Search NASA⌕ Search

SEARCH · Search NASA

Results for “Interpretability”

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 91 records · Page 5
All sourcesNASA NTRSDOE OSTI
NASA NTRS1987

A review of computer aided interpretation technology for the evaluation of radiographs of aluminum welds

Industrial radiography is a well established, reliable means of providing nondestructive structural integrity information. The majority of industrial radiographs are interpreted by trained human eyes using transmitted light and various visual aids. Hundreds of miles of radiographic information are evaluated, documented and archived annually. In many instances, there are serious considerations in terms of interpreter fatigue, subjectivity and limited archival space. Quite often it is difficult to quickly retrieve radiographic information for further analysis or investigation. Methods of improving the quality and efficiency of the radiographic process are being explored, developed and incorporated whenever feasible. High resolution cameras, digital image processing, and mass digital data storage offer interesting possibilities for improving the industrial radiographic process. A review is presented of computer aided radiographic interpretation technology in terms of how it could be used to enhance the radiographic interpretation process in evaluating radiographs of aluminum welds.

Lloyd, J. F., Sr.↗
NASA NTRS1989

Image understanding and interpretation in microwave diversity imaging

The authors investigate microwave imaging of metallic objects using a diversity method and interpret and predict the reconstructed image from an approach based on analysis of the scattering mechanism and a back-projection algorithm used in image retrieval. The connection between the various scattering mechanisms and the reconstructed images is discussed, what the images represent is interpreted, and a prediction is made as to what the image will look like over given spectral and angular windows. A brief description is given of the microwave diversity imaging system and the formulation of the microwave diversity imaging based on the physical optics approximation. The scattering mechanism of a complex shaped metallic object is then briefly reviewed and an alternate approach to interpreting the reconstructed image based on the understanding of the scattering mechanism and the reconstruction algorithm is given. Several numerical and experimental examples are included to support this interpretation approach.

Li, Hsueh-Jyh↗
NASA NTRS1994

Synoptic interpretation of measurements from HALOE

The degree to which the Southern Hemisphere polar vortex is isolated against horizontal (isentropic) mixing is investigated using data from the Halogen Occultation Experiment (HALOE), U.K. Meteorological Office (UKMO) potential vorticity (PV), and contour advection diagnostics. Measurements of methane and water vapor taken by HALOE during a disturbed period in the Southern Hemisphere springtime (21 September - 15 October 1992) are interpreted in light of the prevailing synoptic meteorology. Daily fields of winds and PV are shown to be essential in the interpretation of the data. A climatological high pressure region is responsible for a distorted vortex, and a substantial 'vortex stripping' event is present, associated with the early stages of vortex breakdown. This leads to significant temporal, zonal, and altitudinal variations in the distribution of tracers. The authors point out the difficulties this presents for the interpretation of solar occultation data, especially with regard to the use of zonal average time series. Longitude-height methane distributions from two days during the period are examined. Both days show substantial variations in abundance around a latitude circle. In particular, the authors investigate HALOE measurements at 77 deg S on 15 October 1992, which indicate an abundance of methane in the height region 600-2000 K (approximately 30-1 mb) that is more typical of midlatitude air. Similar distributions, observed in the 1991 HALOE data, have previously been interpreted as evidence for the penetration of midlatitude air into the vortex. Gradients of potential vorticity and contour advection diagnositcs are employed to examine whether the UKMO winds are consistent with this hypothesis in 1992. Although midlatitude air is able to penetrate poleward of the main jet core by advection processes alone, an essentially intact inner core of vortex air remains, which does not mix to any great extent with air from lower latitudes. The authors show that the high-latitude HALOE abundances that are typical of midlatitude air were observed in a region of extensive filamentation and mixing, rather than within the inner, more isolated, core.

Bithell, M.↗
NASA NTRS1991

On the Power of Abstract Interpretation

Increasingly sophisticated applications of static analysis place increased burden on the reliability of the analysis techniques. Often, the failure of the analysis technique to detect some information my mean that the time or space complexity of the generated code would be altered. Thus, it is important to precisely characterize the power of static analysis techniques. We follow the approach of Selur et. al. who studied the power of strictness analysis techniques. Their result can be summarized by saying 'strictness analysis is perfect up to variations in constants.' In other words, strictness analysis is as good as it could be, short of actually distinguishing between concrete values. We use this approach to characterize a broad class of analysis techniques based on abstract interpretation including, but not limited to, strictness analysis. For the first-order case, we consider abstract interpretations where the abstract domain for data values is totally ordered. This condition is satisfied by Mycroft's strictness analysis that of Sekar et. al. and Wadler's analysis of list-strictness. For such abstract interpretations, we show that the analysis is complete in the sense that, short of actually distinguishing between concrete values with the same abstraction, it gives the best possible information. We further generalize these results to typed lambda calculus with pairs and higher-order functions. Note that products and function spaces over totally ordered domains are not totally ordered. In fact, the notion of completeness used in the first-order case fails if product domains or function spaces are added. We formulate a weaker notion of completeness based on observability of values. Two values (including pairs and functions) are considered indistinguishable if their observable components are indistinguishable. We show that abstract interpretation of typed lambda calculus programs is complete up to this notion of indistinguishability. We use denotationally-oriented arguments instead of the detailed operational arguments used by Selur et. al.. Hence, our proofs are much simpler. They should be useful for further future improvements.

Reddy, Uday S.↗
NASA NTRS2004

Understanding AOP through the Study of Interpreters

I return to the question of what distinguishes AOP languages by considering how the interpreters of AOP languages differ from conventional interpreters. Key elements for static transformation are seen to be redefinition of the set and lookup operators in the interpretation of the language. This analysis also yields a definition of crosscutting in terms of interlacing of interpreter actions.

Filman, Robert E.↗
NASA NTRSDate not supplied

Interpretable Convolutional Learning Classifier System (C-LCS) for Higher Dimensional Datasets

The purpose of this paper is to devise an interpretable hybrid classification model for Convolutional Neural Networks (CNN) and a Learning Classifier System (LCS). The presented hybrid system integrates the fundamental attributes from both types of these classifiers. In the proposed hybrid model CNN works as an automatic feature extractor, and LCS works to provide interpretable rule-based classification results. Although LCS has limitations working on higher dimensional datasets, we resolve this limitation by using CNN as a feature extractor. The other concept of the non-interpretability of CNN is addressed by using the LCS rule. Furthermore, our experiment with higher dimensional datasets like CIFAR-10 and Fashion-MNIST shows that extended LCS provides comparable performance to the standard neural network model while also providing interpretable results. We named this extended LCS method Convolutional Learning Classifier Cystem (C-LCS).

Jelani Owens↗
DOE OSTI2022

An interpretable machine learning model for advancing terrestrial ecosystem predictions

We apply an interpretable Long Short-Term Memory (iLSTM) network for land-atmosphere carbon flux predictions based on time series observations of seven environmental variables. iLSTM enables interpretability of variable importance and variable-wise temporal importance to the prediction of targets by exploring internal network structures. The application results indicate that iLSTM not only improves prediction performance by capturing different dynamics of individual variables, but also reasonably interprets the different contribution of each variable to the target and its different temporal relevance to the target. This variable and temporal importance interpretation of iLSTM advances terrestrial ecosystem model development as well as our predictive understanding of the system.

Lu, Dan↗
DOE OSTI2025

AI for Interpreting Nuclear Power Plant Documents for Power Uprates

To reduce the cost and time needed for regulatory compliance, nuclear power plants (NPPs) can utilize artificial intelligence (AI) to assist in interpreting complex and voluminous documents that typically span thousands of pages. Usually, the process of interpreting a plant’s technical specifications (TSs) and associated documents is labor intensive. This study aims to understand what processes state-of-the-art large language models (LLMs) can automate and to identify the pitfalls associated with using LLMs to reduce human labor costs and time. This research uses a recent AI technology called retrieval augmented generation (RAG), which retrieves pages of information from TSs and associated documents to assist with NPP power uprates (cleared to produce more power). LLMs are integral to RAG because they create human-like responses based on the retrieved information, aiding in the interpretation and application processes. A baseline case demonstrates how LLMs can operate successfully for a power uprate application. Then five use cases show five types of potential failures: (1) RAG retrieving the incorrect information, (2) RAG misinterpreting the retrieved information, (3) RAG relying on knowledge not contained in the retrieved information, (4) RAG hallucinating, and (5) RAG refusing to answer. The results of the five use cases suggest that automating the human interpretation of TSs and associated documents with AI should be approached with caution. A subject-matter expert reviewed the AI outputs from the five use cases and concluded that an LLM can produce technical information that is needed to produce power uprate applications in certain instances.

21 - SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLAN↗
DOE OSTI2026

Interpretable Machine Learning for Characterizing Electric Vehicle Charging Behavior: Insights from Real-World Data

As electric vehicle (EV) adoption rises globally, concerns about the impact on aging electrical grids grow, particularly regarding the charging behavior of EV drivers. This study analyzes real-world driving and charging data from Ford battery electric vehicles (BEVs) collected between 2018 and 2019 to develop interpretable models that characterize charging behavior and quantify influencing factors. Prior research has relied on assumptions regarding driver behavior, often overlooking actual charging patterns. By employing generalized linear mixed models (GLMMs), this work offers insights into how various elements, such as next trip distance and state of charge (SOC), influence charging decisions. The dataset comprises over three million park-trip pairs from 1,997 vehicles, revealing that features related to driving behavior significantly dictate charging behavior, while infrastructure and regional factors have lesser impacts. The findings suggest that existing simulation models may oversimplify EV charging behavior assumptions. This work utilizes real-world EV driving and charging data to train interpretable models that describe charging behavior and quantify the factors most associated with how drivers use charging infrastructure. This research underscores the need for interpretable, data-driven methodologies to inform future EV infrastructure planning and grid management.

29 - ENERGY PLANNING, POLICY AND ECONOMY↗
DOE OSTI2018

Recurrent Neural Network Attention Mechanisms for Interpretable System Log Anomaly Detection

Deep learning has recently demonstrated state-of-the art performance on key tasks related to the maintenance of computer systems, such as intrusion detection, denial of service attack detection, hardware and software system failures, and malware detection. In these contexts, model interpretability is vital for administrator and analyst to trust and act on the automated analysis of machine learning models. Deep learning methods have been criticized as black box oracles which allow limited insight into decision factors. In this work we seek to "bridge the gap'' between the impressive performance of deep learning models and the need for interpretable model introspection. To this end we present recurrent neural network (RNN) language models augmented with attention for anomaly detection in system logs. Our methods are generally applicable to any computer system and logging source. By incorporating attention variants into our RNN language models we create opportunities for model introspection and analysis without sacrificing state-of-the art performance. We demonstrate model performance and illustrate model interpretability on an intrusion detection task using the Los Alamos National Laboratory (LANL) cyber security dataset, reporting upward of 0.99 area under the receiver operator characteristic curve despite being trained only on a single day's worth of data.

Anomaly detection, Attention, Recurrent Neural Net↗
NASA NTRS2017

Interpretable Categorization of Heterogeneous Time Series Data

We analyze data from simulated aircraft encounters to validate and inform the development of a prototype aircraft collision avoidance system. The high-dimensional and heterogeneous time series dataset is analyzed to discover properties of near mid-air collisions (NMACs) and categorize the NMAC encounters. Domain experts use these properties to better organize and understand NMAC occurrences. Existing solutions either are not capable of handling high-dimensional and heterogeneous time series datasets or do not provide explanations that are interpretable by a domain expert. The latter is critical to the acceptance and deployment of safety-critical systems. To address this gap, we propose grammar-based decision trees along with a learning algorithm. Our approach extends decision trees with a grammar framework for classifying heterogeneous time series data. A context-free grammar is used to derive decision expressions that are interpretable, application-specific, and support heterogeneous data types. In addition to classification, we show how grammar-based decision trees can also be used for categorization, which is a combination of clustering and generating interpretable explanations for each cluster. We apply grammar-based decision trees to a simulated aircraft encounter dataset and evaluate the performance of four variants of our learning algorithm. The best algorithm is used to analyze and categorize near mid-air collisions in the aircraft encounter dataset. We describe each discovered category in detail and discuss its relevance to aircraft collision avoidance.

Drones↗
NASA NTRSDate not supplied

On the New Optical Constants Database (OCdb) and its Importance for the Interpretation of Observational Data

The Optical Constants database(ocdb.smce.nasa.gov) came online in February 2023 and provides complex refractive indices of laboratory-generated organic refractory materials and ices relevant to (exo) planetary and astrophysical environments.The goal of the OCdb is to centralize published optical constants data to facilitate both their access by the scientific community and their use to analyze observational data returned by space missions and ground-based observatories. Computational tools are also under development to facilitate scientific use of the available OCdb optical constants data sets. Investigators generating laboratory optical constants are therefore encouraged to contribute their data to OCdb in order to increase the availability of their data and to enhance the scientific effectiveness of the database. Optical constants are critical input parameters in models (e.g.,radiative transfer, atmospheric, and reflectance spectral models)that are used to simulate the absorption, reflection, and scattering of light due to solid materials present in planetary and astrophysical environments (planets, their satellites, exoplanets, asteroids, comets, protoplanetary disks, etc.), and are key to the compositional interpretation of observations. We will first present the infrastructure of the OCdb and show how to use and contribute to it. We will introduce the two large NASA projects, namely, the Laboratory Astrophysics Directed Work Package and the NASA Center for Optical Constants, that have been instrumental in (i) developing the OCdb, (ii) generating planetary-and astrophysics-relevant ices and organic refractory materials from gas and ice irradiation in the laboratory, and (iii) determining their optical constants for inclusion in OCdb. We will also present two studies that are making use of these optical constants to interpret observations of Titan’s atmosphere and Pluto’s surface. These studies show the importance of measuring optical constants of laboratory-generated materials, and their impact on the models used to analyze and interpret astronomical observations. These studies also demonstrate the essential importance of such a database and the need for optical constants of a broad range of materials and wavelengths to enable the scientific community and to maximize the scientific return from space missions (e.g., Cassini, New Horizons, SOFIA, JWST).

The Optical Constants database↗
DOE OSTI2020

Evaluation of a visual interpretation method for tau-PET with 18 F-flortaucipir

Introduction: Positron emission tomography targeting tau (tau-PET) is a promising diagnostic tool for the identification of Alzheimer's disease (AD). Currently available data rely on quantitative measures, and a visual interpretation method, critical for clinical translation, is needed. Methods: In this work, we developed a visual interpretation method for 18 F-flortaucipir tau-PET and tested it on 274 individuals (cognitively normal controls, patients with mild cognitive impairment [MCI], AD dementia, and non-AD diagnoses). Two readers interpreted 18 F-flortaucipir PET using two complementary indices: a global visual score and a visual distribution pattern. Results: Global visual scores were reliable, correlated with global cortical 18 F-flortaucipir standardized uptake value ratio (SUVR) and were associated with clinical diagnosis and amyloid status. The AD-like 18 F-flortaucipir pattern had good sensitivity and specificity to identify amyloid-positive patients with AD dementia or MCI. Discussion: This 18 F-flortaucipir visual rating scheme is associated with SUVR quantification, clinical diagnosis, and amyloid status, and constitutes a promising approach to tau measurement in clinical settings.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Implications of the XENON1T excess on the dark matter interpretation

The dark matter interpretation for a recent observation of excessive electron recoil events at the XENON1T detector seems challenging because its velocity is not large enough to give rise to recoiling electrons of O(keV). Fast-moving or boosted dark matter scenarios are receiving attention as a remedy for this issue, rendering the dark matter interpretation a possibility to explain the anomaly. We investigate various scenarios where such dark matter of spin 0 and 1/2 interacts with electrons via an exchange of vector, axial-vector, pseudo-scalar, or scalar mediators. We find parameter values not only to reproduce the excess but to be consistent with existing bounds. Our study suggests that the scales of mass and coupling parameters preferred by the excess can be mostly affected by the type of mediator, and that significantly boosted dark matter can explain the excess depending on the mediator type and its mass choice. The method proposed in this work is general, and hence readily applicable to the interpretation of observed data in the dark matter direct detection experiment.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2021

Interpreting the Lipidome: Bioinformatic Approaches to Embrace the Complexity

Background Improvements in mass spectrometry (MS) technologies coupled with bioinformatics developments have allowed considerable advancement in the measurement and interpretation of lipidomics data in recent years. Since research areas employing lipidomics are rapidly increasing, there is a great need for bioinformatic tools that capture and utilize the complexity of the data. Currently, the diversity and complexity within the lipidome is often concealed by summing over or averaging individual lipids up to (sub)class-based descriptors, losing valuable information about biological function and interactions with other distinct lipids molecules, proteins and/or metabolites. Aim of review To address this gap in knowledge, novel bioinformatics methods are needed to improve identification, quantification, integration and interpretation of lipidomics data. The purpose of this mini-review is to summarize exemplary methods to explore the complexity of the lipidome. Key scientific concepts of review Here we describe six approaches that capture three core focus areas for lipidomics: (1) lipidome annotation including a resolvable database identifier, (2) interpretation via pathway- and enrichment-based methods, and (3) understanding complex interactions to emphasize specific steps in the analytical process and highlight challenges in analyses associated with the complexity of lipidome data.

Kyle, Jennifer E.↗
DOE OSTI2018

On the self-interference in electron scattering: Copenhagen, Bohmian and geometrical interpretations of quantum mechanics

Self-interference embodies the essence of the particle-wave formulation of quantum mechanics (QM). According to the Copenhagen interpretation of QM, self-interference by a double-slit requires a large transverse coherence of the incident wavepacket such that it covers the separation between the slits. Bohmian dynamics provides a first step in the separation of the particle-wave character of matter by introducing deterministic trajectories guided by a pilot wave that follows the time-dependent Schrödinger equation. In this work, I present a new description of the phenomenon of self-interference using the geometrical formulation of QM introduced in Tavernelli (2016). In particular, this formalism removes the need for the concept of wavefunction collapse in the interpretation of the act of measurement i.e., the emergence of the classical world. The three QM formulations (Schrödinger, Bohmian, and geometrical) are applied to the description of the scattering of a free electron by a hydrogen atom and a double-slit. The corresponding interpretations of self-interference are compared and discussed.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2023

How reliable is distribution of relaxation times (DRT) analysis? A dual regression-classification perspective on DRT estimation, interpretation, and accuracy

The distribution of relaxation times (DRT) has gained increasing attention and adoption in recent years as a versatile method for analyzing electrochemical impedance spectroscopy (EIS) data obtained from complex devices like fuel cells, electrolyzers, and batteries. The DRT deconvolutes the impedance without a priori specification of a generative model, which is especially useful for interpretation and model selection when the governing principles of the system under study are not fully understood. However, DRT estimation is an ill-posed inversion problem that must be addressed with a subjective choice of regularization and tuning, which leaves substantial risk of misleading interpretations of EIS data. In this work, we suggest a new classification view of the DRT inversion to clarify DRT estimation and interpretation. We introduce a dual regression-classification framework that unifies the classification and regression views of the DRT inversion with wide-reaching implications for DRT analysis. The dual framework is employed to demonstrate a new kind of DRT inversion algorithm and develop novel evaluation metrics that capture previously ignored aspects of DRT accuracy. These approaches are applied to both synthetic data and experimental spectra collected from a protonic ceramic fuel cell and a lithium-ion battery to illustrate their broad utility. The dual inversion algorithm shows promising performance for accurate DRT estimation and autonomous model identification, while the dual evaluation approach produces metrics that meaningfully assess the strengths and risks of DRT algorithms. Here this work provides valuable insight for both practical application of the DRT to experimental data and further development of EIS analysis methods.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Kinetic Isotope Effects: Interpretation and Prediction Using Degrees of Rate Control

Kinetic isotope effects (KIEs) have been used for decades in catalysis research as a tool for clarifying reaction mechanisms. Significant primary kinetic isotope effects have usually been interpreted as being a result of isotope substitution at a site of bond breaking (or forming) in the rate-determining step in the reaction mechanism. However, quantitative analysis of the magnitude of the KIE in complex multistep reaction mechanisms is seldom reported. We prove here that the logarithm of the rate ratio for two isotopes is the weighted average over all species in the mechanism of their standard-state free energy difference between the two isotopes, divided by RT. The weighting factor is the degree of rate control (DRC) for that species (e.g., transition state, intermediate, reactant) when the rate is measured separately for each isotope. It is instead the degree of selectivity ratio control (DSRC) when the KIE is measured as the product selectivity in a parallel competition between two isotopes within the same reactant molecule. Since only a few species have non-zero DRCs (or DSRCs) for most reactions, only these few contribute to this weighted average and the KIE. We show that this provides a simple way to interpret and quantitatively predict kinetic isotope effects that is powerful in the insights it provides, allowing one to evaluate directly which species contribute most to the KIE. By applying it to H/D KIEs in several example mechanisms, we further show that the traditional way of interpreting KIEs that focuses only on the rate-determining step can easily lead to misunderstanding of KIE and the reaction mechanism. Finally, this highlights the importance to consider the effect of isotope substitution on the energies of all species with large DRCs (i.e, those whose energies are kinetically relevant). This method also offers opportunities for quantitative validation of mechanism-based microkinetic models.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗