Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Scientific method”

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 577 records · Page 32
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Deceptive Infusion of Data: A Novel Data Masking Paradigm for High-Valued Systems

This work addresses how analysts of a high-valued system (e.g., nuclear reactor, aircraft turbine designs) can extract findable, accessible, interoperable, and reusable scientific data for public dissemination to artificial intelligence and machine-learning (AI/ML) researchers in a manner that cannot be reverse-engineered, potentially compromising sensitive or proprietary information. State-of-the-art methods address this problem through data masking techniques, which allow access to a subset of the information while obfuscating private and potentially identifying information (e.g., personally identifying medical data). These methods are unsuitable for industrial engineering processes, where AI/ML tools need explicit access to all the data available to draw the best inference about the system to help optimize its performance and identify its vulnerabilities, etc. Our novel deceptive infusion of data paradigm provides a solution to this conundrum by developing a mathematical approach capable of concealing the identity of the system while providing full access to all the features employed by AI/ML tools to ensure their optimal performance.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2024

A Quantum Approach for Implementing Fixed-Point Arithmetic in Solving Ordinary Differential Equations

Differential equations (DEs) serve as fundamental tools in mathematical modeling across scientific disciplines, yet classical numerical solvers face limitations with large-scale or computationally intensive problems. This study explores a quantum-inspired approach to solving DEs, combining quantum-inspired techniques with classical methods. It focuses on fixed-point arithmetic on quantum circuits, utilizing basic quantum gates to manipulate DE solutions. We expand upon the techniques introduced by Zanger et al. [Quantum, 5, 502 (2021)] by offering a precise computation for a fixed-point signed multiplication scheme, while also presenting a quantum circuit capable of executing the fixed-point division algorithm. We demonstrate the feasibility of our approach through the simulation of a linear Ordinary Differential Equation (ODE), where initial conditions and parameters are encoded into quantum circuits using fixed-point representation. By executing sequences of quantum gates mimicking numerical integration steps, we obtain approximate solutions to the ODE with specified fixed-point precision.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2023

Dataset for 'Ombadi et al. (2023). A warming-induced reduction in snow fraction amplifies rainfall extremes, Nature'

This package contains the main codes, sample input data and main result files to reproduce the analysis and results presented in the article: “Ombadi et al. (2023), A warming-induced reduction in snow fraction amplifies rainfall extremes, Nature”. The folder consists of the following: (1) “Raw data”: a folder that contains sample input data which is used in some of the codes for demonstration purposes. It also contains data that was not pre-processed such as Elevation data; (2) “Results”: this folder contains files of the main results presented in the paper including: “Annual-Max-Series”, “Change-rainfall-extremes”, “Change-snow-fraction”, “Warming levels_By scenario_model_year” and “Masks”. Description of these folders is detailed in the "Readme.rtf" file; (3) Python jupyter notebooks (Extract_Annual Max Series (AMS).ipynb, Elevation-dependent amplification of rainfall extremes.ipynb, Sensitivity_to_global_warming.ipynb) demonstrate the main steps of analysis. Further description of those notebooks is provided in the "Readme.rtf" file; (4) R code for extreme value analysis (Extreme_Value_Analysis.R). The sample and pre-processed dataset in "Raw data" is obtained from publicly available repositories of CMIP6 and ERA5 datasets; see Methods for more detail. This research was supported by Office of Science, Office of Biological and Environmental Research of the US Department of Energy under contract no. DE-AC02-05CH11231 for the CASCADE Scientific Focus (funded by the Regional and Global Model Analysis Program area within the Earth and Environmental Systems Modeling Program) and the iNAIADS Early Career Research Project (funded by the Environmental Systems Science program).

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

AmeriFlux FLUXNET-1F US-xJR NEON Jornada LTER (JORN)

This is the AmeriFlux Management Project (AMP) created FLUXNET-1F version of the carbon flux data for the site US-xJR NEON Jornada LTER (JORN). This is the FLUXNET version of the carbon flux data for the site US-xJR NEON Jornada LTER (JORN) produced by applying the standard ONEFlux (1F) software. Site Description - Land use at Jornada has contrasting livestock grazing histories (including ungrazed lands), dating back to 1902, which are representative of the pervasive land use in desert-southwest wildlands. Historical accounts of the region report significant changes in vegetation starting in the late 1800s coincident with the expansion of livestock grazing. By 1912, sufficiently dramatic changes lead area scientists and private landowners to convince the U.S. government to set aside Public Domain Lands for the Jornada Range Reserve for the purpose of scientific investigation on shrub invasion and subsequent loss of forage grasses. Much of this early research focused on quantifying utilization levels for forage species, developing livestock production strategies to deal with drought, and developing methods for shrub control and grass recovery. Exclosures were constructed and long-term plots were established throughout the range to monitor the continued expansion of shrubs across the landscape. Over the decades, numerous trials of various remediation approaches were implemented, from manual and mechanical shrub removal to herbicide application to construction of terraces or other means of redirecting surface flow of runoff.

Network), NEON (National Ecological Observatory↗
DOE OSTI2025

A kinetic-based regularization method for data science applications

We propose a physics-based regularization technique for function learning, inspired by statistical mechanics. By drawing an analogy between optimizing the parameters of an interpolator and minimizing the energy of a system, we introduce corrections that impose constraints on the lower-order moments of the data distribution. This minimizes the discrepancy between the discrete and continuum representations of the data, in turn allowing to access more favorable energy landscapes, thus improving the accuracy of the interpolator. Our approach improves performance in both interpolation and regression tasks, even in high-dimensional spaces. Unlike traditional methods, it does not require empirical parameter tuning, making it particularly effective for handling noisy data. We also show that thanks to its local nature, the method offers computational and memory efficiency advantages over Radial Basis Function interpolators, especially for large datasets.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Physics-Guided, Physics-Informed, and Physics-Encoded Neural Networks and Operators in Scientific Computing: Fluid and Solid Mechanics

Abstract Advancements in computing power have recently made it possible to utilize machine learning and deep learning to push scientific computing forward in a range of disciplines, such as fluid mechanics, solid mechanics, materials science, etc. The incorporation of neural networks is particularly crucial in this hybridization process. Due to their intrinsic architecture, conventional neural networks cannot be successfully trained and scoped when data are sparse, which is the case in many scientific and engineering domains. Nonetheless, neural networks provide a solid foundation to respect physics-driven or knowledge-based constraints during training. Generally speaking, there are three distinct neural network frameworks to enforce the underlying physics: (i) physics-guided neural networks (PgNNs), (ii) physics-informed neural networks (PiNNs), and (iii) physics-encoded neural networks (PeNNs). These methods provide distinct advantages for accelerating the numerical modeling of complex multiscale multiphysics phenomena. In addition, the recent developments in neural operators (NOs) add another dimension to these new simulation paradigms, especially when the real-time prediction of complex multiphysics systems is required. All these models also come with their own unique drawbacks and limitations that call for further fundamental research. This study aims to present a review of the four neural network frameworks (i.e., PgNNs, PiNNs, PeNNs, and NOs) used in scientific computing research. The state-of-the-art architectures and their applications are reviewed, limitations are discussed, and future research opportunities are presented in terms of improving algorithms, considering causalities, expanding applications, and coupling scientific and deep learning solvers.

Computer Science↗
DOE OSTI2025

Counterpart identification and classification for eRASS1 and characterisation of the active galactic nuclei content

Context. Accurately accounting for the Active Galactic Nucleus (AGN) phase in galaxy evolution requires a large, clean AGN sample. This is now possible with SRG/eROSITA, which completed its first all-sky X-ray survey (eRASS1) on June 12, 2020. The public Data Release 1 (DR1, Jan 31, 2024) includes 930,203 sources from the western Galactic hemisphere. Aims. The data enable the selection of a large AGN sample and the discovery of rare sources. However, scientific return depends on accurate characterisation of the X-ray emitters, requiring high-quality multi-wavelength data. This paper presents the identification and classification of optical and infrared counterparts to eRASS1 sources. Methods. Counterparts to eRASS1 X-ray point sources were identified using Gaia DR3, CatWISE2020, and Legacy Survey DR10 (LS10) with the Bayesian NWAY algorithm and trained priors. Sources were classified as Galactic or extragalactic via a machine-learning model combining optical/IR and X-ray properties, trained on a reference sample. For extragalactic LS10 sources, photometric redshifts were computed using CIRCLEZ. Results. Within the LS10 footprint, all 656,614 eROSITA/DR1 sources have at least one possible optical counterpart; ∼570 000 are extragalactic and likely AGN. Half are new detections compared to AllWISE, Gaia, and Quaia AGN catalogues. Gaia and CatWISE2020 counterparts are less reliable, due to the survey’s shallowness and the limited amount of features available to assess the probability of being an X-ray emitter. In the Galactic plane, where the overdensity of stellar sources also increases the chance of associations, using conservative reliability cuts, we identified approximately 18 000 Gaia and 55 000 CatWISE2020 extragalactic sources. Conclusions. We have released three high-quality counterpart catalogues – plus the training and validation sets – as a benchmark for the field. These datasets have many applications, but in particular, they empower researchers to build AGN samples tailored for completeness and purity, accelerating the hunt for the Universe’s most energetic engines.

X-rays: general↗
DOE OSTI2022

MULTISTEP AND CONTINUOUS PHYSICS-INFORMED NEURAL NETWORK METHODS FOR LEARNING GOVERNING EQUATIONS AND CONSTITUTIVE RELATIONS

In this work, we investigate the applicability and relative merit of discrete and continuous versions of physics-informed neural network (PINN) methods for learning unknown governing equations or constitutive relations in a nonlinear dynamical system. In the case of unknown dynamics, entire right-hand-side (RHS) equations of the ordinary differential equations are unknown. In the case of unknown constitutive relations, however, the RHS equations are known up to the specification of constitutive relations (that may depend on the state of the system). We use a deep neural network to model unknown governing equations or constitutive relations. The discrete PINN approach combines classical multistep discretization methods for dynamical systems with neural-network-based machine learning methods. On the other hand, the continuous versions utilize deep neural networks to minimize the residual function for the continuous governing equations. We use the case of a fedbatch bioreactor system to study the effectiveness of these approaches and discuss conditions for their applicability. Our results indicate that the accuracy of the trained neural network models is much higher for the cases where we only have to learn a constitutive relation instead of all dynamics. This finding corroborates the well-known fact from scientific computing that building as much structural information as is available into an algorithm can enhance its efficiency and/or accuracy.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

A Comprehensive Review of Latent Space Dynamics Identification Algorithms for Intrusive and Non-Intrusive Reduced-Order-Modeling

Numerical solvers of partial differential equations (PDEs) have been widely employed for simulating physical systems. However, the computational cost remains a major bottleneck in various scientific and engineering applications, which has motivated the development of reduced-order models (ROMs). Recently, machine-learning-based ROMs have gained significant popularity and are promising for addressing some limitations of traditional ROM methods, especially for advection dominated systems. In this chapter, we focus on a particular framework known as Latent Space Dynamics Identification (LaSDI), which transforms the high-fidelity data, governed by a PDE, to simpler and low-dimensional latent-space data, governed by ordinary differential equations (ODEs). These ODEs can be learned and subsequently interpolated to make ROM predictions. Each building block of LaSDI can be easily modulated depending on the application, which makes the LaSDI framework highly flexible. In particular, we present strategies to enforce the laws of thermodynamics into LaSDI models (tLaSDI), enhance robustness in the presence of noise through the weak form (WLaSDI), select high-fidelity training data efficiently through active learning (gLaSDI, GPLaSDI), and quantify the ROM prediction uncertainty through Gaussian processes (GPLaSDI). We demonstrate the performance of different LaSDI approaches on Burgers equation, a non-linear heat conduction problem, and a plasma physics problem, showing that LaSDI algorithms can achieve relative errors of less than a few percent and up to thousands of times speed-ups.

Computational Engineering, Finance, and Science (c↗
DOE OSTI2021

Evaluation of the Large-Scale and Regional Climatic Response Across North Africa to Natural Variability in Oceanic Modes and Terrestrial Vegetation Among the CMIP5 Models

Hydrologic variability is a serious threat to the poverty-stricken regions of North Africa. Meanwhile, the scientific community struggles to attribute these extreme climatic episodes to specific oceanic and land drivers. Prior modeling studies have not assessed simulated feedbacks over North Africa against an established observational benchmark, so their results are considered as untested, model-specific findings. The Coupled Model Intercomparison Project Phase Five (CMIP5) archive represents the state-of-the-art in climate projections, with most CMIP5 models now containing interactive vegetation phenology; however, there have been no studies to date of their simulated vegetation feedbacks. The capability of CMIP5 models at accurately simulating the forcing of both sea-surface temperature (SST) and regional leaf area index (LAI) anomalies on North African climate needs to be a key consideration in assessing the models’ overall credibility and determining appropriate model weighting for developing climate projections. The Generalized Equilibrium Feedback Assessment (GEFA) is a promising statistical method, which can be applied to either model output or observations, for isolating the local and remote impacts of individual oceanic or terrestrial forcings on regional climate. We performed a combined observational and modeling assessment of land-ocean-atmosphere interactions across the distinct ecological and moisture gradients of North Africa. We evaluated and demonstrated the reliability of the GEFA statistical method over North Africa using the Community Earth System Model (CESM), applied GEFA to observational data to quantify the observed forcing of ocean basin SST anomalies and regional LAI anomalies on North African climate, evaluated the CMIP5 models’ performance in terms of representing these key observed feedbacks, and formulated CMIP5 feedback performance metrics for weighting North African climate projections. Our study represented the first attempt to separate the observed roles of oceanic and vegetation feedbacks across North Africa, the first systematic assessment and intercomparison of land-ocean-atmosphere feedbacks in CMIP5, and the first exploration of vegetation feedbacks among CMIP5 models. The following work was accomplished. (1) The team generated 14 scientific publications and 30 presentations based on the DOE-funded research. (2) A stepwise version of the GEFA statistical method was developed in which unimportant forcings were dropped to increase the reliability of the results. (3) SGEFA was successfully validated through experiments with CESM in terms of its ability to isolate the atmospheric responses to individual oceanic or land forcings. (4) Observational evidence was revealed for the Sahel’s positive vegetation-rainfall feedback on the seasonal to interannual time scale, and it was attributed to a moisture recycling mechanism rather than an albedo mechanism. (5) An approach with SGEFA was developed in which the individual contributions of soil moisture versus vegetation forcings could be separated, revealing that the former forcing outweighed the latter for sub-Saharan Africa. (6) Tropical ocean temperatures were found to be key regulators of pan-tropical vegetation variability, especially for arid and semi-arid regions, including sub-Saharan Africa. (7) The CMIP5 models largely underestimated the importance of land feedbacks across the Sahel. (8) The general consensus among CMIP5 models indicates, for the late 21st century, a diminished seasonal predictability of sub-Saharan African regional climate and an elevated role of the land surface compared to oceanic drivers in regulating regional climate variability. (9) Land-ocean-atmosphere interactions were demonstrated to be key contributors to the seasonal predictability of African wildfire activity. (10) The El Djouf was determined to be of greater importance than the Bodélé depression in terms of providing trans-Atlantic dust transport to the Americas.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Delving into the depths of NGC 3783 with XRISM II. Cross-calibration of X-ray instruments used in the large, multi-mission observational campaign

Context. Accurate X-ray spectroscopic measurements are fundamental for deriving basic physical parameters of the most abundant baryon components in the Universe. The plethora of X-ray observatories currently operational enables a panchromatic view of the high-energy emission of celestial sources. However, uncertainties in the energy-dependent calibration of the instrument transfer functions (e.g. the effective area, energy redistribution, or gain) can limit - and historically, did limit - the accuracy of X-ray spectroscopic measurements. Aims. We revised the status of the cross-calibration among the scientific payload on board four operation missions: Chandra, NuSTAR, XMM-Newton, and the recently launched XRISM. XRISM carries the micro-calorimeter Resolve, which yields the best energy resolution at energies ≥2 keV. For this purpose, we used the data from a 10-day-long observational campaign targeting the nearby active galactic nucleus NGC 3783, carried out in July 2024. Methods. We present a novel model-independent method for assessing the cross-calibration status that is based on a multi-node spline of the spectra with the highest-resolving power (XRISM/Resolve in our campaign). We also estimated the impact of the intrinsic variability of NGC 3783 on the cross-calibration status due to the different time coverages of participating observatories and performed an empirical reassessment of the Resolve throughput at low energies. Results. Based on this analysis, we derived a set of energy-dependent correction factors of the observed responses, enabling a statistically robust analysis of the whole spectral dataset. They will be employed in subsequent papers describing the astrophysical results of the campaign. Aims. We revised the status of the cross-calibration among the scientific payload on board four operation missions: Chandra, NuSTAR, XMM-Newton, and the recently launched XRISM. XRISM carries the micro-calorimeter Resolve, which yields the best energy resolution at energies ≥2 keV. For this purpose, we used the data from a 10-day-long observational campaign targeting the nearby active galactic nucleus NGC 3783, carried out in July 2024. Methods. We present a novel model-independent method for assessing the cross-calibration status that is based on a multi-node spline of the spectra with the highest-resolving power (XRISM/Resolve in our campaign). We also estimated the impact of the intrinsic variability of NGC 3783 on the cross-calibration status due to the different time coverages of participating observatories and performed an empirical reassessment of the Resolve throughput at low energies. Results. Based on this analysis, we derived a set of energy-dependent correction factors of the observed responses, enabling a statistically robust analysis of the whole spectral dataset. They will be employed in subsequent papers describing the astrophysical results of the campaign.

Active Galactic Nuclei, individual: NGC 3783↗
DOE OSTI2020

Correction: A universal method for sensitive and cell-free detection of CRISPR-associated nucleases

In the original article, incorrect grant information from the Department of Energy was provided. The corrected Acknowledgementssection is provided below, with the correct grant number:This work was supported by the Burroughs Wellcome Fund (Career Award at the Scientific Interface to A. C.), DARPA (BrdiN66001-17-2-4055 to A. C.), NIH (1R21AI126239-01 to A. C.), Army Research Office award W911NF1610586 (to A. C.), and by theDepartment of Energy through grant DE-SC0010595 to E. F., which supported the salary of H. K. K. S. This work is dedicated toProfessor Ronald T. Raines on the occasion of his 60th birthday.The Royal Society of Chemistry apologises for these errors and any consequent inconvenience to authors and readers.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Astronomaly at scale: searching for anomalies amongst 4 million galaxies

ABSTRACT Modern astronomical surveys are producing data sets of unprecedented size and richness, increasing the potential for high-impact scientific discovery. This possibility, coupled with the challenge of exploring a large number of sources, has led to the development of novel machine-learning-based anomaly detection approaches, such as astronomaly. For the first time, we test the scalability of astronomaly by applying it to almost 4 million images of galaxies from the Dark Energy Camera Legacy Survey. We use a trained deep learning algorithm to learn useful representations of the images and pass these to the anomaly detection algorithm isolation forest, coupled with astronomaly’s active learning method, to discover interesting sources. We find that data selection criteria have a significant impact on the trade-off between finding rare sources such as strong lenses and introducing artefacts into the data set. We demonstrate that active learning is required to identify the most interesting sources and reduce artefacts, while anomaly detection methods alone are insufficient. Using astronomaly, we find 1635 anomalies among the top 2000 sources in the data set after applying active learning, including eight strong gravitational lens candidates, 1609 galaxy merger candidates, and 18 previously unidentified sources exhibiting highly unusual morphology. Our results show that by leveraging the human–machine interface, astronomaly is able to rapidly identify sources of scientific interest even in large data sets.

Astronomy & Astrophysics↗
DOE OSTI2025

Kβ X-ray Emission Spectra Analysis Using Bayesian Optimization

The Kβ X-ray emission spectrum of 3 d transition metals is rich with electronic and structural information due to strong exchange interactions with the valence shell of the metal, and has become crucial for understanding their spin and oxidation states. The spectrum is commonly treated using crystal-field multiplet theory, a semi-empirical theory that uses tunable parameters to control the strength of the effects present in X-ray emission spectroscopy (XES). However, determining the experimental values of these parameters remains a challenge. We present a methodology that applies Bayesian optimization to crystal-field multiplet theory to determine parameter values. The algorithm is tested on the X-ray emission spectra of a collection of Mn, Co, and Ni oxides. We are able to find optimal values for the four most impactful parameters: Slater−Condon reduction factors F dd , F pd , and G pd , and crystal field splitting 10 Dq . The algorithm produces significantly improved accuracy compared to current analysis methods, and probes interparameter dependencies by modeling the error landscape. This advancement enhances XES analysis by offering an approach of obtaining quantitative electronic structural information on 3 d transition metal valence shells, facilitating applications across various scientific fields.

Bayesian optimization↗
DOE OSTI2025

Water Mass Transformation Budgets in Finite‐Volume Generalized Vertical Coordinate Ocean Models

Water Mass Transformation (WMT) theory provides conceptual tools that in principle enable innovative analyses of numerical ocean models; in practice, however, these methods can be challenging to implement and interpret, and therefore remain under-utilized. Our aim is to demonstrate the feasibility of diagnosing all terms in the water mass budget and to exemplify their usefulness for scientific inquiry and model development by quantitatively relating water mass changes, overturning circulations, boundary fluxes, and interior mixing. We begin with a pedagogical derivation of key results of classical WMT theory. We then describe best practices for diagnosing each of the water mass budget terms from the output of Finite-Volume Generalized Vertical Coordinate (FV-GVC) ocean models, including the identification of a non-negligible remainder term as the spurious numerical mixing due to advection scheme discretization errors. We illustrate key aspects of the methodology through the analysis of a polygonal region of the Greater Baltic Sea in a regional demonstration simulation using the Modular Ocean Model v6 (MOM6). We verify the convergence of our WMT diagnostics by brute-force, comparing time-averaged (“offline”) diagnostics on various vertical grids to timestep-averaged (“online”) diagnostics on the native model grid. Finally, we briefly describe a stack of xarray-enabled Python packages for evaluating WMT budgets in FV-GVC models (culminating in the new xwmb package), which is intended to be model-agnostic and available for community use and development.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Analysis of Needlet Internal Linear Combination performance on B -mode data from sub-orbital experiments

The observation of primordial B modes in cosmic microwave background (CMB) polarisation data represents the main scientific goal of most of the future CMB experiments. This signal is predicted to be much lower than polarised Galactic emission (foregrounds) in any region of the sky, pointing to the need for effective component separation methods. Aims. Among all the techniques, the blind Needlet Internal Linear Combination (NILC) is of great relevance given our current limited knowledge of the B-mode foregrounds. In this work, we explore the possibility of employing NILC for the analysis of B modes reconstructed from partial-sky data, specifically addressing the complications that such an application yields such as E–B leakage, needlet filtering, and beam convolution. We consider two complementary simulated datasets of future experiments: the balloon-borne Short Wavelength Instrument for the Polarisation Explorer (SWIPE) of the Large Scale Polarisation Explorer, which targets the observation of both reionisation and recombination peaks of the primordial CMB B-mode angular power spectrum, and the ground-based Small Aperture Telescope of Simons Observatory, which, instead, is designed to observe only the recombination bump at ℓ ~ 80. We assessed the performance of the following two alternative techniques to correct for the CMB E–B leakage: the recycling technique and the Zhao-Baskaran method. We find that both techniques reduce the E–B leakage residuals at a negligible level given the sensitivity of the considered experiments, except for the recycling method in the SWIPE footprint at ℓ < 20. Thus, we implemented two extensions of the pipeline, the iterative B decomposition and the diffusive inpainting, which enabled us to recover the input CMB B-mode power for ℓ ≥ 5. For the considered experiments, we demonstrate that needlet filtering and beam convolution do not affect the CMB B-mode reconstruction. Finally, with an appropriate masking strategy, we find that NILC foregrounds subtraction allows one to achieve sensitivities on the tensor-to-scalar ratio in agreement with the targets of the considered CMB experiments.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Enhancing Electron Microscopy Image Classification Using Data Augmentation

Manual labeling for machine learning tasks such as image classification is tedious and labor-intensive; as a result, scientific datasets suitable for deep learning applications are scarce and limited. While data augmentation techniques have shown promise for extending image datasets, very little work has been done to understand the impact of combining multiple augmentation methods sequentially or the limits of their effectiveness when combined. Our work addresses this gap by examining how standard and combinatorial data augmentation affects the performance of machine learning models when trained on small datasets for label classification tasks. For our analysis, we generate single, double and quadruple-augmented datasets for a microscopy image classification task using six standard augmentation methods, and compare the resultant improvements observed in binary classification accuracy with three standard image classification models (DenseNet169, MobileNetV2, ResNet101V2). Our experiments show a non-monotonic relationship between the number of simultaneous augmentation methods and classification accuracy, indicating that there is a trade-off between the degree of augmentation and the model performance. These findings suggest that the optimal number of augmentation methods will vary by domain and use case. We also find that the order in which augmentation methods are applied to a limited dataset matters when combining augmentation schemes, with our use case showing performance differences up to 2.6% when the augmentation order is reversed for double-augmented datasets. Our work offers insights to the limits of data augmentation when working on image classification tasks with limited datasets.

Welsman, Jordan A↗