Search NASA⌕ Search

SEARCH · Search NASA

Results for “model 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 127 records · Page 7

Non-destructive measurement and real-time monitoring of apple hardness during ultrasonic contact drying via portable NIR spectroscopy and machine learning

Portable near-infrared spectrometer in the spectral range of 900–1700 nm was evaluated for the first time to assess and monitor apple hardness in real-time during ultrasonic drying. Calibration models were developed using PLS and ANN, and their performances were evaluated by internal leave-one-out cross-validation and an external dataset. Several pre-treatments including standard normal variate (SNV), multiplicative scatter correction (MSC), Savitzky–Golay first and second derivatives were employed to examine the effects of spectral variations in hardness prediction. Seven important wavelengths were selected using weighted regression coefficients to develop a simple MLR model to facilitate the model interpretation and circumvent noise. The models using PLS, MLR, and ANN with selected wavelengths predicted the apple hardness with R 2 p of 0.91, 0.91, 0.95, and RMSEP of 14.78, 14.85, and 12.46 N, respectively. Finally, the results indicate that portable NIR spectrometers are quite promising for real-time monitoring of apple hardness during ultrasonic drying.

47 OTHER INSTRUMENTATION↗

Intense pulsed neutron emission from a compact pyroelectric driven accelerator

Intense pulsed D-D neutron emission with rates of >10{sup 10} n/s during the pulse, pulse widths of approximately hundreds of nanoseconds and neutron yields of greater than 10 000 per pulse, are demonstrated in a compact pyroelectric accelerator. The accelerator consists of a small pyroelectric LiTaO{sub 3} crystal that provides the accelerating voltage and an independent compact spark plasma ion source. The crystal voltage versus temperature is characterized and compares well with theory. Results show neutron output per pulse that scales with voltage as V{sup {approx}}{sup 1.7}. These neutron yields match a simple model of the system at low voltages but are lower than predicted at higher voltages due to charge losses not accounted for in the model. Interpretation of the data against modeling provides understanding of the accelerator and in general pyroelectric LiTaO{sub 3} crystals operated as charge limited negative high voltage targets. The findings overall serve as the proof of principle and basis for pyroelectric neutron generators that can be pulsed, giving peak neutron rates orders of magnitude greater than previous work and notably increase the potential applications of pyroelectric based neutron generators.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Micro- and meso-scale pore structure in mortar in relation to aggregate content

Mortar is often viewed as a three-phase composite consisting of aggregate, bulk paste, and an interfacial transition zone (ITZ). However, this description is inconsistent with experimental findings because of the basic assumption that larger pores are only present within the ITZ. In this paper, we use backscattered electron (BSE) imaging to investigate the micro- and meso-scale structure of mortar with varying aggregate content. The results indicate that larger pores are present not only within the ITZ but also within areas far from aggregates. This phenomenon is discussed in detail based on a series of analytical calculations, such as the effective water binder ratio and the inter-aggregate spacing. We developed a modified computer model that includes a two-phase structure for bulk paste. This model interprets previous mercury intrusion porosimetry data very well. -- Highlights: •Based on BSE, we examine the HCSS model. •We develop the HCSS-DBLB model. •We use the modified model to interpret the MIP data.

36 MATERIALS SCIENCE↗

Teaching Freight Mode Choice Models New Tricks Using Interpretable Machine Learning Methods

Understanding and forecasting the intricate freight mode choice behavior under various industry, policy, and technology contexts is essential in freight planning and policymaking. Numerous models have been developed in prior studies to provide insights into freight mode selection, the majority of which use discrete choice models such as multinomial logit (MNL) models. However, logit models often rely on linear specifications of independent variables, despite potential nonlinear relationships in the data. Moreover, there often lacks a heuristic and efficient approach to identify such complex relationships to define the logit model specifications. To fill this gap, we developed an MNL model for freight mode choice using the insights from state-of-the- art machine learning (ML) models. ML models can capture the nonlinear nature of the complex decision-making process, and recent advances in 'explainable AI' have greatly improved their interpretability. The interpretable ML methods help enhance the performance of MNL models and advance knowledge of freight mode choice. Specifically, the influential factors and their relationship with individual modes are identified using SHapley Additive exPlanations (SHAP) to improve the MNL's performance. The workflow is demonstrated in a case study of Austin, Texas, and the SHAP results reveal multiple nonlinear relationships predicted by ML models. Incorporating those relationships into MNL model specifications improves the interpretability and accuracy of the MNL model compared to a conventional MNL model. Findings from this study can be used to guide freight planning and inform policymakers and practitioners on how key factors affect freight decision-making.

ADVANCED PROPULSION SYSTEMS,MATHEMATICS AND COMPUT↗

Digital data processing system dynamic loading analysis

Simulation and analysis of the Space Shuttle Orbiter Digital Data Processing System (DDPS) are reported. The mated flight and postseparation flight phases of the space shuttle's approach and landing test configuration were modeled utilizing the Information Management System Interpretative Model (IMSIM) in a computerized simulation modeling of the ALT hardware, software, and workload. System requirements simulated for the ALT configuration were defined. Sensitivity analyses determined areas of potential data flow problems in DDPS operation. Based on the defined system requirements and the sensitivity analyses, a test design is described for adapting, parameterizing, and executing the IMSIM. Varying load and stress conditions for the model execution are given. The analyses of the computer simulation runs were documented as results, conclusions, and recommendations for DDPS improvements.

Lagas, J. J.↗

Space shuttle orbiter digital data processing system timing sensitivity analysis OFT ascent phase

Dynamic loads were investigated to provide simulation and analysis of the space shuttle orbiter digital data processing system (DDPS). Segments of the ascent test (OFT) configuration were modeled utilizing the information management system interpretive model (IMSIM) in a computerized simulation modeling of the OFT hardware and software workload. System requirements for simulation of the OFT configuration were defined, and sensitivity analyses determined areas of potential data flow problems in DDPS operation. Based on the defined system requirements and these sensitivity analyses, a test design was developed for adapting, parameterizing, and executing IMSIM, using varying load and stress conditions for model execution. Analyses of the computer simulation runs are documented, including results, conclusions, and recommendations for DDPS improvements.

Lagas, J. J.↗

A Knowledge-Based Representation Scheme for Environmental Science Models

One of the primary methods available for studying environmental phenomena is the construction and analysis of computational models. We have been studying how artificial intelligence techniques can be applied to assist in the development and use of environmental science models within the context of NASA-sponsored activities. We have identified several high-utility areas as potential targets for research and development: model development; data visualization, analysis, and interpretation; model publishing and reuse, training and education; and framing, posing, and answering questions. Central to progress on any of the above areas is a representation for environmental models that contains a great deal more information than is present in a traditional software implementation. In particular, a traditional software implementation is devoid of any semantic information that connects the code with the environmental context that forms the background for the modeling activity. Before we can build AI systems to assist in model development and usage, we must develop a representation for environmental models that adequately describes a model's semantics and explicitly represents the relationship between the code and the modeling task at hand. We have developed one such representation in conjunction with our work on the SIGMA (Scientists' Intelligent Graphical Modeling Assistant) environment. The key feature of the representation is that it provides a semantic grounding for the symbols in a set of modeling equations by linking those symbols to an explicit representation of the underlying environmental scenario.

Keller, Richard M.↗

Hierarchical Modeling and Statistical Calibration for Photometric Redshifts

The cosmological exploitation of modern photometric galaxy surveys requires both accurate (unbiased) and precise (narrow) redshift probability distributions derived from broadband photometry. Existing methodologies do not meet these requirements. Standard template fitting delivers interpretable models and errors, but lacks the flexibility to learn inaccuracies in the observed photometry or spectral templates. Machine learning addresses those issues, but requires representative training data, and the resulting models and uncertainties cannot be interpreted in the context of a physical model or outside of the training data. We present a hierarchical modeling approach simultaneously addressing the issues of flexibility, interpretability, and generalization. It combines template fitting with flexible (machine-learning-like) models to correct the spectral templates, model their redshift distributions, and recalibrate the photometric observations. By optimizing the full posterior distribution of the model and solving for its (thousands of) parameters, one can perform a global statistical calibration of the data and the spectral energy distribution (SED) model. We apply this approach to the public Dark Energy Survey Science Verification data and show that it provides more accurate and compact redshift posterior distributions than existing methods, as well as insights into residual photometric and SED systematics. Here, the model is causal and makes predictions for future data (e.g., additional photometric bandpasses), and its internal parameters and components are interpretable. This approach does not formally require the training data to be complete or representative; in principle, it can even work in regimes in which few or no spectroscopic redshifts are available.

79 ASTRONOMY AND ASTROPHYSICS↗

An investigation of MAGSAT and complementary data emphasizing precambrian shields and adjacent areas of West Africa and South America

Some MAGSAT data and preliminary results emphasizing Africa, Latin America, and the Caribbean were presented to geologists in Jamaica, Guyana, Brazil, and Argentina who then expressed interest in collaborating on linking computer modeling of MAGSAT data to tectionic and metallogenic interpretation for South America and in making the final interpretation. Modeling continues with the preparation of a synopsis tectonic map of South America and the initiation of semi-final modeling of northern South America.

Hastings, D. A.↗

Sparsifying priors for Bayesian uncertainty quantification in model discovery

We propose a probabilistic model discovery method for identifying ordinary differential equations governing the dynamics of observed multivariate data. Our method is based on the sparse identification of nonlinear dynamics (SINDy) framework, where models are expressed as sparse linear combinations of pre-specified candidate functions. Promoting parsimony through sparsity leads to interpretable models that generalize to unknown data. Instead of targeting point estimates of the SINDy coefficients, we estimate these coefficients via sparse Bayesian inference. The resulting method, uncertainty quantification SINDy (UQ-SINDy), quantifies not only the uncertainty in the values of the SINDy coefficients due to observation errors and limited data, but also the probability of inclusion of each candidate function in the linear combination. UQ-SINDy promotes robustness against observation noise and limited data, interpretability (in terms of model selection and inclusion probabilities) and generalization capacity for out-of-sample forecast. Sparse inference for UQ-SINDy employs Markov chain Monte Carlo, and we explore two sparsifying priors: the spike and slab prior, and the regularized horseshoe prior. UQ-SINDy is shown to discover accurate models in the presence of noise and with orders-of-magnitude less data than current model discovery methods, thus providing a transformative method for real-world applications which have limited data.

97 MATHEMATICS AND COMPUTING↗

Neural Network Based Modeling and Analysis of LP Control Surface Allocation

This paper presents an approach to interpretive modeling of LP based control allocation in intelligent flight control. The emphasis is placed on a nonlinear interpretation of the LP allocation process as a static map to support analytical study of the resulting closed loop system, albeit in approximate form. The approach makes use of a bi-layer neural network to capture the essential functioning of the LP allocation process. It is further shown via Lyapunov based analysis that under certain relatively mild conditions the resulting closed loop system is stable. Some preliminary conclusions from a study at Ames are stated and directions for further research are given at the conclusion of the paper.

Langari, Reza↗

Search for high-mass dilepton resonances in p p collisions at s = 8 TeV with the ATLAS detector

Here, the ATLAS detector at the Large Hadron Collider is used to search for high-mass resonances decaying to dielectron or dimuon final states. Results are presented from an analysis of proton-proton (pp) collisions at a center-of-mass energy of 8 TeV corresponding to an integrated luminosity of 20.3 fb –1 in the dimuon channel. A narrow resonance with Standard Model Z couplings to fermions is excluded at 95% confidence level for masses less than 2.79 TeV in the dielectron channel, 2.53 TeV in the dimuon channel, and 2.90 TeV in the two channels combined. Limits on other model interpretations are also presented, including a grand-unification model based on the E 6 gauge group, Z* bosons, minimal Z' models, a spin-2 graviton excitation from Randall-Sundrum models, quantum black holes, and a minimal walking technicolor model with a composite Higgs boson.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Explainable Neural Architecture Search (XNAS)

Code for the paper Learning Interpretable Models Through Multi-Objective Neural Architecture Search by Zachariah Carmichael, Tim Moon, and Sam Ade Jacobs. Monumental advances in deep learning have led to unprecedented achievements across a multitude of domains. While the performance of deep neural networks is indubitable, the architectural design and interpretability of such models are nontrivial. Research has been introduced to automate the design of neural network architectures through neural architecture search (NAS). Recent progress has made these methods more pragmatic by exploiting distributed computation and novel optimization algorithms. However, there is little work in optimizing architectures for interpretability. To this end, we propose a multiobjective distributed NAS framework that optimizes for both task performance and introspection. We leverage the non-dominated sorting genetic algorithm (NSGA-II) and explainable AI (XAI) techniques to reward architectures that can be better comprehended by humans. The framework is evaluated on several image classification datasets. We demonstrate that jointly optimizing for introspection ability and task error leads to more disentangled architectures that perform within tolerable error.

Carmichael, ZachariahJ↗