XSub: Explanation-Driven Adversarial Attack against Blackbox Classifiers via Feature Substitution

Despite its significant benefits in enhancing the transparency and trustworthiness of artificial intelligence (AI) systems, explainable AI (XAI) can unintentionally provide adversaries with insights into blackbox models, increasing their vulnerability to various attacks. In this paper, we develop a novel explanation-driven adversarial attack against blackbox classifiers based on feature substitution, called XSub. The key idea of XSub is to strategically replace important features (identified via XAI) in the original sample with corresponding important features of a different label, thereby increasing the likelihood of the model misclassifying the perturbed sample. XSub only requires a minimal number of queries and can be easily extended to launch backdoor attacks in case the attacker has access to the model's training data. Our evaluation shows that XSub is not only effective and stealthy but also low-cost, showcasing its feasibility across a wide range of AI applications.

adversarial attack↗
DOE OSTI2025

Data Agnostic Feature-Target Analysis & Ranking Machine Learning Pipeline (DAFTAR-ML) v0.1.0

DAFTAR-ML is a specialized machine-learning pipeline that identifies relevant features based on their relationship to a target variable. Many ML pipelines focus solely on prediction, and feature ranking is often absent or lacks robust statistical methods. DAFTAR-ML performs its tasks with this outcome in mind. Model training is robust, using nested cross-validation and hyperparameter tuning. Instead of relying on native feature-importance scores, it employs SHAP (SHapley Additive exPlanations) to quantify feature importance. The pipeline also produces comprehensive results, including publication-quality visualizations.

Melie, Tina [Lawrence Berkeley National Laboratory↗
DOE OSTI2024

listenr

SAND2024-13902O listenr is an R package containing code that allows users to fit echo state networks (a machine learning model) on data to obtain predictions. The code also allows users to compute spatio-temporal feature importance on the echo state network, as described in “Characterizing Climate Pathways Using Feature Importance on Echo State Networks.” The package also provides a function for computing principal components.

Ries, Daniel↗
DOE OSTI2020

Different tertiary interactions create the same important 3D features in a distinct flavivirus xrRNA

During infection by a flavivirus (FV), cells accumulate noncoding subgenomic flavivirus RNAs (sfRNAs) that interfere with several antiviral pathways. These sfRNAs are formed by structured RNA elements in the 3' untranslated region (UTR) of the viral genomic RNA, which block the progression of host cell exoribonucleases that have targeted the viral RNA. Previous work on these exoribonuclease-resistant RNAs (xrRNAs) from mosquito-borne FVs revealed a specific three-dimensional fold with a unique topology in which a ring-like structure protectively encircles the 5' end of the xrRNA. Conserved nucleotides make specific tertiary interactions that support this fold. Examination of more divergent FVs reveals differences in their 3' UTR sequences, raising the question of whether they contain xrRNAs and if so, how they fold. To answer this, we demonstrated the presence of an authentic xrRNA in the 3' UTR of the Tamana bat virus (TABV) and solved its structure by X-ray crystallography. The structure reveals conserved features from previously characterized xrRNAs, but in the TABV version these features are created through a novel set of tertiary interactions not previously seen in xrRNAs. This includes two important A–C interactions, four distinct backbone kinks, several ordered Mg 2+ ions, and a C + –G–C base triple. The discovery that the same overall architecture can be achieved by very different sequences and interactions in distantly related flaviviruses provides insight into the diversity of this type of RNA and will inform searches for undiscovered xrRNAs in viruses and beyond.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Classifying thermodynamic cloud phase using machine learning models

Vertically resolved thermodynamic cloud-phase classifications are essential for studies of atmospheric cloud and precipitation processes. The Department of Energy (DOE) Atmospheric Radiation Measurement (ARM) Thermodynamic Cloud Phase (THERMOCLDPHASE) value-added product (VAP) uses a multi-sensor approach to classify the thermodynamic cloud phase by combining lidar backscatter and depolarization, radar reflectivity, Doppler velocity, spectral width, microwave-radiometer-derived liquid water path, and radiosonde temperature measurements. The measured pixels are classified as ice, snow, mixed phase, liquid (cloud water), drizzle, rain, and liq_driz (liquid+drizzle). We use this product as the ground truth to train three machine learning (ML) models to predict the thermodynamic cloud phase from multi-sensor remote sensing measurements taken at the ARM North Slope of Alaska (NSA) observatory: a random forest (RF), a multi-layer perceptron (MLP), and a convolutional neural network (CNN) with a U-Net architecture. Evaluations against the outputs of the THERMOCLDPHASE VAP with 1 year of data show that the CNN outperforms the other two models, achieving the highest test accuracy, F1 score, and mean intersection over union (IOU). Analysis of ML confidence scores shows that ice, rain, and snow have higher confidence scores, followed by liquid, while mixed, drizzle, and liq_driz have lower scores. Feature importance analysis reveals that the mean Doppler velocity and vertically resolved temperature are the most influential data streams for ML thermodynamic cloud-phase predictions. Lidar measurements exhibit lower feature importance due to rapid signal attenuation caused by the frequent presence of persistent low-level clouds at the NSA site. The ML models' generalization capacity is further evaluated by applying them at another Arctic ARM site in Norway using data taken during the ARM Cold-Air Outbreaks in the Marine Boundary Layer Experiment (COMBLE) field campaign. The models demonstrated similar performance to that observed at the NSA site. Finally, we evaluate the ML models' response to simulated instrument outages and signal degradation and show that a CNN U-Net model trained with input channel dropouts performs better when input fields are missing.

ARM Aerial Facility↗
DOE OSTI2021

Development of a Deep Learning Model for Predicting the Drag Coefficients of Spherical and Non-Spherical Particles,

There is yet to be a well-established drag model for non-spherical particles required in a particle-laden flow that could cover a wide range of sphericities. This talk will explore the development of a general drag model for non-spherical particles by applying deep learning using available experimental data available in the literature. The integration of several raw experimental measurements from different sources and research directions allows the training of robust Artificial Intelligence and Machine Learning (ML) models. Neural networks are an ML approach inspired by the inner biological workings of the brain. This work aims to develop a Deep Neural Network (DNN) that predicts drag coefficient values with the ability to adapt appropriately to unseen data. Given the limited number of data points available and the variance found within the data collected from various sources, challenges may arise when looking to train the model. Our study tests and implements various model regularization techniques and assesses different loss and activation functions for the proposed DNN. The proposed model considers a broader range of features other than sphericity and Reynold number. These features include density ratio, solid volume fraction, lengthwise and crosswise sphericity, and more. Furthermore, we present the features that play a significant role in predicting different drag coefficients through feature importance. Within the investigated parameter ranges in this study, the following conclusions can be achieved and summarized below: • An improved drag coefficient model can be developed by considering more features such as, aspect ratio, lengthwise sphericity, crosswise sphericity, and density ratio. • DNN model can predict better results compared to traditional methods using MAE metric. • The proposed model addresses data challenges such as limited data and extreme data points through expanded feature-set and regularization. • Three major features that mostly affect the drag coefficient were identified from a feature importance analysis.

Presa-Reyes, Maria↗
DOE OSTI2022

SDSS-IV MaNGA: Unveiling Galaxy Interaction by Merger Stages with Machine Learning

We use machine-learning techniques to classify galaxy merger stages, which can unveil physical processes that drive the star formation and active galactic nucleus (AGN) activities during galaxy interaction. The sample contains 4690 galaxies from the integral field spectroscopy survey SDSS-IV MaNGA and can be separated into 1060 merging galaxies and 3630 nonmerging or unclassified galaxies. For the merger sample, there are 468, 125, 293, and 174 galaxies (1) in the incoming pair phase, (2) in the first pericentric passage phase, (3) approaching or just passing the apocenter, and (4) in the final coalescence phase or post-mergers. With the information of projected separation, line-of-sight velocity difference, Sloan Digital Sky Survey (SDSS) gri images, and MaNGA Hα velocity map, we are able to classify the mergers and their stages with good precision, which is the most important score to identify interacting galaxies. For the two-phase classification (binary; nonmerger and merger), the performance can be high (precision > 0.90) with LGBMClassifier . We find that sample size can be increased by rotation, so the five-phase classification (nonmerger, and merger stages 1, 2, 3, and 4) can also be good (precision > 0.85). The most important features come from SDSS gri images. The contribution from the MaNGA Hα velocity map, projected separation, and line-of-sight velocity difference can further improve the performance by 0%–20%. In other words, the image and the velocity information are sufficient to capture important features of galaxy interactions, and our results can apply to all the MaNGA data, as well as future all-sky surveys.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Machine learning enables identification of an alternative yeast galactose utilization pathway

How genomic differences contribute to phenotypic differences is a major question in biology. The recently characterized genomes, isolation environments, and qualitative patterns of growth on 122 sources and conditions of 1,154 strains from 1,049 fungal species (nearly all known) in the yeast subphylum Saccharomycotina provide a powerful, yet complex, dataset for addressing this question. We used a random forest algorithm trained on these genomic, metabolic, and environmental data to predict growth on several carbon sources with high accuracy. Known structural genes involved in assimilation of these sources and presence/absence patterns of growth in other sources were important features contributing to prediction accuracy. By further examining growth on galactose, we found that it can be predicted with high accuracy from either genomic (92.2%) or growth data (82.6%) but not from isolation environment data (65.6%). Prediction accuracy was even higher (93.3%) when we combined genomic and growth data. After the GALactose utilization genes, the most important feature for predicting growth on galactose was growth on galactitol, raising the hypothesis that several species in two orders, Serinales and Pichiales (containing the emerging pathogen Candida auris and the genus Ogataea, respectively), have an alternative galactose utilization pathway because they lack the GAL genes. Growth and biochemical assays confirmed that several of these species utilize galactose through an alternative oxidoreductive D-galactose pathway, rather than the canonical GAL pathway. Machine learning approaches are powerful for investigating the evolution of the yeast genotype–phenotype map, and their application will uncover novel biology, even in well-studied traits.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

ML Pipeline (Machine Learning Pipeline) [SWR-22-35]

The Machine Learning Pipeline (ML Pipeline) allows the user to fit a model to predict a dependent variable, y, based on a feature set, X. ML Pipeline gives the ability to automatically generate features with its Feature Engineering function, automatically select the most important features with its Feature Selection function, fit the model with its Fit function, and evaluate the model with its Evaluate function. ML Pipeline's functionality can not only help the user predict values for the desired variable it also gives the user a better understanding of which features are important and how effective the model is. ML Pipeline accomplishes all this, while remaining simple to use and interpret.

Schiek, Andrew↗
DOE OSTI2026

Data for FUN-PROSE: A Deep Learning Approach to Predict Condition-Specific Gene Expression in Fungi

mRNA levels of all genes in a genome is a critical piece of information defining the overall state of the cell in a given environmental condition. Being able to reconstruct such condition-specific expression in fungal genomes is particularly important to metabolically engineer these organisms to produce desired chemicals in industrially scalable conditions. Most previous deep learning approaches focused on predicting the average expression levels of a gene based on its promoter sequence, ignoring its variation across different conditions. Here we present FUN-PROSE—a deep learning model trained to predict differential expression of individual genes across various conditions using their promoter sequences and expression levels of all transcription factors. We train and test our model on three fungal species and get the correlation between predicted and observed condition-specific gene expression as high as 0.85. We then interpret our model to extract promoter sequence motifs responsible for variable expression of individual genes. We also carried out input feature importance analysis to connect individual transcription factors to their gene targets. A sizeable fraction of both sequence motifs and TF-gene interactions learned by our model agree with previously known biological information, while the rest corresponds to either novel biological facts or indirect correlations.

Genomics↗
DOE OSTI2021

Data-Driven Distributed Algorithms for Estimating Eigenvalues and Eigenvectors of Interconnected Dynamical Systems

Here, the paper presents data-driven algorithms to estimate in a distributed manner the eigenvalues, right and left eigenvectors of an unknown linear (or linearized) interconnected dynamic system. In particular, the proposed algorithms do not require the identification of the system model in advance before performing the estimation. As a first step, we consider interconnected dynamical system with distinct eigenvalues. The proposed strategy first estimates the eigenvalues using the well-known Prony method. The right and left eigenvectors are then estimated by solving distributively a set of linear equations. One important feature of the proposed algorithms is that the topology of communication network used to perform the distributed estimation can be chosen arbitrarily, given that it is connected, and is also independent of the structure or sparsity of the system (state) matrix. The proposed distributed algorithms are demonstrated via a numerical example.

97 MATHEMATICS AND COMPUTING↗