SEARCH · Search NASA

Results for “learning classifiers”

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.

SourceResearch connectionFrom yearThrough year

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

At least 253 records · Page 14

All sources NASA NTRS DOE OSTI

DOE OSTI2026

Data Summarization and Inference at Scale

This is the final report for the DOE ASCR grant SC-0022260, Data Summarization and Inference at Scale, PI: Alex Pothen, Purdue University. The goal of the project was to solve data-intensive and compute-intensive problems in the physical sciences, engineering, information science, data science, etc. by designing and implementing new algorithms that could work with a subset of the data. The four subgoals were: (a) The solution of problems where the data is too large to be stored in the memory of a computer. In this streaming model of computation, the data arrives as a stream of elements to the computer, each element is processed as it arrives, and a decision is made to discard the data or to store it; only a small subset of the data proportional to the size of the output solution is stored, and when all the data has been streamed, a solution to the problem is computed from the stored subset. (b) The use of machine learning methods to compute solutions to data-intensive problems. The use of GPUs is critical to obtain high performance on machine learning tasks, but their memory sizes are smaller relative to that of CPUs. For large-scale problems, the data is sampled many times, and small samples are used with repetition, for robustness, to compute solutions to inference tasks. This sampling reduces the memory required to solve the problem, but attention is needed to avoid slow convergence to the solutions, and reduced accuracy of inference. We propose submodular optimization, Large Language Models, and physics-informed neural networks to enable GPU computations here. (c) Modeling and visualization of high-dimensional data using interpretable features. Clinical proteomic data sets from immunology for the detection of cancer and other diseases are temporal and high-dimensional, and algorithms for visualizing these data sets using clinically interpretable features are lacking. We propose methods that compute distances based on the optimal transportation problem and graph edit distances to address this problem. We also propose the use of optimal transport-based distances, spatial statistics, and network structure to classify image data sets, We apply these algorithms to electron micrographs of the peripheral nervous system in the digestive tract. (d) The design of data-intensive algorithms on emerging architectures, specifically, noisy, intermediate-scale quantum (NISQ) devices. Quantum computers offer the possibility of exploring large solution spaces due to the principle of superposition, but current quantum computers are limited by few qubits, short coherence times due to noise, poor interconections among the qubits, etc. We propose the use of the divide and conquer paradigm to solve large-scale problems, wherein collections of small subproblems are solved on the quantum devices, and the solutions to the subproblems are integrated into a solution for the original problem on a classical computer.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2026

An Iterative Machine Learning Framework for Event Classification and Monte Carlo Tuning in SpinQuest

The E1039/SpinQuest experiment at Fermi National Accelerator Laboratory uses a 120~GeV proton beam from the Main Injector incident on transversely polarized proton and deuteron targets, using $NH_3$ and $ND_3$, respectively. In addition to measuring the Sivers asymmetry in Drell--Yan $pp$ and $pd$ scattering from sea quarks, SpinQuest will study transverse-spin effects, particularly the transverse single-spin asymmetry (TSSA) in $J/\psi$ production. The angular distributions from the $J/\psi$ decay could play an important role in understanding the gluon contribution to the proton spin structure. However, before extracting these angular distributions, it is necessary to isolate signal events originating from the target from events produced by other sources and from the combinatorial background. To effectively and accurately classify the target events, it is important to ensure that the simulated events are properly tuned to the experimental physics channels. We have introduced an iterative technique to match simulated and experimental events and to classify the physics channels using deep neural networks and a generative model based on normalizing flows.

Hossain, Forhad [Virginia U. (main)] (ORCID:000000↗

DOE OSTI2024

Anomaly detection with flow-based fast calorimeter simulators

Recently, several normalizing flow-based deep generative models have been proposed to accelerate the simulation of calorimeter showers. Using caloflow as an example, we show that these models can simultaneously perform unsupervised anomaly detection with no additional training cost. As a demonstration, we consider electromagnetic showers initiated by one (background) or multiple (signal) photons. The caloflow model is designed to generate single-photon showers, but it also provides access to the shower likelihood. We use this likelihood as an anomaly score and study the showers tagged as being unlikely. As expected, the tagger struggles when the signal photons are nearly collinear but is otherwise effective. This approach is complementary to a supervised classifier trained on only specific signal models using the same low-level calorimeter inputs. While the supervised classifier is also highly effective at unseen signal models, the unsupervised method is more sensitive in certain regions, and, thus, we expect that the ultimate performance will require a combination of these approaches.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2024

Exploring electron-beam induced modifications of materials with machine-learning assisted high temporal resolution electron microscopy

Directed atomic fabrication using an aberration-corrected scanning transmission electron microscope (STEM) opens new pathways for atomic engineering of functional materials. In this approach, the electron beam is used to actively alter the atomic structure through electron beam induced irradiation processes. One of the impediments that has limited widespread use thus far has been the ability to understand the fundamental mechanisms of atomic transformation pathways at high spatiotemporal resolution. Here, we develop a workflow for obtaining and analyzing high-speed spiral scan STEM data, up to 100 fps, to track the atomic fabrication process during nanopore milling in monolayer MoS 2 . An automated feedback-controlled electron beam positioning system combined with deep convolution neural network (DCNN) was used to decipher fast but low signal-to-noise datasets and classify time-resolved atom positions and nature of their evolving atomic defect configurations. Through this automated decoding, the initial atomic disordering and reordering processes leading to nanopore formation was able to be studied across various timescales. Using these experimental workflows a greater degree of speed and information can be extracted from small datasets without compromising spatial resolution. This approach can be adapted to other 2D materials systems to gain further insights into the defect formation necessary to inform future automated fabrication techniques utilizing the STEM electron beam.

36 MATERIALS SCIENCE↗

DOE OSTI2024

Scalar bounded-from-below conditions from Bayesian active learning

We present a procedure leveraging Bayesian deep active learning to rapidly produce highly accurate approximate bounded-from-below conditions for arbitrary renormalizable scalar potentials, in the form of a neural network which may be saved and exported for use in arbitrary parameter space scans. We explore the performance of our procedure on three different scalar potentials with either highly nontrivial or unknown symbolic bounded-from-below conditions (the most general two-Higgs doublet model, the three-Higgs doublet model, and a version of the Georgi-Machacek model without custodial symmetry). We find that we can produce fast and highly accurate binary classifiers for all three potentials. Furthermore, for the potentials for which no known symbolic necessary and sufficient conditions on boundedness-from-below exist, our classifiers substantially outperform some common approximate analytical methods, such as producing tractable sufficient but not necessary conditions or evaluating boundedness-from-below conditions for scenarios in which only a subset of the theory’s fields achieve vacuum expectation values. Our methodology can be readily adapted to any renormalizable scalar field theory. For the community’s use, we have developed a package, BFBrain, which allows for the rapid implementation of our analysis procedure on user-specified scalar potentials with a high degree of customizability. Published by the American Physical Society 2024

Astronomy & Astrophysics↗

DOE OSTI2025

Benchmarking a Tunable Quantum Neural Network on Trapped-Ion and Superconducting Hardware

We implement a quantum generalization of a neural network on trapped-ion and IBM superconducting quantum computers to classify MNIST images, a common benchmark in computer vision. The network feedforward involves qubit rotations whose angles depend on the results of measurements in the previous layer. The network is trained via simulation, but inference is performed experimentally on quantum hardware. The classical-to-quantum correspondence is controlled by an interpolation parameter, $a$, which is zero in the classical limit. Increasing $a$ introduces quantum uncertainty into the measurements, which is shown to improve network performance at moderate values of the interpolation parameter. We then focus on particular images that fail to be classified by a classical neural network but are detected correctly in the quantum network. For such borderline cases, we observe strong deviations from the simulated behavior. We attribute this to physical noise, which causes the output to fluctuate between nearby minima of the classification energy landscape. Such strong sensitivity to physical noise is absent for clear images. We further benchmark physical noise by inserting additional single-qubit and two-qubit gate pairs into the neural network circuits. Our work provides a springboard toward more complex quantum neural networks on current devices: while the approach is rooted in standard classical machine learning, scaling up such networks may prove classically non-simulable and could offer a route to near-term quantum advantage.

FOS: Physical sciences↗

DOE OSTI2024

SSTDR and FDR Detection of Un-Energized and Energized Cable Anomalies Including Thermal Degradation Using Machine Learning

Historically, cables are initially qualified for nuclear power plant use for 40 years. As plants extend their operating license to 60 and 80 years, continued use of these cables must shift to a performance-based approach since it is cost prohibitive to completely replace cables that are likely still capable of performing their design function. A variety of cable tests are available and are commonly applied during outages when the cables can be taken out of service. Frequency domain reflectometry (FDR) is one of these test methods that is being more broadly accepted and used because it not only detects anomalies along the cable with a low-voltage signal that does not stress the cable insulation, but the technique also locates the anomalies. This supports follow-up local inspection and local repair or partial replacement of a damaged cable segment. Currently, FDR testing is only applied to cables that are taken out of service since the test instrument would be damaged by operational voltages. A related technology that has found some acceptance in the aircraft and rail industry is spread spectrum time domain reflectometry (SSTDR). This technology has been implemented with a custom commercial instrument by LiveWire Innovation that is designed to operate on live cables up to 1000 volts and with a bandwidth of 48 MHz. Initial evaluation by the Pacific Northwest National Laboratory (PNNL) of the Live Wire system indicated that a broader bandwidth (BW) SSTDR may be better for many kinds of flaws. This led PNNL to develop an SSTDR laboratory instrument suitable for tests up to 500 MHz bandwidth. Testing on energized cables is also desirable for online monitoring systems so an inductive clamshell coupler was developed that allows energized cables to be tested up to at least 5 kV and likely higher voltage levels. Dielectric spectroscopy and tan delta testing plus various laboratory destructive tests were included in this data acquisition campaign directed to feed a machine learning (ML) study. With these kinds of developments, online energized cable tests may be possible with industrial adoption of such hardware advances but it will be completely impractical to have highly skilled data analysts continually examine these complex signals for indications of damage or compromised conditions. If online testing is to be implemented in new test hardware, it must be accompanied by software that can interpret the signals and alert plant operators of changing or degraded conditions. The thermally aged, shielded cable investigated here was separately treated for ML analysis. Visual analysis of electrical data showed generally increasing peaks where the cable entered and exited the oven. These peaks were not exactly aligned with expected locations, but these differences were attributed to velocity of propagation calibration errors. Only supervised ML was applied to the thermally aged data as this data was only available shortly before the committed publication date of this report. The supervised ML was structured to divide the 0 to 70-day responses as ‘normal’ from 0 to 35 days or ‘anomalous’ from 36 to 70 days, based on cable tensile elongation at break (EAB) insulation characterization. Using 80% of the data for training and 20% for testing, the supervised ML predicted normal versus anomalous was 70% accurate. Important conclusions include: • Accuracy to predict the presence of cable damage is improved from the 2023 effort by more training data. Weighted accuracies for comparisons among the instruments ranged from 67 to 89 % for unsupervised ML and 71 to 99% for supervised ML. • Based on the synthetic data tests, the unsupervised models are more generalizable to unseen anomalies. The Multi-Layer Perceptron classifier (MLP) model reported as high as 99.7% accuracy on the test data, but this dropped to 58.3% when tested on the synthetic data. In contrast, the unsupervised Pointwise model only achieved 89.7% accuracy on the experimental data but reported 78.3% accuracy on the synthetic data. • The best anomaly indicators are higher frequency (400 MHz BW) FDR data. Other tests may be interesting but for this study, this was the best predicter.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

DOE OSTI2026

Wasserstein Normalized Autoencoder for Anomaly Detection in ProtoDUNE Vertical-Drift Detector

ProtoDUNE Vertical Drift needs a selective triggering algorithm. The detector sits on Earth's surface, so cosmic activity dominates its data. Our goal in this paper is to trigger on neutrino events more robustly than the current deployed Analog-to-Digital Converter Simple Window (ADCSW) model and, eventually, search for signals of Beyond Standard Model (BSM) physics at DUNE as our ultimate North Star objective. As a step towards this goal, we evaluate a Wasserstein Normalized Autoencoder (WNAE) on simulated collection-plane only windows of shape $1\times10\times10$ where Neutrinos act as our BSM-proxy and Cosmic-ray Muons serve as our learned background. The network parameters are fitted using only cosmic-ray muon events as background in order to maintain an unsupervised pipeline. Training uses finite-step Langevin $x^-$ samples, positive-sample reconstruction energy, and an empirical sliced $2$-Wasserstein objective to learn a normalized Boltzmann energy model. We then calibrate on a nominal $5\,\mathrm{Hz}$ operating threshold calculated from cosmic validation data. Both WNAE and ADCSW accept 311 of 194,083 held-out cosmic background events at this $5\,\mathrm{Hz}$ threshold. We found that WNAE accepts 9,677 of 34,634 neutrino-proxy events $(27.9\pm0.24)\%$, compared with 10,076 $(29.1\pm0.24)\%$ for ADCSW, an observed WNAE-minus-ADCSW difference of $-1.15\%$. At another nominal $2\,\mathrm{Hz}$ target threshold, the corresponding efficiencies are $(20.5\pm0.22)\%$ and $(22.6\pm0.22)\%$, respectively. Of the WNAE-selected neutrino proxies at $5\,\mathrm{Hz}$, $(20.8\pm0.4)\%$ of the classified neutrino-proxy events are unique to WNAE, where the uncertainty is an absolute binomial standard error of $0.4\%$.

Zheng, Jake [U. Chicago (main)] (ORCID:00090002189↗

DOE OSTI2026

Wasserstein Normalized Autoencoder for Anomaly Detection in ProtoDUNE Vertical-Drift Detector

Zheng, Jake [Chicago U.] (ORCID:0009000218901379)↗

DOE OSTI2024

Circumventing data imbalance in magnetic ground state data for magnetic moment predictions

Abstract Magnetic materials play a crucial role in the transition to more sustainable forms of energy and electric vehicles. There is an anticipated shortage in magnetic materials in the future, and as a result there is an urgent need to discover and design new magnetic materials. Computational magnetic material design using density functional theory is daunting because of the challenge in identifying magnetic ground states from a combinatorially large set of possibilities. Machine learning offers a path forward by enabling efficient surrogate models that can more readily enumerate these states, but there is a dearth of training data available, and what is available tends to be imbalanced with too much non-magnetic data. In this work we show that the discrete and previously tackled data imbalance that exists at the level of the magnetic ordering leads to an imbalanced continuous distribution with many zeros when the data is unraveled at the atomic magnetic moment level, which subsequently leads to models with low accuracy for magnetic properties. We mitigate this by using a two-part model framework. Our scheme is able to classify atoms into magnetic and non-magnetic with an F1 score and Matthew’s correlation coefficient (MCC) of ~91% and then to provide an implicit embedding representation that maps directly onto the magnitude of the magnetic moment with a mean absolute error of 0.1 μ B . Beyond screening for new magnetic materials, we demonstrate an additional practical use case of our scheme: the provision of good initial guesses for magnetic moments in first-principles electronic relaxations. Such initialization is shown to lead to faster convergence to configurations that lie closer to the ground state.

Computer Science↗

DOE OSTI2024

Machine Learning Correlation of Electron Micrographs and ToF-SIMS for the Analysis of Organic Biomarkers in Mudstone

The spatial distribution of organics in geological samples can be used to determine when and how these organics were incorporated into the host rock. Mass spectrometry (MS) imaging can rapidly collect a large amount of data, but ions produced are mixed without discrimination, resulting in complex mass spectra that can be difficult to interpret. Here, we apply unsupervised and supervised machine learning (ML) to help interpret spectra from time-of-flight-secondary ion mass spectrometry (ToF-SIMS) of an organic-carbon-rich mudstone of the Middle Jurassic of England (UK). It was previously shown that the presence of sterane molecular biomarkers in this sample can be detected via ToF-SIMS (Pasterski, M. J. et al., Astrobiology 2023, 23, 936). We use unsupervised ML on scanning electron microscopy–electron dispersive spectroscopy (SEM-EDS) measurements to define compositional categories based on differences in elemental abundances. We then test the ability of four ML algorithms─k-nearest neighbors (KNN), recursive partitioning and regressive trees (RPART), eXtreme gradient boost (XGBoost), and random forest (RF)─to classify the ToF-SIM spectra using (1) the categories assigned via SEM-EDS, (2) organic and inorganic labels assigned via SEM-EDS, and (3) the presence or absence of detectable steranes in ToF-SIMS spectra. In terms of predictive accuracy and balanced accuracy, KNN was the best performing model and RPART the worst. The feature importance, or the specific features of the ToF-SIM spectra used by the models to make classifications, cannot be determined for KNN, preventing posthoc model interpretation. Nevertheless, the feature importance extracted from the other models was useful for interpreting spectra. In conclusion, we determined that some of the organic ions used to classify biomarker containing spectra may be fragment ions derived from kerogen which is abundant in this mudstone sample.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2026

Transforming jet flavour tagging at ATLAS

Jet flavour tagging enables the identification of jets originating from heavy-flavour quarks in proton–proton collisions at the Large Hadron Collider, playing a critical role in its physics programmes. This paper presents GN2, a transformer-based flavour tagging algorithm deployed by the ATLAS Collaboration that represents a different methodology compared to previous approaches. Designed to classify jets based on the flavour of their constituent particles, GN2 processes low-level tracking information in an end-to-end architecture and incorporates physics-informed auxiliary training objectives to enhance both interpretability and performance. Its performance is validated in both simulation and collision data. The measured c-jet (light-jet) rejection in data is improved by a factor of 3.5 (1.8) for a 70% b-jet tagging efficiency, compared to the previous algorithm. GN2 provides substantial benefits for physics analyses involving heavy-flavour jets, such as measurements of Higgs boson pair production and the couplings of bottom and charm quarks to the Higgs boson, and demonstrates the impact of advanced machine learning methods in experimental particle physics.

Characterization and analytical techniques↗

DOE OSTI2026

Hyperdimensional computing for image classification (HDC) v1.0

This is an implementation of the hyperdimensional computing technique to classify images. It consists of a python script that trains the system for a set of images from a set of images (dataset) specified by the user. This training produces hardware configuration parameters and description vectors that are then loaded into the hardware description part of the project. The hardware description consists of hardware described in Verilog (a well known language for this purpose) that is synthesizable and can be implemented in a real chip. This hardware received the training information generated by python, and then is able to accept images to produce answers for each image on which category (class) from the pre-=trained ones the image belongs to. The hardware and python training scripts are configurable and documented. The advantage of hyperdimensional computing is its robustness to errors and the easy capability for online learning (refining the training during inference slowly over time), which this implementation supports.

Michelogiannakis, Georgios [Lawrence Berkeley Nati↗

DOE OSTI2024

Scalable Risk Assessment of Rare Events in Power Systems With Uncertain Wind Generation and Loads

Risk assessment of rare events has become increasingly important in power system planning and operation with the increasing integration of renewable energy and the presence of system uncertainties. However, quantifying the risk posed by rare events via the traditional method, i.e., Monte Carlo sampling (MCS), incurs substantial computational expense stemming from the vast ensemble of power flow simulations. To accelerate the assessment, this paper proposes a Deep Neural Network (DNN)-kernelized vector-valued Gaussian Process (VVGP) approach with excellent computational efficiency while maintaining high accuracy. Consequently, serving as a surrogate model for the power flow solver, the DNN-kernelized VVGP enables significantly faster but accurate risk assessment compared to the power flow solver. The developed surrogate model evaluates low-order N - k events that contain more than 90% instances by adeptly capturing the topological features while the high-order N - k events are assessed via a power flow solver, thereby striking a balance between computational efficiency and uncertainty quantification accuracy. Moreover, the model incorporates a Support Vector Machine (SVM) classifier to resample concerning low-probability tail events to counteract the biases potentially introduced during the DNN-kernelized VVGP evaluations. Simulations conducted on the modified IEEE 24-bus, 118-bus, and European 1354-bus systems demonstrate that the proposed method maintains the accuracy benchmark set by MCS while significantly reducing computational demands in large-scale power systems as compared to other state-of-the-art methods.

17 WIND ENERGY↗

DOE OSTI2024

Ferroelectric capacitors and field-effect transistors as in-memory computing elements for machine learning workloads

Abstract This study discusses the feasibility of Ferroelectric Capacitors (FeCaps) and Ferroelectric Field-Effect Transistors (FeFETs) as In-Memory Computing (IMC) elements to accelerate machine learning (ML) workloads. We conducted an exploration of device fabrication and proposed system-algorithm co-design to boost performance. A novel FeCap device, incorporating an interfacial layer (IL) and$$\text {Hf}_{0.5}\text {Zr}_{0.5}\text {O}_2$$ Hf 0.5 Zr 0.5 O 2 (HZO), ensures a reduction in operating voltage and enhances HZO scaling while being compatible with CMOS circuits. The IL also enriches ferroelectricity and retention properties. When integrated into crossbar arrays, FeCaps and FeFETs demonstrate their effectiveness as IMC components, eliminating sneak paths and enabling selector-less operation, leading to notable improvements in energy efficiency and area utilization. However, it is worth noting that limited capacitance ratios in FeCaps introduced errors in multiply-and-accumulate (MAC) computations. The proposed co-design approach helps in mitigating these errors and achieves high accuracy in classifying the CIFAR-10 dataset, elevating it from a baseline of 10% to 81.7%. FeFETs in crossbars, with a higher on-off ratio, outperform FeCaps, and our proposed charge-based sensing scheme achieved at least an order of magnitude reduction in power consumption, compared to prevalent current-based methods.

Science & Technology - Other Topics↗

DOE OSTI2026

Reconstructing Quasar Spectra and Measuring the Lyα Forest with SpenderQ

Quasar spectra carry the imprint of foreground intergalactic medium (IGM) through absorption features. In particular, absorption caused by neutral hydrogen gas, the "Lyα forest," is a key spectroscopic tracer for cosmological analyses used to measure cosmic expansion and test physics beyond the standard model. Despite their importance, current methods for measuring Lyα absorption cannot directly derive the intrinsic quasar continuum and make strong assumptions on its shape, thus distorting the measured Lyα clustering. We present SpenderQ , a ML-based approach for directly reconstructing the intrinsic quasar spectra and measuring the Lyα forest from observations. SpenderQ uses the Spender spectrum autoencoder to learn a compact and redshift-invariant latent encoding of quasar spectra, combined with an iterative procedure to identify and mask absorption regions. To demonstrate its performance, we apply SpenderQ to 400,000 synthetic quasar spectra created to validate the Dark Energy Spectroscopic Instrument Year 1 Lyα cosmological analyses. SpenderQ accurately reconstructs the true intrinsic quasar spectra, including the broad Lyβ, Lyα, SiIV, CIV, and CIII emission lines. Redward of Lyα, SpenderQ provides percent-level reconstructions of the true quasar spectra. Blueward of Lyα, SpenderQ reconstructs the true spectra to < 5%. SpenderQ reproduces the shapes of individual quasar spectra more robustly than the current state-of-the-art. We, thus, expect it will significantly reduce biases in Lyα clustering measurements and enable studies of quasars and their physical properties. SpenderQ also provides informative latent variable encodings that can be used to, e.g., classify quasars with Broad Absorption Lines. Overall, SpenderQ provides a new data-driven approach for unbiased Lyα forest measurements in cosmological, quasar, and IGM studies.

Hahn, ChangHoon [Arizona U., Astron. Dept. - Stewa↗

DOE OSTI2025

Leveraging explainable AI to characterize floating-point exceptions in linear solvers

Linear solver packages are central to many scientific, engineering, and machine learning applications. When floating-point exceptions occur in these solvers, e.g., division by zero or overflow, numerical results are compromised and become unreliable. Existing static and dynamic analysis tools can detect such exceptions, but they do not explain why the exceptions occur in terms of the solver inputs. Here, we present a study to characterize the inputs that cause numerical exceptions in linear solver packages. Our approach uses explainable AI (XAI) to find the most relevant characteristics of input matrices that explain the occurrence of exceptions in the solvers. Since training data in this domain is scarce, we perform extensive data gathering and data augmentation to obtain exception-inducing inputs. Our approach uses a repair strategy on the features blamed by XAI to validate that such features indeed explain the exceptions. We compare the LIME and SHAP XAI techniques using a dozen matrix features with three classifiers. We evaluate the approach on three widely used linear solver packages and find that some input characteristics can explain the occurrence of exceptions 100% of the time, in specific solvers and preconditioners.

Explainable AI↗

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↗