Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “post-hoc”

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 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

NuGraph2 with explainability: post-hoc explanations for geometric neural network predictions

With the growing popularity of artificial intelligence (AI) used for scientific applications, the ability of attribute a result to a reasoning process from the network is in high demand for robust scientific generalizations to hold. In this work we aim to motivate the need for and demonstrate the use of post-hoc explainability methods when applied to AI methods used in scientific applications. To this end, we introduce explainability add-ons to the existing graph neural network (GNN) for neutrino tagging, NuGraph2. The explanations take the form of a suite of techniques examining the output of the network (node classifications) and the edge connections between them, and probing of the latent space using novel general-purpose tools applied to this network. We show how none of these methods are singularly sufficient to show network ‘understanding’, but together can give insights into the processes used in classification. While these methods are tested on the NuGraph2 application, they can be applied to a broad range of networks, not limited to GNNs. The code for this work is publicly available on GitHub at https://github.com/voetberg/XNuGraph.

Voetberg, Margaret [Fermilab] (ORCID:0009000527154↗
DOE OSTI2025

Post-hoc reweighting of hadron production in the Lund string model

We present a method for reweighting flavor selection in the Lund string fragmentation model. This is the process of calculating and applying event weights enabling fast and exact variation of hadronization parameters on pre-generated event samples. The procedure is post hoc, requiring only a small amount of additional information stored per event, and allowing for efficient estimation of hadronization uncertainties without repeated simulation. Weight expressions are derived from the hadronization algorithm itself, and validated against direct simulation for a wide range of observables and parameter shifts. The hadronization algorithm can be viewed as a hierarchical Markov process with stochastic rejections, a structure common to many complex simulations outside of high-energy physics. This perspective makes the method modular, extensible, and potentially transferable to other domains. We demonstrate the approach in Pythia, including both coverage considerations and timing benefits. For the purpose of this paper, our goal is to develop and demonstrate the the formalism, and we therefore exclude several model variations for baryon production (popcorn model, junction production) needed for proton collisions. These will be the topic of a future paper.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2026

Toward the “platinum standard” of quantum chemistry on quantum computers: Perturbative quadruple corrections in unitary coupled cluster theory

We propose a non-iterative, post-hoc correction to the unitary coupled cluster theory with the single, double, and triple excitations (UCCSDT) Ansatz, which considers the leading-order effects of neglected quadruple excitations. We present two ways to derive this correction, henceforth referred to as [Q-6], which leads to an improvement in the correlation energy shown to be truncated to sixth-order in many-body perturbation theory. Furthermore, a comparison between the UCC-based [Q-6] correction proposed in this work and analogous, “platinum standard” quadruple corrections proposed in conventional coupled cluster theory recognizes that [Q-6] is distinct from prior corrections since it is constructed entirely from internally connected components. Although trotterized (t) and full operator variants of UCCSDT exhibit errors in scans of small molecule potential energy surfaces that routinely exceed 1.6 mH, we find that t/UCCSDT[Q-6] is, nevertheless, able to achieve chemical accuracy as measured by the mean unsigned error.

Correlation energy↗
DOE OSTI2025

Where do the fish go in winter? A year of observing seasonal changes in Sequim Bay’s nearshore fish community

The use of environmental DNA (eDNA) sampling has been proposed as a complementary method to monitor fish species in marine environments. eDNA offers a non-invasive, cost-effective, and scalable way to detect aquatic species. It is appealing in environments where traditional methods are limited by access or visibility, especially in complex or sensitive habitats such as tidal channels and other marine energy sites. Before eDNA can be fully relied upon, we must verify its accuracy against established methods, like underwater photography. In this study, we collected eDNA samples and concurrently deployed a 360-degree camera near the floating dock of PNNL-Sequim in the tidal channel of Sequim Bay once a month for twelve consecutive months. During deployments, the camera remained on the seafloor for several hours overlapping slack tide and captured time-lapse photographs at ten second intervals. Counts and identifications of fish species observed in the images were used to calculate monthly Shannon diversity and Pielou evenness indices. These values were compared across the months using a Kruskal-Wallis test paired with a Conover-Iman post-hoc test. Cliff’s Delta was also calculated to quantify the effect size of the monthly differences. No fish were observed from December through April, likely due to seasonal behavior changes within the local fish community. Fish returned to the shoreline in May, with the greatest diversity and evenness recorded in September. These findings reveal substantial seasonal variation in nearshore fish communities. Many monthly comparisons were found to contain statistically significant differences within the diversity and evenness, and even more were found to have large effect sizes, signifying large ecological changes throughout the seasons. The absence of fish observed during the winter months is a key outcome of this survey that will hopefully be reflected in the eDNA results still to come, which would help validate the eDNA approach.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Explainable Machine Learning for Functional Data

Black-box machine learning models are recognized as useful tools for prediction applications, but the algorithmic complexity of some models causes interpretation challenges. Explainability methods have been proposed to provide insight into these models, but there is little research focused on supervised modeling with functional data inputs. We argue that, especially in applications of high consequence, it is important to explicitly model the functional dependence in a black-box analysis to not obscure or misrepresent patterns in explanations. As such, we propose the V ariable importance E xplainable E lastic S hape A nalysis (VEESA) pipeline for training supervised machine learning models with functional inputs. The pipeline is an analysis process that includes the data preprocessing, modeling, and post-hoc explanations. The preprocessing is done using elastic functional principal components analysis, which accounts for vertical and horizontal variability in functional data and, ultimately, allows for explanations in the original data space that identify the important functional variability without bias due to correlated variables. Here, we demonstrate the pipeline on two high-consequence applications: explosives classification for national security and inkjet printer identification in forensic science. The applications exhibit the VEESA pipeline’s ability to provide an understanding of the characteristics of the functional data useful for prediction. Code for implementing the pipeline is available in the veesa R package (and supplemental python code).

Elastic Shape Analysis↗
NASA NTRS1993

Comparison of total body water estimates from O-18 and bioelectrical response prediction equations

Identification of an indirect, rapid means to measure total body water (TBW) during space flight may aid in quantifying hydration status and assist in countermeasure development. Bioelectrical response testing and hydrostatic weighing were performed on 27 subjects who ingested O-18, a naturally occurring isotope of oxygen, to measure true TBW. TBW estimates from three bioelectrical response prediction equations and fat-free mass (FFM) were compared to TBW measured from O-18. A repeated measures MANOVA with post-hoc Dunnett's Test indicated a significant (p less than 0.05) difference between TBW estimates from two of the three bioelectrical response prediction equations and O-18. TBW estimates from FFM and the Kushner & Schoeller (1986) equation yielded results that were similar to those given by O-18. Strong correlations existed between each prediction method and O-18; however, standard errors, identified through regression analyses, were higher for the bioelectrical response prediction equations compared to those derived from FFM. These findings suggest (1) the Kushner & Schoeller (1986) equation may provide a valid measure of TBW, (2) other TBW prediction equations need to be identified that have variability similar to that of FFM, and (3) bioelectrical estimates of TBW may prove valuable in quantifying hydration status during space flight.

Barrows, Linda H.↗
NASA NTRS1996

Synaptic Plasticity In Mammalian Gravity Sensors: Preliminary Results From SLS-2

Sensory conflict is the prevalent theoretical explanation for space adaptation syndrome. This ultrastructural study tests the hypothesis that peripheral gravity sensors (maculae) play a role. Results were obtained from the medial part of utricular maculae of adult rats exposed to microgravity for 14 days, and from controls. Means and statistical significance of synapse counts were calculated using SUPERANOVA(Trademark) and Scheffe's procedure for post-hoc comparisons. Preliminary findings are from 2 sets of 100 serial sections for each dataset. Synapses were doubled numerically in type II hair cells of utricular maculae collected on day 13 inflight compared to controls (11.4 +/- 7.1 vs. 5.3 +/- 3.8; p < 0.0001). Flight mean synaptic number declined rapidly postflight and became comparable to means of controls. Synapses also increased numerically in type I cells inflight (2.4 +/- 1.6 vs. 1.7 +/- 1.0; p < 0.0341). Postflight there were no significant differences in counts. Results concerning shifts in ribbon type and distribution are also largely replicating previous findings from flight studies. Results indicate that mammalian maculae are adaptive endorgans that retain the property of synaptic plasticity into the adult stage. Macular plasticity has clinical implications for balance disorders of peripheral origin.

Ross, M. D.↗
NASA NTRS2002

Skeletal Phenotype of Transgenic Mice Expressing the Beta1 Integrin Cytoplasmic Tail In Osteoblasts

To define the physiologic role of beta1 integrin in bone formation and mechanical loading, transgenic mice were generated by expressing the cytoplasmic tall and transmembrane domain of Beta1 integrin under the control of the osteocalcin promoter. In cultured cells, this truncated fragment of Beta1 can act as a dominant negative. Previously, the matrix of calvariae was shown to be abnormal in transgenic (TG) compared to wildtype (WT) mice. In this study, we analyzed appendicular bone in TG and WT, male and female mice at 14, 35, 63, 90 and 365 days old (n=8-12/gp). To assess beta1 integrin function in mechanical loading, a pilot study using hindlimb unloading by tail suspension was performed. 35d old TG and WT females were hindlimb unloaded for 4 wks (n=3-5). Body mass, bone mineral content, histomorphometric (distal femur) and biomechanical parameters were analyzed. Statistical significance (P less than.05) was defined by ANOVA using the Tukey-Kramer post-hoc test. We confirmed transgene expression by immunoprecipitating then immunoblotting bone lysates using an antibody against the beta1 tail. Body masses of TG mice at 63, 90 and 365d old were greater (16-25%) than WT. Some TG female mice at 365d appeared obese; mean abdominal fat mass was 415% greater in TG than WT mice. Tibiae were longer (5-7%) in TG than WT mice at 63 and 90d. Tibial mineral mass of 35d males was 7% lower in TG than WT mice, but at 63d was 21% higher. The % osteoblast surface in 35d TG mice was 20% higher than WT, and at 63d was 17% lower, while % osteoclast surface did not differ. In 365d mice, cancellous bone volume (125%) and endocortical mineral apposition rate (40%) were greater in TG than WT males but not females. In WT mice, hindlimb unloading caused a reduction in mineral mass of tibiae (-20%) and lumbar vertebrae (-22%) relative to normally loaded controls. Surprisingly, hindlimb unloading also caused a relative reduction (-13%) in humerus mass. The effects of hindlimb unloading on tibia and humerus mass were less obvious in TG than in WT mice. Since hindlimb unloading caused skeletal changes in both loaded and unloaded bones, systemic changes may contribute to bone responses observed using this animal model. In conclusion, transgene expression resulted in marked metabolic changes during growth and in the aged female. Our results demonstrate that expression of the Beta1 integrin cytoplasmic tail in vivo causes gender- and age-specific changes in select morphometric parameters, bone length, and bone mass.

Globus, R. K.↗
NASA NTRS2000

Augmentation of blood circulation to the fingers by warming distant body areas

Future activities in space will require greater periods of time in extreme environments in which the body periphery will be vulnerable to chilling. Maintaining the hands and fingers in comfortable conditions enhances finger flexibility and dexterity, and thus effects better work performance. We have evaluated the efficacy of promoting heat transfer and release by the extremities by increasing the blood flow to the periphery from more distant parts of the body. The experimental garment paradigm developed by the investigators was used to manipulate the temperature of different body areas. Six subjects, two females and four males, were evaluated in a stage-1 baseline condition, with the inlet temperature of the circulating water in the liquid cooling/warming garment (LCWG) at 33 degrees C. At stage 2 the total LCWG water inlet temperature was cooled to 8 degrees C, and at stage 3 the inlet water temperature in specific segments of the LCWG was warmed (according to protocol) to 45 degrees C, while the inlet temperature in the rest of the LCWG was maintained at 8 degrees C. The following four body-area-warming conditions were studied in separate sessions: (1) head, (2) upper torso/arm, (3) upper torso/arm/head, and (4) legs/feet. Skin temperature, heat flux and blood perfusion of the fingers, and subjective perception of thermal sensations and overall physical comfort were assessed. Finger temperature (T(fing)) analyses showed a statistically significant condition x stage interaction. Post-hoc comparisons (T(fing)) indicated that at stage 3, the upper torso/arm/head warming condition was significantly different from the head, upper torso/arm and legs/feet conditions, showing an increase in T(fing). There was a significant increase in blood perfusion in the fingers at stage 3 in all conditions. Subjective perception of hand warmth, and overall physical comfort level significantly increased in the stage 3 upper torso/arm/head condition. The findings indicate that physiological methods to enhance heat transfer by the blood to the periphery within protective clothing provide an additional tool for increasing total and local human comfort in extreme environments.

NASA Discipline Life Sciences Technologies↗