Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Performance Rating Method”

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 811 records · Page 45
All sourcesNASA NTRSDOE OSTI
DOE OSTI2014

Nuclear resonant scattering measurements on {sup 57}Fe by multichannel scaling with a 64-pixel silicon avalanche photodiode linear-array detector

We developed a silicon avalanche photodiode (Si-APD) linear-array detector for use in nuclear resonant scattering experiments using synchrotron X-rays. The Si-APD linear array consists of 64 pixels (pixel size: 100 × 200 μm{sup 2}) with a pixel pitch of 150 μm and depletion depth of 10 μm. An ultrafast frontend circuit allows the X-ray detector to obtain a high output rate of >10{sup 7} cps per pixel. High-performance integrated circuits achieve multichannel scaling over 1024 continuous time bins with a 1 ns resolution for each pixel without dead time. The multichannel scaling method enabled us to record a time spectrum of the 14.4 keV nuclear radiation at each pixel with a time resolution of 1.4 ns (FWHM). This method was successfully applied to nuclear forward scattering and nuclear small-angle scattering on {sup 57}Fe.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2014

Characterization of the 80-mm diameter Hamamatsu PMTs for the KM3NeT project

The optical module designed for the KM3NeT project consists of 31 photomultipliers of 3-inch diameter housed into a 17-inch diameter glass sphere. A proposed photomultiplier was the R12199-02 Hamamatsu 80-mm diameter. 203 of such PMTs have been delivered from Hamamatsu and tested by the KM3NeT groups of NIKHEF-Amsterdam, ECAP-Erlangen and INFN-Catania. Tests have been performed to measure the main parameters, such as gain, transit time spread, dark pulses rate, fraction of spurious pulses, quantum efficiency and effective photocathode size. The main results matched with the requirements of the project. Methods and results are presented in this report.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2015

The effect of intersystem crossings in N({sup 2}D) + H{sub 2} collisions

The transitions between quartet and doublet states of the NH{sub 2} molecule are studied for the first time, allowing the evaluation of the N({sup 4}S) + H{sub 2} reactive channel. High level ab initio calculations of the spin-orbit coupling are performed over the whole configurational space of the NH{sub 2} molecule and fitted to a proposed analytic form. Quasiclassical trajectories coupled with the surface hopping method are employed to calculate reaction cross section and rate constants. The reaction is largely affected by the initial rovibrational states of H{sub 2}, while the formation of long-lived complexes enhances the reaction probability.

74 ATOMIC AND MOLECULAR PHYSICS↗
DOE OSTI2016

Ratcheted electrophoresis of Brownian particles

The realization of nanoscale machines requires efficient methods by which to rectify unbiased perturbations to perform useful functions in the presence of significant thermal noise. The performance of such Brownian motors often depends sensitively on their operating conditions—in particular, on the relative rates of diffusive and deterministic motions. In this letter, we present a type of Brownian motor that uses contact charge electrophoresis of a colloidal particle within a ratcheted channel to achieve directed transport or perform useful work against an applied load. We analyze the stochastic dynamics of this model ratchet to show that it functions under any operating condition—even in the limit of strong thermal noise and in contrast to existing ratchets. The theoretical results presented here suggest that ratcheted electrophoresis could provide a basis for electrochemically powered, nanoscale machines capable of transport and actuation of nanoscale components.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2016

Thermal relaxation of molecular oxygen in collisions with nitrogen atoms

Investigation of O{sub 2}–N collisions is performed by means of the quasi-classical trajectory method on the two lowest ab initio potential energy surfaces at temperatures relevant to hypersonic flows. A complete set of bound–bound and bound–free transition rates is obtained for each precollisional rovibrational state. Special attention is paid to the vibrational and rotational relaxations of oxygen as a result of chemically non-reactive interaction with nitrogen atoms. The vibrational relaxation of oxygen partially occurs via the formation of an intermediate NO{sub 2} complex. The efficient energy randomization results in rapid vibrational relaxation at low temperatures, compared to other molecular systems with a purely repulsive potential. The vibrational relaxation time, computed by means of master equation studies, is nearly an order of magnitude lower than the relaxation time in N{sub 2}–O collisions. The rotational nonequilibrium starts to play a significant effect at translational temperatures above 8000 K. The present work provides convenient relations for the vibrational and rotational relaxation times as well as for the quasi-steady dissociation rate coefficient and thus fills a gap in data due to a lack of experimental measurements for this system.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Accelerating FEM-Based Corrosion Predictions Using Machine Learning

Atmospheric corrosion of metallic parts is a widespread materials degradation phenomena that is challenging to predict given its dependence on many factors (e.g. environmental, physiochemical, and part geometry). For materials with long expected service lives, accurately predicting the degree to which corrosion will degrade part performance is especially difficult due to the stochastic nature of corrosion damage spread across years or decades of service. The Finite Element Method (FEM) is a computational technique capable of providing accurate estimates of corrosion rate by numerically solving complex differential Eqs. characterizing this phenomena. Nevertheless, given the iterative nature of FEM and the computational expense required to solve these complex equations, FEM is ill-equipped for an efficient exploration of the design space to identify factors that accelerate or deter corrosion, despite its accuracy. In this work, a machine learning based surrogate model capable of providing accurate predictions of corrosion with significant computational savings is introduced. Specifically, this work leverages AdaBoosted Decision trees to provide an accurate estimate of corrosion current per width given different values of temperature, water layer thickness, molarity of the solution, and the length of the cathode for a galvanic couple of aluminum and stainless steel.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Noble Gas Management: SBMOF 1 vs. NUCON Carbon

Pacific Northwest National Laboratory (PNNL), in collaboration with Flibe Energy, Inc., demonstrated the fabrication of metal organic frameworks (MOFs) into engineered beads to manage off-gases released from molten salt reactors. Among the many MOF options, SBMOF 1 was selected as an ideal sorbent to manage off-gases released from a molten salt reactor because its pore size is optimal. The SBMOF 1 was synthesized and fabricated into engineered particles using three different methods, with and without polymers. All the engineered particles were exposed to Co-60 radiation (1,000 kGy) to demonstrate their radiation stability. Single column breakthrough experiments were performed at different temperatures, activation conditions, flow rates, and off-gas compositions to optimize the Xe capacity at the breakthrough point and at saturation. The Xe loading in SBMOF 1 was compared with that of standard solid sorbent (Nucon carbon) under identical conditions. Based on the single column experiments, SBMOF 1 outperforms the Nucon carbon at all the temperatures and reduces the mass (to as little as 55%) of sorbent needed to achieve the same performance as carbon. The bed volume is also reduced, down to 48% of that of Nucon carbon, when the MOF is used. These results clearly demonstrate improvement in the amount of adsorbent needed and show a reduction in bed volume compared to traditional sorbent material.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2017

Three-flavour neutrino oscillations with MINOS and CHIPS

MINOS was a long-baseline neutrino oscillation experiment comprising two functionally identical detectors that observed Fermilab's NuMI neutrino beam in its low energy tune, at distances of 1 km and 735 km. When NuMI switched to a higher-power medium energy tune in 2012, the MINOS detectors continued to operate as MINOS+. Since its commissioning in 2003, the MINOS Far Detector has also been able to detect atmospheric neutrinos. Atmospheric neutrino oscillations are sensitive to the mass splitting m2 32 and mixing angle 23, and are also subject to the Matter E ect as the neutrinos pass through the Earth, which a ects neutrinos and antineutrinos di erently in a way that depends upon the mass hierarchy. This thesis presents the rst atmospheric neutrino analysis using data from the MINOS+ era, and the rst dedicated MINOS atmospheric neutrino analysis to use a full three- avour mixing model. It includes 10.79 kiloton years of new data and encompasses almost an entire period of the 11-year solar cycle, from 2003 to 2014. The CHIPS experiment aims to reduce construction costs of large water Cherenkov detectors to $200-300k per kiloton by submerging detectors with a lightweight structure in bodies of water on the surface of the Earth. Such detectors could reach masses of 1 Mton and would assist with the search for CP violation in the neutrino sector by measuring the rate of e appearance in a beam. A detailed reconstruction framework for CHIPS has been developed, incorporating a novel method based on the timing of PMT hits. This framework has been used to study the performance of di erent designs for a 10 kiloton CHIPS R&D module, and to demonstrate that e events can be identi ed in a sparsely-instrumented detector with a 6% coverage of 3" PMTs.

Perch, Andrew John↗
DOE OSTI2018

Self-optimized construction of transition rate matrices from accelerated atomistic simulations with Bayesian uncertainty quantification

Here, a massively parallel method to build large transition rate matrices from temperature-accelerated molecular dynamics trajectories is presented. Bayesian Markov model analysis is used to estimate the expected residence time in the known state space, providing crucial uncertainty quantification for higher-scale simulation schemes such as kinetic Monte Carlo or cluster dynamics. The estimators are additionally used to optimize where exploration is performed and the degree of temperature acceleration on the fly, giving an autonomous, optimal procedure to explore the state space of complex systems. The method is tested against exactly solvable models and used to explore the dynamics of C15 interstitial defects in iron. Our uncertainty quantification scheme allows for accurate modeling of the evolution of these defects over timescales of several seconds.

36 MATERIALS SCIENCE↗
DOE OSTI2016

Long-Term Outcomes Among Patients Who Achieve Complete or Near-Complete Responses After the Induction Phase of Bladder-Preserving Combined-Modality Therapy for Muscle-Invasive Bladder Cancer: A Pooled Analysis of NRG Oncology/RTOG 9906 and 0233

Purpose: To investigate the differences in outcomes among patients with muscle-invasive bladder cancer on NRG Oncology Radiation Therapy Oncology Group protocols 9906 and 0233 who achieved complete response and near-complete response after induction chemoradiation and then completed bladder-preserving therapy with chemoradiation therapy (chemo-RT) to full dose (60-64 Gy). Patients and Methods: A pooled analysis was performed on 119 eligible patients with muscle-invasive bladder cancer enrolled on NRG Oncology Radiation Therapy Oncology Group trials 9906 and 0233, who were classified as having a complete (T0) or near-complete (Ta or Tis) response after induction chemo-RT and completed consolidation with a total RT dose of at least 60 Gy. Bladder recurrence, salvage cystectomy rates, and disease-specific survival were estimated by the cumulative incidence method and bladder-intact and overall survivals by the Kaplan-Meier method. Results: Among the 119 eligible patients, 101 (85%) achieved T0, and 18 (15%) achieved Ta or Tis after induction chemo-RT and proceeded to consolidation. After a median follow-up of 5.9 years, 36 of 101 T0 patients (36%) versus 5 of 18 Ta or Tis patients (28%) experienced bladder recurrence (P=.52). Thirteen patients among complete responders eventually required late salvage cystectomy for tumor recurrence, compared with 1 patient among near-complete responders (P=.63). Disease-specific, bladder-intact, and overall survivals were not significantly different between T0 and Ta/Tis cases. Conclusions: The bladder recurrence and salvage cystectomy rates of the complete and the near-complete responders were similar. Therefore it is reasonable to recommend that patients with Ta or Tis after induction chemo-RT continue with bladder-sparing therapy with consolidation chemo-RT to full dose (60-64 Gy).

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2020

Wearable radio-frequency sensing of respiratory rate, respiratory volume, and heart rate

Abstract Many health diagnostic systems demand noninvasive sensing of respiratory rate, respiratory volume, and heart rate with high user comfort. Previous methods often require multiple sensors, including skin-touch electrodes, tension belts, or nearby off-the-body readers, and hence are uncomfortable or inconvenient. This paper presents an over-clothing wearable radio-frequency sensor study, conducted on 20 healthy participants (14 females) performing voluntary breathing exercises in various postures. Two prototype sensors were placed on the participants, one close to the heart and the other below the xiphoid process to couple to the motion from heart, lungs and diaphragm, by the near-field coherent sensing principle. We can achieve a satisfactory correlation of our sensor with the reference devices for the three vital signs: heart rate ( r = 0.95), respiratory rate ( r = 0.93) and respiratory volume ( r = 0.84). We also detected voluntary breath-hold periods with an accuracy of 96%. Further, the participants performed a breathing exercise by contracting abdomen inwards while holding breath, leading to paradoxical outward thorax motion under the isovolumetric condition, which was detected with an accuracy of 83%.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2017

Age and Method of Release Affect Migratory Performance of Hatchery Steelhead

Abstract Hatchery programs that are designed to aid recovery of natural populations of anadromous salmonids, including steelhead Oncorhynchus mykiss, require locally derived, natural-origin broodstock. In such programs, achieving smoltification size thresholds may require extending hatchery rearing beyond age 1. We compared out-migration survival and travel rates of 142,990 PIT-tagged steelhead smolts released at age 1 (S1 rearing strategy) or age 2 (S2 rearing strategy) over five release years at Winthrop National Fish Hatchery (WNFH, Okanogan County, Washington). An S2 rearing cycle produced larger smolts with more uniform size distributions, resulting in higher survival during the first portion of their out-migration than for S1 smolts in three of the five release years. The S2 smolts migrated more rapidly to the ocean than S1 smolts in all years except 2011 and arrived in the Columbia River estuary 5.4 d earlier on average than the S1 smolts. The S1 steelhead that did not leave during the volitional release were subsequently forced from the hatchery to measure their survival. Nonvolitional S1 migrants were smaller and had survival rates that were 2.3–66.3 times lower than those of S1 steelhead that left WNFH on their own. The same was true for S2 steelhead, but the magnitude of the survival difference between volitional migrants and fish forced from the raceways was less variable and ranged from 2.5- to 4.6-fold. We conclude that S2 rearing can be a successful strategy for producing smolts from local natural-origin broodstock, with out-migration survival and travel times that are equivalent to or better than those of S1 smolts produced from nonlocal broodstock. Received July 25, 2016; accepted March 31, 2017 Published online May 31, 2017

Tatara, Christopher P.↗
DOE OSTI2015

Beam’s-eye-view imaging during non-coplanar lung SBRT

Purpose: Beam’s-eye-view (BEV) imaging with an electronic portal imaging device (EPID) can be performed during lung stereotactic body radiation therapy (SBRT) to monitor the tumor location in real-time. Image quality for each patient and treatment field depends on several factors including the patient anatomy and the gantry and couch angles. The authors investigated the angular dependence of automatic tumor localization during non-coplanar lung SBRT delivery. Methods: All images were acquired at a frame rate of 12 Hz with an amorphous silicon EPID. A previously validated markerless lung tumor localization algorithm was employed with manual localization as the reference. From ten SBRT patients, 12 987 image frames of 123 image sequences acquired at 48 different gantry–couch rotations were analyzed. δ was defined by the position difference of the automatic and manual localization. Results: Regardless of the couch angle, the best tracking performance was found in image sequences with a gantry angle within 20° of 250° (δ = 1.40 mm). Image sequences acquired with gantry angles of 150°, 210°, and 350° also led to good tracking performances with δ = 1.77–2.00 mm. Overall, the couch angle was not correlated with the tracking results. Among all the gantry–couch combinations, image sequences acquired at (θ = 30°, ϕ = 330°), (θ = 210°, ϕ = 10°), and (θ = 250°, ϕ = 30°) led to the best tracking results with δ = 1.19–1.82 mm. The worst performing combinations were (θ = 90° and 230°, ϕ = 10°) and (θ = 270°, ϕ = 30°) with δ > 3.5 mm. However, 35% (17/48) of the gantry–couch rotations demonstrated substantial variability in tracking performances between patients. For example, the field angle (θ = 70°, ϕ = 10°) was acquired for five patients. While the tracking errors were ≤1.98 mm for three patients, poor performance was found for the other two patients with δ ≥ 2.18 mm, leading to average tracking error of 2.70 mm. Only one image sequence was acquired for all other gantry–couch rotations (δ = 1.18–10.29 mm). Conclusions: Non-coplanar beams with gantry–couch rotation of (θ = 30°, ϕ = 330°), (θ = 210°, ϕ = 10°), and (θ = 250°, ϕ = 30°) have the highest accuracy for BEV lung tumor localization. Additionally, gantry angles of 150°, 210°, 250°, and 350° also offer good tracking performance. The beam geometries (θ = 90° and 230°, ϕ = 10°) and (θ = 270°, ϕ = 30°) are associated with substantial automatic localization errors. Overall, lung tumor visibility and tracking performance were patient dependent for a substantial number of the gantry–couch angle combinations studied.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2019

Kinetics of Recycling CF/EP Composites by Thermal Excitation of Cr{sub 2}O{sub 3}

Recycling high performance carbon fibers (CFs) by thermal excitation of Cr{sub 2}O{sub 3} semiconductor was proposed. The kinetic model of decomposition of epoxy resin was established and the method for solving kinetic parameters was proposed. The influences of temperature and treatment time on the decomposition rate of epoxy resin were investigated. The kinetic equation was established and the accuracy was verified experimentally. The morphology of the recycled CF was observed and monofilament tensile strength was determined. Results indicated that temperature and treatment time had positive correlation with decomposition rate of the epoxy resin. The reaction order n was 1.5, with excitation energy E of 85.03 kJ·mol{sup −1} and pre-exponential factor k{sub 0} of 1.26 × 10{sup 6} min{sup −1}. The established kinetic equation could be used to estimate the temperature and treatment time. Prolongation of treatment time could improve the monofilament tensile strength. The monofilament tensile strength of the recycled CF at 480 °C for 30 min was 102.72% of that of the original CF, indicating that surface heat treatment was performed on CF during the recycling process.

36 MATERIALS SCIENCE↗
DOE OSTI2026

Simulation of Channel Flow with Square Ribs for Blanket First-Wall Cooling: Geometry-Specific Tuning of k-ω Model Using Adjoint Method

Cooling of the plasma-facing first wall is challenging in the design of blanket components because of the high heat flux (on the order of 𝑀𝑊/𝑚2) from the plasma, especially when a low thermal mass medium like helium is chosen as the coolant. Therefore, heat transfer enhancement in which the convective heat transfer rate is augmented by the addition of turbulence-promoting structures becomes a key initiative for providing sufficient cooling capability with helium. Previously, computational fluid dynamics simulations had been performed on pipe flows with different transverse and longitudinal ribbed geometries at Oak Ridge National Laboratory to compare the enhancement performance among different ribbed geometries. Rib shape morphing had been conducted to obtain an optimized rib profile. In the work presented here, the adjoint method is adopted in the ANSYS Fluent solver for turbulence model augmentation, and the Generalized k-ω (GEKO) turbulence model is employed because of its ability of tuning the turbulence model. The Nusselt number and pressure drop obtained from the channel flow with bottom ribbed wall experiments are used as the targets. Sensitivity analysis provides information as guidance to improve the turbulence model accuracy. The augmented GEKO model is tuned for the studied ribbed channel geometry and flow conditions, providing improved predictive accuracy within this context. Extension to other configurations offers potential but may require additional tuning and validation.

Xu, Tracy [ORNL] (ORCID:0009000193700887)↗
DOE OSTI2015

Comparison and uncertainty evaluation of different calibration protocols and ionization chambers for low-energy surface brachytherapy dosimetry

Purpose: A surface electronic brachytherapy (EBT) device is in fact an x-ray source collimated with specific applicators. Low-energy (<100 kVp) x-ray beam dosimetry faces several challenges that need to be addressed. A number of calibration protocols have been published for x-ray beam dosimetry. The media in which measurements are performed are the fundamental difference between them. The aim of this study was to evaluate the surface dose rate of a low-energy x-ray source with small field applicators using different calibration standards and different small-volume ionization chambers, comparing the values and uncertainties of each methodology. Methods: The surface dose rate of the EBT unit Esteya (Elekta Brachytherapy, The Netherlands), a 69.5 kVp x-ray source with applicators of 10, 15, 20, 25, and 30 mm diameter, was evaluated using the AAPM TG-61 (based on air kerma) and International Atomic Energy Agency (IAEA) TRS-398 (based on absorbed dose to water) dosimetry protocols for low-energy photon beams. A plane parallel T34013 ionization chamber (PTW Freiburg, Germany) calibrated in terms of both absorbed dose to water and air kerma was used to compare the two dosimetry protocols. Another PTW chamber of the same model was used to evaluate the reproducibility between these chambers. Measurements were also performed with two different Exradin A20 (Standard Imaging, Inc., Middleton, WI) chambers calibrated in terms of air kerma. Results: Differences between surface dose rates measured in air and in water using the T34013 chamber range from 1.6% to 3.3%. No field size dependence has been observed. Differences are below 3.7% when measurements with the A20 and the T34013 chambers calibrated in air are compared. Estimated uncertainty (with coverage factor k = 1) for the T34013 chamber calibrated in water is 2.2%–2.4%, whereas it increases to 2.5% and 2.7% for the A20 and T34013 chambers calibrated in air, respectively. The output factors, measured with the PTW chambers, differ by less than 1.1% for any applicator size when compared to the output factors that were measured with the A20 chamber. Conclusions: Measurements using both dosimetric protocols are consistent, once the overall uncertainties are considered. There is also consistency between measurements performed with both chambers calibrated in air. Both the T34013 and A20 chambers have negligible stem effect. Any x-ray surface brachytherapy system, including Esteya, can be characterized using either one of these calibration protocols and ionization chambers. Having less correction factors, lower uncertainty, and based on measurements, performed in closer to clinical conditions, the TRS-398 protocol seems to be the preferred option.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2025

Openpronghorn

OpenPronghorn is a simulation tool specifically tailored for modeling thermal-hydraulic phenomena in advanced nuclear reactors. It is built on the Multiphysics Object-Oriented Simulation Environment (MOOSE), an open-source platform that facilitates the development of high-performance scientific computing applications. OpenPronghorn solves the Navier-Stokes equations, which describe the conservation of mass, momentum, and energy in fluid flows, using the finite volume numerical method. The code supports a wide range of fluid flow conditions that are applicable to nuclear reactors, including incompressible and weakly compressible flows, as well as single-phase and multiphase flows. It is capable of modeling diverse flow regimes, including laminar and turbulent flows, using various turbulence models such as the standard k-epsilon models, the v2f model, and the mixing length model. For multiphase flows, OpenPronghorn employs a mixture a Eulerian modeling approach with mixture, drift-flux, and full Eulerian models, and includes open-sourced interfacial transfer correlations for drag, exchange, and heat transfer coming from the scientific literature. OpenPronghorn's modular design allows it to handle multiscale simulations, ranging from detailed Reynolds-Averaged Navier Stokes (RANS) simulations to coarse-mesh and lumped parameter models. This flexibility enables users to perform high-fidelity simulations of specific reactor components as well as system-level analyses of entire reactor circuits. The code can be coupled with other MOOSE-based tools using the MultiApp system, allowing for the transfer of coupling quantities such as mass flow rates, heat fluxes, and boundary conditions between different simulation scales. One of the main features of OpenPronghorn is the it includes built-in validation cases from the open-source scientific literature and supports the implementation of user-defined models and correlations through MOOSE's FunctorMaterial system. OpenPronghorn is designed to be computationally efficient, leveraging the SIMPLE projection method for large-scale problems, and can be run on high-performance computing systems to handle the extensive computational demands of detailed reactor simulations. Overall, OpenPronghorn is a versatile and robust tool that provides critical insights into the thermal-hydraulic behavior of advanced nuclear reactors, supporting the design, safety, and optimization of next-generation nuclear energy systems.

Retamales, Mauricio Eduardo Tano [Idaho National L↗
DOE OSTI2025

Explaining word embeddings with perfect fidelity: a case study in predicting research impact

The best-performing approaches for scholarly document quality prediction are based on embedding models. In addition to their performance when used in classifiers, embedding models can also provide predictions even for words that were not contained in the labelled training data for the classification model, which is important in the context of the ever-evolving research terminology. Although model-agnostic explanation methods, such as Local interpretable model-agnostic explanations, can be applied to explain machine learning classifiers trained on embedding models, these produce results with questionable correspondence to the model. We introduce a new feature importance method, Self-Model Entities Rated (SMER), for logistic regression-based classification models trained on word embeddings. We show that SMER has theoretically perfect fidelity with the explained model, as the average of logits of SMER scores for individual words (SMER explanation) exactly corresponds to the logit of the prediction of the explained model. Quantitative and qualitative evaluation is performed through five diverse experiments conducted on 50,000 research articles (papers) from the CORD-19 corpus. In conclusion, through an AOPC curve analysis, we experimentally demonstrate that SMER produces better explanations than LIME, SHAP and global tree surrogates.

Coarse-grained models↗