Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “classification models”

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 631 records · Page 35
All sourcesNASA NTRSDOE OSTI
NASA NTRS1990

Remote sensing of Earth terrain

Remote sensing of earth terrain is examined. The layered random medium model is used to investigate the fully polarimetric scattering of electromagnetic waves from vegetation. The model is used to interpret the measured data for vegetation fields such as rice, wheat, or soybean over water or soil. Accurate calibration of polarimetric radar systems is essential for the polarimetric remote sensing of earth terrain. A polarimetric calibration algorithm using three arbitrary in-scene reflectors is developed. In the interpretation of active and passive microwave remote sensing data from the earth terrain, the random medium model was shown to be quite successful. A multivariate K-distribution is proposed to model the statistics of fully polarimetric radar returns from earth terrain. In the terrain cover classification using the synthetic aperture radar (SAR) images, the applications of the K-distribution model will provide better performance than the conventional Gaussian classifiers. The layered random medium model is used to study the polarimetric response of sea ice. Supervised and unsupervised classification procedures are also developed and applied to synthetic aperture radar polarimetric images in order to identify their various earth terrain components for more than two classes. These classification procedures were applied to San Francisco Bay and Traverse City SAR images.

Kong, Jin AU↗
NASA NTRS1989

Modeling and control of flexible space platforms with articulated payloads

The first steps in developing a methodology for spacecraft control-structure interaction (CSI) optimization are identification and classification of anticipated missions, and the development of tractable mathematical models in each mission class. A mathematical model of a generic large flexible space platform (LFSP) with multiple independently pointed rigid payloads is considered. The objective is not to develop a general purpose numerical simulation, but rather to develop an analytically tractable mathematical model of such composite systems. The equations of motion for a single payload case are derived, and are linearized about zero steady-state. The resulting model is then extended to include multiple rigid payloads, yielding the desired analytical form. The mathematical models developed clearly show the internal inertial/elastic couplings, and are therefore suitable for analytical and numerical studies. A simple decentralized control law is proposed for fine pointing the payloads and LFSP attitude control, and simulation results are presented for an example problem. The decentralized controller is shown to be adequate for the example problem chosen, but does not, in general, guarantee stability. A centralized dissipative controller is then proposed, requiring a symmetric form of the composite system equations. Such a controller guarantees robust closed loop stability despite unmodeled elastic dynamics and parameter uncertainties.

Graves, Philip C.↗
NASA NTRS2004

Fast Query-Optimized Kernel-Machine Classification

A recently developed algorithm performs kernel-machine classification via incremental approximate nearest support vectors. The algorithm implements support-vector machines (SVMs) at speeds 10 to 100 times those attainable by use of conventional SVM algorithms. The algorithm offers potential benefits for classification of images, recognition of speech, recognition of handwriting, and diverse other applications in which there are requirements to discern patterns in large sets of data. SVMs constitute a subset of kernel machines (KMs), which have become popular as models for machine learning and, more specifically, for automated classification of input data on the basis of labeled training data. While similar in many ways to k-nearest-neighbors (k-NN) models and artificial neural networks (ANNs), SVMs tend to be more accurate. Using representations that scale only linearly in the numbers of training examples, while exploring nonlinear (kernelized) feature spaces that are exponentially larger than the original input dimensionality, KMs elegantly and practically overcome the classic curse of dimensionality. However, the price that one must pay for the power of KMs is that query-time complexity scales linearly with the number of training examples, making KMs often orders of magnitude more computationally expensive than are ANNs, decision trees, and other popular machine learning alternatives. The present algorithm treats an SVM classifier as a special form of a k-NN. The algorithm is based partly on an empirical observation that one can often achieve the same classification as that of an exact KM by using only small fraction of the nearest support vectors (SVs) of a query. The exact KM output is a weighted sum over the kernel values between the query and the SVs. In this algorithm, the KM output is approximated with a k-NN classifier, the output of which is a weighted sum only over the kernel values involving k selected SVs. Before query time, there are gathered statistics about how misleading the output of the k-NN model can be, relative to the outputs of the exact KM for a representative set of examples, for each possible k from 1 to the total number of SVs. From these statistics, there are derived upper and lower thresholds for each step k. These thresholds identify output levels for which the particular variant of the k-NN model already leans so strongly positively or negatively that a reversal in sign is unlikely, given the weaker SV neighbors still remaining. At query time, the partial output of each query is incrementally updated, stopping as soon as it exceeds the predetermined statistical thresholds of the current step. For an easy query, stopping can occur as early as step k = 1. For more difficult queries, stopping might not occur until nearly all SVs are touched. A key empirical observation is that this approach can tolerate very approximate nearest-neighbor orderings. In experiments, SVs and queries were projected to a subspace comprising the top few principal- component dimensions and neighbor orderings were computed in that subspace. This approach ensured that the overhead of the nearest-neighbor computations was insignificant, relative to that of the exact KM computation.

Mazzoni, Dominic↗
DOE OSTI2024

Chemical signature characterization with hyperspectral imagery: novel deep learning model architectures and physically-motivated data augmentation techniques

The high spectral resolution afforded by Hyperspectral Imaging (HSI) sensors is poised to bring unprecedented advancements to signature characterization applications. Thus far, much of the research in the machine learning field devoted to HSI applications has focused on a few specific tasks like land-use land-cover classification. In land classification tasks, spatial information is very important, and model architectures are often designed to leverage spatial contexts. However, it is unclear how well these spatially-tuned models will translate to tasks where spectral information is critical, like the detection and characterization of chemicals. In this work, we compare spectral models (inputs are 1D spectra) and spatial-spectral models (inputs are 3D cubes) in the context of predicting chemical concentration maps. We find that spatial-spectral models perform the best, though we find a wide range in performance across the different architectures tested. Additionally, we find that model performance is impacted by the availability of training data, particularly in scenarios where the training data doesn't fully capture the true variance of real-world conditions. We find that data augmentation can help mitigate sparse coverage of observed parameter space (e.g., seasonal or geographic variability in ground cover), and present augmentation strategies that are tailored to hyperspectral data.

• Artificial intelligence (AI) / machine learning ↗
DOE OSTI2022

Disentangling Alzheimer’s disease neurodegeneration from typical brain ageing using machine learning

Abstract Neuroimaging biomarkers that distinguish between changes due to typical brain ageing and Alzheimer’s disease are valuable for determining how much each contributes to cognitive decline. Supervised machine learning models can derive multivariate patterns of brain change related to the two processes, including the Spatial Patterns of Atrophy for Recognition of Alzheimer’s Disease (SPARE-AD) and of Brain Aging (SPARE-BA) scores investigated herein. However, the substantial overlap between brain regions affected in the two processes confounds measuring them independently. We present a methodology, and associated results, towards disentangling the two. T1-weighted MRI scans of 4054 participants (48–95 years) with Alzheimer’s disease, mild cognitive impairment (MCI), or cognitively normal (CN) diagnoses from the Imaging-based coordinate SysTem for AGIng and NeurodeGenerative diseases (iSTAGING) consortium were analysed. Multiple sets of SPARE scores were investigated, in order to probe imaging signatures of certain clinically or molecularly defined sub-cohorts. First, a subset of clinical Alzheimer’s disease patients (n = 718) and age- and sex-matched CN adults (n = 718) were selected based purely on clinical diagnoses to train SPARE-BA1 (regression of age using CN individuals) and SPARE-AD1 (classification of CN versus Alzheimer’s disease) models. Second, analogous groups were selected based on clinical and molecular markers to train SPARE-BA2 and SPARE-AD2 models: amyloid-positive Alzheimer’s disease continuum group (n = 718; consisting of amyloid-positive Alzheimer’s disease, amyloid-positive MCI, amyloid- and tau-positive CN individuals) and amyloid-negative CN group (n = 718). Finally, the combined group of the Alzheimer’s disease continuum and amyloid-negative CN individuals was used to train SPARE-BA3 model, with the intention to estimate brain age regardless of Alzheimer’s disease-related brain changes. The disentangled SPARE models, SPARE-AD2 and SPARE-BA3, derived brain patterns that were more specific to the two types of brain changes. The correlation between the SPARE-BA Gap (SPARE-BA minus chronological age) and SPARE-AD was significantly reduced after the decoupling (r = 0.56–0.06). The correlation of disentangled SPARE-AD was non-inferior to amyloid- and tau-related measurements and to the number of APOE ε4 alleles but was lower to Alzheimer’s disease-related psychometric test scores, suggesting the contribution of advanced brain ageing to the latter. The disentangled SPARE-BA was consistently less correlated with Alzheimer’s disease-related clinical, molecular and genetic variables. By employing conservative molecular diagnoses and introducing Alzheimer’s disease continuum cases to the SPARE-BA model training, we achieved more dissociable neuroanatomical biomarkers of typical brain ageing and Alzheimer’s disease.

Hwang, Gyujoon↗
DOE OSTI2022

Predicting defects in laser powder bed fusion using in-situ thermal imaging data and machine learning

Variation in the local thermal history during the Laser Powder Bed Fusion (LPBF) process in Additive Manufacturing (AM) can cause micropore defects, which add to the uncertainty of the mechanical properties (e.g., fatigue life, tensile strength) of the built materials. In-situ sensing has been proposed for monitoring the AM process to minimize defects, but successful minimization requires establishing a quantitative relationship between the sensing data and the porosity, which is particularly challenging with a large number of variables (e.g., laser speed, power, scan path, powder property). Physics-based modeling can simulate such an in-situ sensing-porosity relationship, but it is computationally costly. In this work, we develop Machine Learning (ML) models that can use in-situ thermographic data to predict the micropore of LPBF stainless steel materials. This work considers two identified key features from the thermal histories: the time above the apparent melting threshold ($\tau$) and the maximum radiance ( T max ). These features are computed, stored for each voxel in the built material, and then used as inputs. The binary state of each voxel, either defective or normal, is the output. Different ML models are trained and tested for the binary classification task. In addition to using the thermal features of each voxel to predict its own state, the thermal features of neighboring voxels are also included as inputs. This is shown to improve the prediction accuracy, which is consistent with thermal transport physics around each voxel contributing to its final state. Among the models trained, the F1 scores on test sets reach above 0.96 for Random Forests. Feature importance analysis based on the ML models shows that T max is more important to the voxel state than $\tau$ . The analysis also finds that the thermal history of the voxels above the present voxel is more influential than those beneath it. Our study significantly extends the capability of using in-situ thermographic data to predict porosity in LPBF materials. Finally, since ML models are fast, they may play integral roles in the optimization and control of such AM technologies.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Monitoring Operational States of a Nuclear Reactor Using Seismoacoustic Signatures and Machine Learning

Monitoring nuclear reactors is an important safety and security task with growing requirements. We explore the possibility of using seismic and acoustic data for inferring the power level of an operating reactor. Continuous data recorded at a single seismoacoustic station that is located about 50 m away from a research reactor was visualized and analyzed. The data show a clear correlation between seismoacoustic features and reactor main operational states. We designed a workflow that includes two machine learning (ML) models to classify the reactor operational states (OFF, transition, and ON) and estimate reactor power levels (10%, 30%, 50%, 70%, and 90%). We applied and compared five ML algorithms for the reactor OFF-transition-ON and four approaches for the power level classification. We also compared the performance of ML models trained with seismic-only, acoustic-only, and both types of data. Five-fold cross validations were implemented to assure a thorough evaluation of the model performances. Additionally, the results show the extreme boosting gradient algorithm worked best for the first model, whereas random forests performed best for the second model. Combining seismic and acoustic data leads to better performance than using a single type of data. Seismic data contributed more than acoustic data for both models. We reached an accuracy of 0.98 for reactor OFF and ON. The accuracies for the transition state and power levels are less optimal with a minimum accuracy of 0.66. However, our results suggest seismic and acoustic data contain useful information about the transition state as well as power levels. Seismic and acoustic data could be integrated with other observations to improve monitoring performance.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2025

Regime Characterization of Offshore Wind Resource Using Unsupervised Learning

Predictability of wind resource conditions is critical for offshore wind design and operations. While many studies of extreme wind conditions focus on specific events such as low-level jets or ramps, these rely on threshold definitions that limit generality. Here we present a data-driven framework that combines principal component analysis (PCA), self-organizing maps (SOM), and k-means clustering to classify wind resource conditions as typical and anomalous from climatological data. Anomalies are defined not by fixed thresholds but by flagging samples located far from SOM node centers inside the baseline SOM structure. This reframes extremes as rare ebents and hence, likely difficult to anticipate by numerical weather prediction models. We applied this approach to 23 years (2000–2022) of hourly profiles from the NOW-23 hindcast model at the Humboldt Wind Energy Area. Classification is conducted on a feature space consisting of 10 m wind speed and direction, bulk shear and veer across 30–270 m, and a low-level jet index. Dimensionality reduction is achieved through PC. A 2 × 3 OM lattice trained on the PCA vectors identified six baseline regimes spanning weak to strong flow states. High quantization-error profiles are identified and re-clustered into four anomalous regimes. The baseline regimes exhibited clear seasonal and diurnal cycles. Meanwhile, the anomalous regimes represented <10 % of all hours but showed distinct combinations of speed, shear, and veer, when compared to the baseline regimes. Anomalous regimes are typically short-lived (~few hours), yet their transitions can lead to hub-height wind changes of −18 to +9 m s -1 . For a representative 15 MW turbine, these shifts imply rapid swings in capacity factor from near-full output to negligible generation. Validation with lidar buoy data showed 51% agreement in SOM labels across ~6,000 overlapping hours, with most mismatches confined to adjacent speed classes. HRRR comparisons further revealed that anomalous regimes were disproportionately associated with forecast biases exceeding 5 m s -1 . Together, these results reframe extremes in offshore wind from absolute maxima or minima to weather states that are difficult to anticipate from models.

17 WIND ENERGY↗
NASA NTRS2009

Quantifying Parameter Sensitivity, Interaction and Transferability in Hydrologically Enhanced Versions of Noah-LSM over Transition Zones

We use sensitivity analysis to identify the parameters that are most responsible for shaping land surface model (LSM) simulations and to understand the complex interactions in three versions of the Noah LSM: the standard version (STD), a version enhanced with a simple groundwater module (GW), and version augmented by a dynamic phenology module (DV). We use warm season, high-frequency, near-surface states and turbulent fluxes collected over nine sites in the US Southern Great Plains. We quantify changes in the pattern of sensitive parameters, the amount and nature of the interaction between parameters, and the covariance structure of the distribution of behavioral parameter sets. Using Sobol s total and first-order sensitivity indexes, we show that very few parameters directly control the variance of the model output. Significant parameter interaction occurs so that not only the optimal parameter values differ between models, but the relationships between parameters change. GW decreases parameter interaction and appears to improve model realism, especially at wetter sites. DV increases parameter interaction and decreases identifiability, implying it is overparameterized and/or underconstrained. A case study at a wet site shows GW has two functional modes: one that mimics STD and a second in which GW improves model function by decoupling direct evaporation and baseflow. Unsupervised classification of the posterior distributions of behavioral parameter sets cannot group similar sites based solely on soil or vegetation type, helping to explain why transferability between sites and models is not straightforward. This evidence suggests a priori assignment of parameters should also consider climatic differences.

Rosero, Enrique↗
NASA NTRS1981

Multiresource analysis and information system concepts for incorporating LANDSAT and GIS technology into large area forest surveys

A method of relating different classifications at each stage of a multistage, multiresource inventory using remotely sensed imagery is discussed. A class transformation matrix allowing the conversion of a set of proportions at one stage, to a set of proportions at the subsequent stage through use of a linear model, is described. The technique was tested by applying it to Kershaw County, South Carolina. Unsupervised LANDSAT spectral classifications were correlated with interpretations of land use aerial photography, the correlations employed to estimate land use classifications using the linear model, and the land use proportions used to stratify current annual increment (CAI) field plot data to obtain a total CAI for the county. The estimate differed by 1% from the published figure for land use. Potential sediment loss and a variety of land use classifications were also obtained.

Langley, P. G.↗
DOE OSTI2021

Monitoring leaf phenology in moist tropical forests by applying a superpixel-based deep learning method to time-series images of tree canopies

Tropical leaf phenology-particularly its variability at the tree-crown scale-dominates the seasonality of carbon and water fluxes. However, given enormous species diversity, accurate means of monitoring leaf phenology in tropical forests is still lacking. Time series of the Green Chromatic Coordinate (GCC) metric derived from tower-based red-green-blue (RGB) phenocams have been widely used to monitor leaf phenology in temperate forests, but its application in the tropics remains problematic. To improve monitoring of tropical phenology, we explored the use of a deep learning model (i.e. superpixel-based Residual Networks 50, SP-ResNet50) to automatically differentiate leaves from non-leaves in phenocam images and to derive leaf fraction at the tree-crown scale. To evaluate our model, we used a year of data from six phenocams in two contrasting forests in Panama. Here, we first built a comprehensive library of leaf and non-leaf pixels across various acquisition times, exposure conditions and specific phenocams. We then divided this library into training and testing components. We evaluated the model at three levels: 1) superpixel level with a testing set, 2) crown level by comparing the model-derived leaf fractions with those derived using image-specific supervised classification, and 3) temporally using all daily images to assess the diurnal stability of the model-derived leaf fraction. Finally, we compared the model-derived leaf fraction phenology with leaf phenology derived from GCC. Our results show that: 1) the SP-ResNet50 model accurately differentiates leaves from non-leaves (overall accuracy of 93%) and is robust across all three levels of evaluations; 2) the model accurately quantifies leaf fraction phenology across tree-crowns and forest ecosystems; and 3) the combined use of leaf fraction and GCC helps infer the timing of leaf emergence, maturation and senescence, critical information for modeling photosynthetic seasonality of tropical forests. Collectively, this study offers an improved means for automated tropical phenology monitoring using phenocams.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Multisource Mobile Transfer Learning Algorithm Based on Dynamic Model Compression

With the development of the Internet of Things, the application of computer vision on mobile phones is becoming more and more extensive and people have higher and higher requirements for the timeliness of the recognition results returned and the processing capabilities of the mobile phone for image recognition. However, the processing capability and storage capability of the user terminal equipment cannot meet the needs of identifying and storing a large number of pictures, and the data transmission process will cause high energy consumption of the terminal equipment. At the same time, multisource deep transfer learning has outstanding performance in computer vision and image classification. However, due to the huge amount of calculation of the deep network model, it is impossible to use the existing excellent network model to realize image recognition and classification on the mobile terminal. In order to solve the abovementioned problems, we propose a multisource mobile transfer learning algorithm based on dynamic model compression, this algorithm considers the realization of multisource transfer learning computing in the case of multiple mobile device computing source domains, and the method also guarantees data privacy and security for each device (origin domain). Meanwhile, extensive experiments show that our method can achieve remarkable results in popular image classification datasets.

Gao, Peng↗
DOE OSTI2024

Modern deep neural networks for Direct Normal Irradiance forecasting: A classification approach

The escalating energy demand and the adverse environmental impacts of fossil-fuel use necessitate a shift towards cleaner and renewable alternatives. Concentrated Solar Power (CSP) technology emerges as a promising solution, offering a carbon-free alternative for power generation. The efficiency and profitability of CSP depend on the Direct Normal Irradiance (DNI) component of solar radiation; hence, accurate DNI forecasting can help optimize CSP plants’ operations and performance. The unpredictable nature of weather phenomena, particularly cloud cover, introduces uncertainty into DNI projections. Existing DNI forecasting models use meteorological factors, which are both challenging to estimate numerically over short prediction windows and expensive to model through data at a sufficiently high spatial and temporal resolution. This research addresses the challenge by presenting a novel approach that formulates DNI prediction as a multi-class classification problem, departing from conventional regression-based methods. The primary objective of this classification framework is to identify optimal periods aligning with specific operational thresholds for CSP plants, contributing to enhanced dispatch optimization strategies. We model the DNI classification problem using four advanced deep neural networks – rectified linear unit (ReLU) networks, 1D residual networks (ResNets), bidirectional long short-term memory (BiLSTM) networks, and transformers – achieving accuracies up to 93.5% without requiring meteorological parameters.

14 SOLAR ENERGY↗
DOE OSTI2022

The dark energy survey 5-yr photometrically identified type Ia supernovae

ABSTRACT As part of the cosmology analysis using Type Ia Supernovae (SN Ia) in the Dark Energy Survey (DES), we present photometrically identified SN Ia samples using multiband light curves and host galaxy redshifts. For this analysis, we use the photometric classification framework SuperNNovatrained on realistic DES-like simulations. For reliable classification, we process the DES SN programme (DES-SN) data and introduce improvements to the classifier architecture, obtaining classification accuracies of more than 98 per cent on simulations. This is the first SN classification to make use of ensemble methods, resulting in more robust samples. Using photometry, host galaxy redshifts, and a classification probability requirement, we identify 1863 SNe Ia from which we select 1484 cosmology-grade SNe Ia spanning the redshift range of 0.07 < z < 1.14. We find good agreement between the light-curve properties of the photometrically selected sample and simulations. Additionally, we create similar SN Ia samples using two types of Bayesian Neural Network classifiers that provide uncertainties on the classification probabilities. We test the feasibility of using these uncertainties as indicators for out-of-distribution candidates and model confidence. Finally, we discuss the implications of photometric samples and classification methods for future surveys such as Vera C. Rubin Observatory Legacy Survey of Space and Time.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
NASA NTRS2007

Application of High Resolution Multispectral Imagery for Levee Slide Detection and Monitoring

The objective is to develop methods to detect and monitor levee slides using commercially available high resolution multispectral imagery. High resolution multispectral imagery like IKONOS and QuickBird are suitable for detecting and monitoring levee slides. IKONOS is suitable for visual inspection, image classification and Tasseled Cap transform based slide detection. Tasseled Cap based model was found to be the best method for slide detection. QuickBird was suitable for visual inspection and image classification.

Hossain, A. K. M. Azad↗
DOE OSTI2025

DriveSense: A Noise-Resilient Framework for Driving Mode Identification

Accurate drive mode classification is essential for enhancing the reliability and predictive maintenance of heavy-duty electric trucks. This study proposes a novel fuzzy logic-based framework, DriveSense, for real-time drive mode classification, addressing key challenges such as sensor noise, transitional behaviors, and computational efficiency. The proposed approach integrates a two-stage filtering pipeline, combining adaptive outlier removal and a dynamic Kalman filter to enhance data quality. A fuzzy inference system with smoothened trapezoidal membership functions is then applied to classify driving modes into standstill, constant speed, acceleration, and deceleration while mitigating the effects of noise and edge cases. Performance evaluation using real-world and simulated drive cycles demonstrates significant improvements in classification accuracy (up to 97.8%), F1-score (up to 0.97), and robustness against noise, while reducing false positives. Comparative analysis against baseline models, demonstrates DriveSense’s superior accuracy and generalizability across diverse driving patterns. The framework’s lightweight and interpretable fuzzy inference engine operates with low computational latency, ensuring compatibility with real-time embedded systems typical of heavy-duty electric trucks. Moreover, DriveSense models transitional behaviors through overlapping fuzzy sets and adaptive borderline classification logic, enabling smooth identification of subtle shifts such as rolling stops or gradual deceleration. These results highlight DriveSense’s potential to enhance predictive maintenance strategies, reduce downtime, and support scalable, fleet-wide diagnostics.

Kumar, Praveen [Oak Ridge National Laboratory (ORN↗
DOE OSTI2025

Domain Shift Analysis in Chest Radiographs Classification in a Veterans Healthcare Administration Population

This study aims to assess the impact of domain shift on chest X-ray classification accuracy and to analyze the influence of ground truth label quality and demographic factors such as age group, sex, and study year. We used a DenseNet121 model pre-trained MIMIC-CXR dataset for deep learning-based multi-label classification using ground truth labels from radiology reports extracted using the CheXpert and CheXbert Labeler. We compared the performance of the 14 chest X-ray labels on the MIMIC-CXR and Veterans Healthcare Administration chest X-ray dataset (VA-CXR). The validation of ground truth and the assessment of multi-label classification performance across various NLP extraction tools revealed that the VA-CXR dataset exhibited lower disagreement rates than the MIMIC-CXR datasets. Additionally, there were notable differences in AUC scores between models utilizing CheXpert and CheXbert. When evaluating multi-label classification performance across different datasets, minimal domain shift was observed in the unseen VA dataset, except for the label “Enlarged Cardiomediastinum.” The subgroup with the most significant variations in multi-label classification performance was study year. These findings underscore the importance of considering domain shift in chest X-ray classification tasks, paying particular attention to the temporality of the exam. Our study reveals the significant impact of domain shift and demographic factors on chest X-ray classification, emphasizing the need for improved transfer learning and robust model development. Addressing these challenges is crucial for advancing medical imaging research and improving patient care.

chest X-ray image classification↗
DOE OSTI2026

A curated benchmark for cofolding models on kinase conformational states

Abstract Protein kinases are critical drug targets, requiring therapeutics that can modulate their active and inactive conformational states. While cofolding models can generate global folds directly from kinase sequences and ligand SMILES strings, these models have not yet been tested on their ability to recover ligand-induced-fit conformational states of the kinase proteins. Here, we introduce KinConfBench, a curated benchmark of 2225 high-quality human kinase chains to evaluate the ability of four state-of-the-art cofolding models—Boltz-2, Chai-1, Protenix, and RoseTTAFold-All-Atom—to recover both canonical and rare conformational states. We show that geometric success metrics of a ligand pose in the active site do not correlate strongly with the correct kinase conformational state, motivating a new set of dynamical benchmarks for assessing cofolding models. While all four cofolding models achieve ~60–80% prediction accuracy for kinase conformational classification, they exhibit severe mode collapse when performing multiple inferences, show negligible structural diversity in sampling induced-fit motions, and display a prevalent “apo-drift” in which most cofolding models predominantly predict the kinase to be in its ligand-free state. Our results highlight that capturing ligand-induced protein conformational diversity, not just geometric fit, is critical for next-generation structure-based drug discovery.

Sun, Kunyang↗