SEARCH · Engineering Papers

Results for “explainability”

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 199 records · Page 11

All sources NASA NTRS DOE OSTI

DOE OSTI2018

Explaining low ℓ anomalies in the CMB power spectrum with resonant superstring excitations during inflation

We explore the possibility that both the suppression of the ℓ = 2 multipole moment of the power spectrum of cosmic microwave background temperature fluctuations and the possible dip for ℓ = 10 –30 can be explained as well as a possible new dip for ℓ ≈ 60 as the result of the resonant creation of sequential excitations of a fermionic (or bosonic) closed superstring that couples to the inflaton field. We consider a D = 26 closed bosonic string with one toroidal compact dimension as an illustration of how string excitations might imprint themselves on the CMB. We analyze the existence of successive momentum states, winding states or oscillations on the string as the source of the three possible dips in the power spectrum. Although the evidence of these dips are of marginal statistical significance, this might constitute the first observational evidence of successive superstring excitations in Nature.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2019

AWE-ML: Averaged Weights for Explainable Machine Learning v. 1.0

This software uses averaged weighted estimators to classify instances and then explicate the individual decision process in classification. The software has state of the art performance in accuracy, while maintaining full explainability. It does this by calculating hierarchically averaged probabilities for a large class of feature combinations and updating them ‘on-the-fly’. On the fly updating also allows the model to be accurate, relative to the total dataset, maintaining accuracy in real-time, without a need to retrain the classifier.

Agarwal, Sapan↗

DOE OSTI2021

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↗

DOE OSTI2024

Vistransformers Explained

The Vistransformers Explained library is a collection of python notebooks that demonstrate the internal mechanics and uses of visual-transformer (ViT) machine learning models. The code implements, with mild modifications, ViT models that have been made publicly available through publication and GitHub code. The value added by this code is in-depth explanations of the mathematics behind the sub-modules of the ViT models, including original figures. Additionally, the library contains the code necessary to implement and train the ViT models. The library does not include example training data for the models; instead, it would rely on users generating their own datasets. The code is based on the PyTorch python library. It does not include any files other than python scripts, modules, or notebooks.

Callis, Skylar↗

DOE OSTI2022

Understanding Predictability of Daily Southeast U.S. Precipitation Using Explainable Machine Learning

Abstract We investigate the predictability of the sign of daily southeastern U.S. (SEUS) precipitation anomalies associated with simultaneous predictors of large-scale climate variability using machine learning models. Models using index-based climate predictors and gridded fields of large-scale circulation as predictors are utilized. Logistic regression (LR) and fully connected neural networks using indices of climate phenomena as predictors produce neither accurate nor reliable predictions, indicating that the indices themselves are not good predictors. Using gridded fields as predictors, an LR and convolutional neural network (CNN) are more accurate than the index-based models. However, only the CNN can produce reliable predictions that can be used to identify forecasts of opportunity. Using explainable machine learning we identify which variables and grid points of the input fields are most relevant for confident and correct predictions in the CNN. Our results show that the local circulation is most important as represented by maximum relevance of 850-hPa geopotential heights and zonal winds to making skillful, high-probability predictions. Corresponding composite anomalies identify connections with El Niño–Southern Oscillation during winter and the Atlantic multidecadal oscillation and North Atlantic subtropical high during summer.

Pegion, Kathy↗

DOE OSTI2021

Explaining Globally Inhomogeneous Future Changes in Monsoons Using Simple Moist Energy Diagnostics

Here, this study examines the annual cycle of monsoon precipitation simulated by models from phase 6 of the Coupled Model Intercomparison Project (CMIP6), then uses moist energy diagnostics to explain globally inhomogeneous projected future changes. Rainy season characteristics are quantified using a consistent method across the globe. Model bias is shown to include rainy season onsets tens of days later than observed in some monsoon regions (India, Australia, and North America) and overly large summer precipitation in others (North America, South America, and southern Africa). Projected next-century changes include rainy season lengthening in the two largest Northern Hemisphere monsoon regions (South Asia and central Sahel) and shortening in the two largest Southern Hemisphere regions (South America and southern Africa). Changes in the North American and Australian monsoons are less coherent across models. To understand these changes, relative moist static energy (MSE) is defined as the difference between local and tropical-mean surface air MSE. Future changes in relative MSE in each region correlate well with onset and demise date changes. Furthermore, Southern Hemisphere regions projected to undergo rainy season shortening are spanned by an increasing equator-to-pole MSE gradient, suggesting their rainfall will be increasingly inhibited by fluxes of dry extratropical air; Northern Hemisphere regions with projected lengthening of rainy seasons undergo little change in equator-to-pole MSE gradient. Thus, although model biases raise questions as to the reliability of some projections, these results suggest that globally inhomogeneous future changes in monsoon timing may be understood through simple measures of surface air MSE.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2021

Human Readiness Levels Explained

The Human Readiness Level scale complements and supplements the existing technology readiness level scale to support comprehensive and systematic evaluation of human system aspects throughout a system’s life cycle. The objective is to ensure humans can use a fielded technology or system as intended to support mission operations safely and effectively. This article defines the nine human readiness levels in the scale, explains their meaning, and illustrates their application using a helmet-mounted display example.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2020

Potentially adaptive SARS-CoV-2 mutations discovered with novel spatiotemporal and explainable AI models

Abstract Background A mechanistic understanding of the spread of SARS-CoV-2 and diligent tracking of ongoing mutagenesis are of key importance to plan robust strategies for confining its transmission. Large numbers of available sequences and their dates of transmission provide an unprecedented opportunity to analyze evolutionary adaptation in novel ways. Addition of high-resolution structural information can reveal the functional basis of these processes at the molecular level. Integrated systems biology-directed analyses of these data layers afford valuable insights to build a global understanding of the COVID-19 pandemic. Results Here we identify globally distributed haplotypes from 15,789 SARS-CoV-2 genomes and model their success based on their duration, dispersal, and frequency in the host population. Our models identify mutations that are likely compensatory adaptive changes that allowed for rapid expansion of the virus. Functional predictions from structural analyses indicate that, contrary to previous reports, the Asp 614 Gly mutation in the spike glycoprotein (S) likely reduced transmission and the subsequent Pro 323 Leu mutation in the RNA-dependent RNA polymerase led to the precipitous spread of the virus. Our model also suggests that two mutations in the nsp13 helicase allowed for the adaptation of the virus to the Pacific Northwest of the USA. Finally, our explainable artificial intelligence algorithm identified a mutational hotspot in the sequence of S that also displays a signature of positive selection and may have implications for tissue or cell-specific expression of the virus. Conclusions These results provide valuable insights for the development of drugs and surveillance strategies to combat the current and future pandemics.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2021

Explainable Artificial Intelligence in Endocrinological Medical Research

Artificial Intelligence (AI) has been a part of the medical community for decades in the form of Clinical Decision Support Systems (CDSS) to aide physicians in diagnosis and categorization of patients (1). Recent years have seen a shift from expert-derived models to the integration of machine learning (ML) algorithms to drive the output of these AI systems due to the ability of ML models to more accurately make predictions by exploiting higher dimensional and often complex data. In many cases ML models gain their advantage in accuracy by capture complex and often non-linear relationships between features being used to make the prediction. However, the hype and excitement around these methods are tempered by the clinical utility of often black-box solutions driven by skepticism of results that are difficult for practitioners to not only interpret but explain to their patients (1-3). This skepticism is not unfounded as multiple examples of black-box solutions identifying incidental correlates as the key predictors have highlighted the potential bias in a training set, or reward system, that a ML model may exploit; for example a model discerning wolves from huskies based on snow in the background rather than features of the dogs (1, 4, 5).

Endocrinological, artificial intelligence, diabete↗

DOE OSTI2023

Explaining machine-learning models for gamma-ray detection and identification

As more complex predictive models are used for gamma-ray spectral analysis, methods are needed to probe and understand their predictions and behavior. Recent work has begun to bring the latest techniques from the field of Explainable Artificial Intelligence (XAI) into the applications of gamma-ray spectroscopy, including the introduction of gradient-based methods like saliency mapping and Gradient-weighted Class Activation Mapping (Grad-CAM), and black box methods like Local Interpretable Model-agnostic Explanations (LIME) and SHapley Additive exPlanations (SHAP). In addition, new sources of synthetic radiological data are becoming available, and these new data sets present opportunities to train models using more data than ever before. In this work, we use a neural network model trained on synthetic NaI(Tl) urban search data to compare some of these explanation methods and identify modifications that need to be applied to adapt the methods to gamma-ray spectral data. We find that the black box methods LIME and SHAP are especially accurate in their results, and recommend SHAP since it requires little hyperparameter tuning. We also propose and demonstrate a technique for generating counterfactual explanations using orthogonal projections of LIME and SHAP explanations.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2024

Data and Scripts Associated with the Manuscript “Water Column Respiration in the Yakima River Basin is Explained by Temperature, Nutrients and Suspended Solids”

This data package is associated with the publication “Water Column Respiration in the Yakima River Basin is Explained by Temperature, Nutrients and Suspended Solids” published in EGU Biogeochemistry (Laan et al. 2025). In this research, water column respiration (ERwc) data, surface water chemistry data, organic matter (OM) chemistry data, and publicly available geospatial data were used in analysis to evaluate the variability in ERwc at 47 sites across the Yakima River basin in Washington, USA. In addition to this readme, this data package also includes a file-level metadata (FLMD) file that describes each file and a data dictionary (DD) that describes all column/row headers and variable definitions. The data package includes the data inputs, and outputs, and R scripts to reproduce all the analyses performed in the manuscript and create manuscript figures. The data package is comprised of three main folders (Code, Data, and Figures). The Code folder is comprised of four scripts and three analysis-specific subfolders that contain the R scripts to perform the analyses described in the publication and create publication figures. The Data folder is comprised of two “.csv” files and four subfolders that contain data input and output files. The Published_Data folder contains a readme that directs the user to download the appropriate files and add to this folder when using scripts. The Figures folder includes figures from the manuscript in “.pdf” and “.png” formats and a folder with intermediate figure files. This data package is associated with a GitHub repository which can be found at https://github.com/river-corridors-sfa/rcsfa-RC2-SPS-ERwc. We acknowledge the Yakama Nation as owners and caretakers of the lands where we collected some of these data. We thank the Confederated Tribes and Bands of the Yakama Nation Tribal Council and Yakama Nation Fisheries for working with us to facilitate sample collection and optimization of data usage according to their values and worldview.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2020

Methods for Explainable Artificial Intelligence

We explored ways of quantifying information in a neural network. This can be used to determine the right size of a network or to infer the way in which a network is processing information. The first year and a half was somewhat exploratory while the last half of the project focused on approaches that seemed to show the most promise. The introduce a new way of computing explainable artificial intelligence (XAI) saliency maps that is several orders of magnitude faster than methods with similar fidelity We call it FastCAM. The method works be combining a Class Activation Map (CAM) method such as GradCAM with a forward activation map computed with a statistic we call SMOE Scale. The addition of the forward activation maps to CAM methods seems to always improve their fidelity. At the same time, computational overhead is not increased by very much. While Gradients with SmoothGrad scores better on some fidelity measures, it is overall not as good and requires more than 1500 times to compute. We demonstrate two completed applications of FastCAM on tasks outside of the LDRD at LLNL. The source code for FastCAM is currently being implanted into Captum, the official XAI toolkit for the popular deep learning toolkit PyTorch. The LDRD currently has 13 publications released to the public. 10 of them are journal length.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

Development of Explainable, Knowledge-Guided AI Models to Enhance the E3SM Land Model Development and Uncertainty Quantification

Focal Area(s): (2)Predictive modeling using AI techniques and AI-derived model components; use of AI and other tools to design a prediction system comprising of a hierarchy of models. (3) Insight gleaned from complex data (both observed and simulated) using AI, big data analytics, and other advanced methods, including explainable AI and physics- or knowledge- guided AI. Science Challenge: The Energy Exascale Earth System Model (E3SM) is a fully coupled, state-of-the-science Earth system model that uses code optimized for DOE's advanced computers to address the most critical scientific questions facing our nation and society (Golaz et al., 2019). The E3SM Land model (ELM) is designed to understand how the changes in terrestrial land surfaces will interact with other Earth system components and has been used to understand hydrologic cycles, biogeophysics, and ecosystem dynamics. In spite of great successes, the ELM has several known issues that restrain rapid improvements. For example, the ELM uses equilibrium models to simulate dynamic land-climate interactions and it requires long model spin-up time to identify suitable initial conditions for transient simulations. The ELM lacks built-in uncertainty mechanisms that can improve the robustness of model predictions. The ELM is a holistic, deterministic model system with a rigid design, and in many situations, it is hard to modify the ELM system to incorporate new theory/hypothesis and new data across scales to address emerging science problems (such as predicting the impacts of water cycle extremes). In addition, The ELM is technically optimized for traditional CPU-centric computers and it cannot fully utilize the current and incoming leadership computers for model simulations and uncertainty quantification (UQ). The success of artificial intelligence (AI) has inspired scientists to use AI models to discover intrinsic features from simulation data (Chattopadhyay et al., 2020) and observational data (Reichstein et al., 2019) to gain further process understanding of Earth science problems. However, autonomous AI model training through deep learning usually requires a huge amount of annotated data. To overcome the limitations from the data and computing resources, knowledge-guided AI models are necessary where human-knowledge is ingested in model construction (Banino et al., 2018) and training process (Silver et al., 2016) for efficient learning. Herein, we present a new way that leverages the process understanding from the ELM to guide AI model development for the ELM enhancement and UQ. We hope this study can inspire further Earth and environmental system model developments and transformations.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2023

Identifying and Explaining Anomalous Activity in Surveillance Video with Compression Algorithms

The primary purpose of this document is to outline the progress made on the LDRD titled “Identifying and Explaining Anomalous Activity in Surveillance Video with Compression Algorithms” in FY22 and FY23. In this LDRD, we explored the usage of compression-based analytics to identify anomalous activity in video. We developed a novel algorithm, Spatio-Temporal N-Gram PPM (STNG PPM) that accounts for spatially and temporally aware anomalies in video. We extracted features using motions vectors from video as well as operating on the raw features. STNG PPM is comparable to many deep learning approaches but does not require specialized hardware (GPUs) to run efficiently. We also examine the evaluation metrics and propose novel measures addressing faults in the current evaluation measures.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Explainable and Differentiable Reinforcement Learning for Multi-objective Optimization in Particle Accelerators

Operating particle accelerators involves optimizing multiple goals simultaneously, which can be challenging due to trade-offs among objectives. While evolutionary algorithms like the genetic algorithm (GA) have been used for various Multi-Objective Optimization (MOO) tasks, they are not inherently suited for complex control problems. This talk highlights two variations of Reinforcement Learning (RL) for concurrently optimizing heat load and trip rates at the Continuous Electron Beam Accelerator Facility (CEBAF). The problem involves strict constraints on individual states, actions, and overall energy requirements of the beam. First, this talk highlights how differentiability can be harnessed through a Deep Differentiable Reinforcement Learning (DDRL) approach to address MOO issues within particle accelerators. We examine the DDRL method alongside Model Free Reinforcement Learning (MFRL), GA, and Bayesian Optimization (BO). The performance of these methods is assessed by generating a Pareto-front for two objectives. Our findings indicate that DDRL excels in handling high-dimensional problems more effectively than MFRL, BO, and GA. Next, we will show integration of explainable physics-based constraints into RL algorithms to enhance trans- parency and trust in decision-making processes by enabling users to verify that agents adhere to established physical principles. This surrogate function can be modeled using neural networks or sparse dictionary mod- els. By examining the mathematical form of the learned constraint function, we are able to confirm the agent has learned to use the established physics of each environment provided but the surrogate model. In addi- tion, we find that the introduction of a mathematical functional dictionary based surrogate model enables our reinforcement learning algorithms to reliably converge for difficult high-dimensional accelerator controls environments.

Rajput, Kishansingh [Thomas Jefferson National Acc↗

DOE OSTI2023

Data to Accompany: PM2.5 is insufficient to explain personal PAH exposure

Fine particulate matter (PM2.5) air quality index (AQI) data from outdoor stationary monitors and Hazard Mapping System (HMS) smoke density data from satellites are often used as proxies for personal chemical exposure. Silicone wristbands can quantify more individualized exposure data than stationary air monitors or smoke satellites. However, it is not understood how these proxy measurements compare to chemical data measured from wristbands. We hypothesized that predictive models for personal chemical exposure would be significantly improved by expanding beyond stationary PM2.5 AQI data or satellite HMS data to also include environmental and behavioral information. In Eugene, Oregon, participants wore daily wristbands, carried a phone that recorded locations, and answered daily questionnaires for a seven-day period in multiple seasons. We gathered publicly available daily PM2.5 AQI data and HMS data. We analyzed wristbands for 94 organic chemicals, including 53 polycyclic aromatic hydrocarbons (PAHs). Wristband chemical detections and concentrations, behavioral variables (e.g., time spent indoors), and environmental conditions (e.g., PM2.5 AQI) significantly differed between seasons. Machine learning models were fit to predict personal chemical exposure using PM2.5 AQI only, HMS only, and a multivariate feature set including PM2.5 AQI, HMS, and other environmental and behavioral information. On average, the multivariate models increased predictive accuracy by approximately 70% compared to either the AQI model or the HMS model for all chemicals modeled. This study provides evidence that PM2.5 AQI data alone or HMS data alone is insufficient to explain personal chemical exposures. Our results identify additional key predictors of personal chemical exposure.

Bramer, Lisa M↗

DOE OSTI2019

A Dynamic Optimality Principle for Water Use Strategies Explains Isohydric to Anisohydric Plant Responses to Drought

Optimality principles that underlie models of stomatal kinetics require identifying and formulating the gain and the costs involved in opening stomata. While the gain has been linked to larger carbon acquisition, there is still a debate as to the costs that limit stomatal opening. This work presents an Euler-Lagrange framework that accommodates water use strategy and various costs through the formulation of constraints. The reduction in plant hydraulic conductance due to cavitation is added as a new constraint above and beyond the soil hydrological balance and is analyzed for three different types of whole-plant vulnerability curves. Model results show that differences in vulnerability curves alone lead to relatively iso- and aniso-hydric stomatal behavior. Moreover, this framework explains how the presence of competition (biotic or abiotic) for water alters stomatal response to declining soil water content. This contribution corroborates previous research that predicts that a plant's environment (e.g., competition, soil processes) significantly affects its response to drought and supplies the required mathematical machinery to represent this complexity. The method adopted here disentangles cause and effect of the opening and closure of stomata and complements recent mechanistic models of stomatal response to drought.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2023

Seasonal decline in leaf photosynthesis in perennial switchgrass explained by sink limitations and water deficit

Leaf photosynthesis of perennial grasses usually decreases markedly from early to late summer, even when the canopy remains green and environmental conditions are favorable for photosynthesis. Understanding the physiological basis of this photosynthetic decline reveals the potential for yield improvement. We tested the association of seasonal photosynthetic decline in switchgrass (Panicum virgatum L.) with water availability by comparing plants experiencing ambient rainfall with plants in a rainfall exclusion experiment in Michigan, USA. For switchgrass exposed to ambient rainfall, daily net CO 2 assimilation ( A' net ) declined from 0.9 mol CO 2 m -2 day -1 in early summer to 0.43 mol CO 2 m -2 day -1 in late summer (53% reduction; P<0.0001). Under rainfall exclusion shelters, soil water content was 73% lower and A' net was 12% and 26% lower in July and September, respectively, compared to those of the rainfed plants. Despite these differences, the seasonal photosynthetic decline was similar in the season-long rainfall exclusion compared to the rainfed plants; A' net in switchgrass under the shelters declined from 0.85 mol CO 2 m -2 day -1 in early summer to 0.39 mol CO 2 m -2 day -1 (54% reduction; P<0.0001) in late summer. These results suggest that while water deficit limited A' net late in the season, abundant late-season rainfalls were not enough to restore A' net in the rainfed plants to early-summer values suggesting water deficit was not the sole driver of the decline. Alongside change in photosynthesis, starch in the rhizomes increased 4-fold (P<0.0001) and stabilized when leaf photosynthesis reached constant low values. Additionally, water limitation under shelters had no negative effects on the timing of rhizome starch accumulation, and rhizome starch content increased ~6-fold. These results showed that rhizomes also affect leaf photosynthesis during the growing season. Towards the end of the growing season, when vegetative growth is completed and rhizome reserves are filled, diminishing rhizome sink activity likely explained the observed photosynthetic declines in plants under both ambient and reduced water availability.

59 BASIC BIOLOGICAL SCIENCES↗