Search NASA⌕ Search

SEARCH · Search NASA

Results for “higher accuracy”

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 649 records · Page 36
All sourcesNASA NTRSDOE OSTI
NASA NTRS2024

Uncertainty Budget for Detector-Based Absolute Radiometric Calibration With GLAMR

The accuracy of the absolute radiometric calibration (RadCal) for remote sensing instruments is essential to their wide range of applications. The uncertainty associated to the traditional source-based RadCal method is assessed at a 2% (k=1) or higher level for radiance measurement. To further improve the accuracy to meet the demands of climate studies, a detector-based approach using tunable lasers as a light source has been devised. The Goddard Laser for Absolute Measurement of Radiance, known as the GLAMR system, is a notable example of the incorporation of such technology. Using transfer radiometers calibrated at NIST as calibration standards, the absolute spectral response (ASR) function of a remote sensing instrument is measured with its uncertainty traceable to the International System of Units. This paper presents a comprehensive uncertainty analysis of the detector-based absolute RadCal using the GLAMR system. It identifies and examines uncertainty sources during the GLAMR RadCal test, including those from the GLAMR system, the testing configuration, and data processing methodologies. Analysis is carried out to quantify the contribution of each source and emphasize the most influential factors. It is shown that the calibration uncertainty of GLAMR RadCal can be better than 0.3% (k=1) in the wavelength range of 350-950 nm and 0.6% (k=1) between 950-2300 nm, with the exception of regions with strong water absorption. In addition, recommendations are made to refine the calibration process to further reduce the uncertainty.

Zhipeng Wang↗
DOE OSTI2024

Accelerating defect predictions in semiconductors using graph neural networks

First-principles computations reliably predict the energetics of point defects in semiconductors but are constrained by the expense of using large supercells and advanced levels of theory. Machine learning models trained on computational data, especially ones that sufficiently encode defect coordination environments, can be used to accelerate defect predictions. Here, we develop a framework for the prediction and screening of native defects and functional impurities in a chemical space of group IV, III–V, and II–VI zinc blende semiconductors, powered by crystal Graph-based Neural Networks (GNNs) trained on high-throughput density functional theory (DFT) data. Using an innovative approach of sampling partially optimized defect configurations from DFT calculations, we generate one of the largest computational defect datasets to date, containing many types of vacancies, self-interstitials, anti-site substitutions, impurity interstitials and substitutions, as well as some defect complexes. We applied three types of established GNN techniques, namely crystal graph convolutional neural network, materials graph network, and Atomistic Line Graph Neural Network (ALIGNN), to rigorously train models for predicting defect formation energy (DFE) in multiple charge states and chemical potential conditions. We find that ALIGNN yields the best DFE predictions with root mean square errors around 0.3 eV, which represents a prediction accuracy of 98% given the range of values within the dataset, improving significantly on the state-of-the-art. We further show that GNN-based defective structure optimization can take us close to DFT-optimized geometries at a fraction of the cost of full DFT. The current models are based on the semi-local generalized gradient approximation-Perdew–Burke–Ernzerhof (PBE) functional but are highly promising because of the correlation of computed energetics and defect levels with higher levels of theory and experimental data, the accuracy and necessity of discovering novel metastable and low energy defect structures at the PBE level of theory before advanced methods could be applied, and the ability to train multi-fidelity models in the future with new data from non-local functionals. The DFT-GNN models enable prediction and screening across thousands of hypothetical defects based on both unoptimized and partially optimized defective structures, helping identify electronically active defects in technologically important semiconductors.

Rahman, Md Habibur (ORCID:000000027705984X)↗
DOE OSTI2024

A hybrid finite volume method and smoothed particle hydrodynamics approach for efficient and accurate blast simulations

Modeling strong shock waves in fluids remains a persistent challenge in computational physics. Essential to research efforts in industry and defense, numerous methods have been devised to improve the accuracy and efficiency of shock simulations. A novel, hybrid Finite Volume Method (FVM)-Smoothed Particle Hydrodynamics (SPH) approach is capable of further improving efficiency and retaining accuracy by exploiting the favorable characteristics of each respective method. This hybrid approach is presented for shock capturing in compressible fluids. The Python framework Pyro2 is employed to simulate a coarse FVM mesh, while the Python framework PySPH is utilized to model the fluid in regions with high gradients through SPH particles. The performance of the hybrid FVM-SPH scheme, compared to the individual FVM and SPH methods, is assessed in 1 kt and 10 kt blast simulations. Our results indicate that the hybrid approach offers higher computational efficiency than SPH while preserving its accuracy and characteristics. The hybrid approach had a relative speedup of 11.3x and 22.3x over the FVM and SPH approaches for the 1 kt simulation and a relative speedup of 14.7x and 20.9x over the FVM and SPH approaches for the 10 kt simulation. The hybrid SPH algorithm enables future compressible fluid simulations with more extensive capabilities than grid-based methods alone, presenting potential applications in modeling fluid-structure interactions and solid deformation and fracturing in blast simulations.

Myers, Conner↗
DOE OSTI2021

Improving scalability of parallel CNN training by adaptively adjusting parameter update frequency

Synchronous SGD with data parallelism, the most popular parallelization strategy for CNN training, suffers from the expensive communication cost of averaging gradients among all workers. The iterative parameter updates of SGD cause frequent communications and it becomes the performance bottleneck. In this paper, we propose a lazy parameter update algorithm that adaptively adjusts the parameter update frequency to address the expensive communication cost issue. Our algorithm accumulates the gradients if the difference of the accumulated gradients and the latest gradients is sufficiently small. Here, the less frequent parameter updates reduce the per-iteration communication cost while maintaining the model accuracy. Our experimental results demonstrate that the lazy update method remarkably improves the scalability while maintaining the model accuracy. For ResNet50 training on ImageNet, the proposed algorithm achieves a significantly higher speedup (739.6 on 2048 Cori KNL nodes) as compared to the vanilla synchronous SGD (276.6) while the model accuracy is almost not affected (<0.2% difference).

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Dynamical structure factors of warm dense matter from time-dependent orbital-free and mixed-stochastic-deterministic density functional theory

Abstract We present the first calculations of the inelastic part of the dynamical structure factor (DSF) for warm dense matter (WDM) using time-dependent orbital-free density functional theory (TD-OF-DFT) and mixed-stochastic-deterministic (mixed) Kohn Sham TD-DFT (KS TD-DFT). WDM is an intermediate phase of matter found in planetary cores and laser-driven experiments, where the accurate calculation of the DSF is critical for interpreting x-ray Thomson scattering measurements. Traditional TD-DFT methods, while highly accurate, are computationally expensive, motivating the exploration of TD-OF-DFT and mixed TD-KS-DFT as more efficient alternatives. We applied these methods to experimentally measured WDM systems, including solid-density aluminum and beryllium, compressed beryllium, and carbon–hydrogen mixtures. Our results show that TD-OF-DFT requires a dynamical kinetic energy potential in order to qualitatively capture the plasmon response. Additionally, it struggles with capturing bound electron contributions. In contrast, mixed TD-KS-DFT offers greater accuracy in distinguishing bound and free electron effects, aligning well with experimental data, though at a higher computational cost. This study highlights the trade-offs between computational efficiency and accuracy, demonstrating that TD-OF-DFT remains a valuable tool for rapid scans of parameter space, while mixed TD-KS-DFT should be preferred for high-fidelity simulations. Our findings provide insight into the future development of DFT methods for WDM and suggest potential improvements for TD-OF-DFT.

36 MATERIALS SCIENCE↗
NASA NTRS2013

A Kinematic Calibration Process for Flight Robotic Arms

The Mars Science Laboratory (MSL) robotic arm is ten times more massive than any Mars robotic arm before it, yet with similar accuracy and repeatability positioning requirements. In order to assess and validate these requirements, a higher-fidelity model and calibration processes were needed. Kinematic calibration of robotic arms is a common and necessary process to ensure good positioning performance. Most methodologies assume a rigid arm, high-accuracy data collection, and some kind of optimization of kinematic parameters. A new detailed kinematic and deflection model of the MSL robotic arm was formulated in the design phase and used to update the initial positioning and orientation accuracy and repeatability requirements. This model included a higher-fidelity link stiffness matrix representation, as well as a link level thermal expansion model. In addition, it included an actuator backlash model. Analytical results highlighted the sensitivity of the arm accuracy to its joint initialization methodology. Because of this, a new technique for initializing the arm joint encoders through hardstop calibration was developed. This involved selecting arm configurations to use in Earth-based hardstop calibration that had corresponding configurations on Mars with the same joint torque to ensure repeatability in the different gravity environment. The process used to collect calibration data for the arm included the use of multiple weight stand-in turrets with enough metrology targets to reconstruct the full six-degree-of-freedom location of the rover and tool frames. The follow-on data processing of the metrology data utilized a standard differential formulation and linear parameter optimization technique.

Collins, Curtis L.↗
DOE OSTI2022

MULTISTEP AND CONTINUOUS PHYSICS-INFORMED NEURAL NETWORK METHODS FOR LEARNING GOVERNING EQUATIONS AND CONSTITUTIVE RELATIONS

In this work, we investigate the applicability and relative merit of discrete and continuous versions of physics-informed neural network (PINN) methods for learning unknown governing equations or constitutive relations in a nonlinear dynamical system. In the case of unknown dynamics, entire right-hand-side (RHS) equations of the ordinary differential equations are unknown. In the case of unknown constitutive relations, however, the RHS equations are known up to the specification of constitutive relations (that may depend on the state of the system). We use a deep neural network to model unknown governing equations or constitutive relations. The discrete PINN approach combines classical multistep discretization methods for dynamical systems with neural-network-based machine learning methods. On the other hand, the continuous versions utilize deep neural networks to minimize the residual function for the continuous governing equations. We use the case of a fedbatch bioreactor system to study the effectiveness of these approaches and discuss conditions for their applicability. Our results indicate that the accuracy of the trained neural network models is much higher for the cases where we only have to learn a constitutive relation instead of all dynamics. This finding corroborates the well-known fact from scientific computing that building as much structural information as is available into an algorithm can enhance its efficiency and/or accuracy.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1985

An analysis of Landsat Thematic Mapper P-Product internal geometry and conformity to earth surface geometry

Performance requirements regarding geometric accuracy have been defined in terms of end product goals, but until recently no precise details have been given concerning the conditions under which that accuracy is to be achieved. In order to achieve higher spatial and spectral resolutions, the Thematic Mapper (TM) sensor was designed to image in both forward and reverse mirror sweeps in two separate focal planes. Both hardware and software have been augmented and changed during the course of the Landsat TM developments to achieve improved geometric accuracy. An investigation has been conducted to determine if the TM meets the National Map Accuracy Standards for geometric accuracy at larger scales. It was found that TM imagery, in terms of geometry, has come close to, and in some cases exceeded, its stringent specifications.

Bryant, N. A.↗
NASA NTRS2018

Automated Pneumothorax Diagnosis using Deep Neural Networks

Thoracic ultrasound can provide information leading to rapid diagnosis of pneumothorax with improved accuracy over the standard physical examination and with higher sensitivity than anteroposterior chest radiography. However, the clinical We have Furthermore, remote environments, such as the battlefield or deep-space exploration, may lack expertise for diagnosing developed an automated image interpretation pipeline for the analysis of thoracic ultrasound data and the classification of pneumothorax events to provide decision support in such situations. Our pipeline consists of image preprocessing, data augmentation, and deep learning architectures for medical diagnosis. In this work, we demonstrate that robust, accurate interpretation of chest images and video can be achieved using deep neural networks. A number of novel image processing techniques were employed to achieve this result. Affine transformations were applied for data augmentation. Hyperparameters were optimized for learning rate, dropout regularization, batch size, and epoch iteration by a sequential model-based Bayesian approach. In addition, we utilized pretrained architecturesinterpretation of a patient medical image is highly operator dependent. certain pathologies., applying transfer learning and fine-tuning techniques to fully connected layers. Our pipeline yielded binary classification validation accuracies of 98.3% for M-mode images and 99.8% with B-mode video frames.

US Army collaboration↗
NASA NTRS1992

Research in computational fluid dynamics and analysis of algorithms

Recently, higher-order compact schemes have seen increasing use in the DNS (Direct Numerical Simulations) of the Navier-Stokes equations. Although they do not have the spatial resolution of spectral methods, they offer significant increases in accuracy over conventional second order methods. They can be used on any smooth grid, and do not have an overly restrictive CFL dependence as compared with the O(N(exp -2)) CFL dependence observed in Chebyshev spectral methods on finite domains. In addition, they are generally more robust and less costly than spectral methods. The issue of the relative cost of higher-order schemes (accuracy weighted against physical and numerical cost) is a far more complex issue, depending ultimately on what features of the solution are sought and how accurately they must be resolved. In any event, the further development of the underlying stability theory of these schemes is important. The approach of devising suitable boundary clusters and then testing them with various stability techniques (such as finding the norm) is entirely the wrong approach when dealing with high-order methods. Very seldom are high-order boundary closures stable, making them difficult to isolate. An alternative approach is to begin with a norm which satisfies all the stability criteria for the hyperbolic system, and look for the boundary closure forms which will match the norm exactly. This method was used recently by Strand to isolate stable boundary closure schemes for the explicit central fourth- and sixth-order schemes. The norm used was an energy norm mimicking the norm for the differential equations. Further research should be devoted to BC for high order schemes in order to make sure that the results obtained are reliable. The compact fourth order and sixth order finite difference scheme had been incorporated into a code to simulate flow past circular cylinders. This code will serve as a verification of the full spectral codes. A detailed stability analysis by Carpenter (from the fluid Mechanics Division) and Gottlieb gave analytic conditions for stability as well as asymptotic stability. This had been incorporated in the code in form of stable boundary conditions. Effects of the cylinder rotations had been studied. The results differ from the known theoretical results. We are in the middle of analyzing the results. A detailed analysis of the effects of the heating of the cylinder on the shedding frequency had been studied using the above schemes. It has been found that the shedding frequency decreases when the wire was heated. Experimental work is being carried out to affirm this result.

Gottlieb, David↗
DOE OSTI2025

Hierarchical Bayesian Modeling for Cosmology: Can NPE reliably replace MCMC?

Hierarchical neural posterior estimation has its place Hierarchical Bayesian Modeling (HBM) combined with MCMC algorithms has been shown to provide more robust and accurate inference for real-world phenomena in which nature takes a nested form. However, MCMC-based inference can be computationally expensive, and its performance often suffers for complex posterior geometries. These costs are especially pertinent for HBM. Studies have recently demonstrated the potential for a flexible, expressive, and amortized hierarchical neural posterior estimator (HNPE) built on Normalizing Flows. These studies have mostly been performed on simple datasets, or they focus on a single parameter from each level of the hierarchy. A systematic study analyzing how both hierarchical methods compare for more complex and realistic datasets is necessary before applying HNPE for scientific measurements. Here, we re-explore the theory behind HNPE and conduct comparative numerical experiments of HNPE and MCMC-based HBM methods on real and synthetic data, including strong gravitational lensing simulations. In particular, we use a suite of diagnostics to show trade-offs in terms of accuracy, precision, time to train or sample, reproducibility, and the need for expert domain knowledge. Especially for higher dimensional and complex posteriors, HNPE is expected to drastically improve on time for inference, accuracy, and precision with an upfront training time cost.

Hur, Rachel [Chicago U.] (ORCID:000900089890445X)↗
DOE OSTI2025

Bayesian modeling of traffic-related air pollutants: A case study of urban transportation and air quality dynamics in Columbia, South Carolina

Traffic emissions significantly impact near-road air quality and public health. This research applies a Bayesian modeling framework to investigate these impacts using high-resolution traffic and air pollutant data from an urban corridor in Columbia, South Carolina. Despite a data collection period truncated by the COVID-19 lockdown, the Bayesian approach successfully identified significant predictors and quantified model uncertainty. Employing Bayesian Model Selection and Averaging enhanced prediction accuracy and evaluated model uncertainty. Findings indicate that higher temperatures and increased moisture levels elevate particulate matter (PM 1.0 , PM 2.5 , PM 10 ) concentrations, while traffic speed significantly affects nitrogen dioxide (NO 2 ) levels. Specifically, higher average traffic speeds (indicative of smoother flow) correspond to lower NO 2 concentrations, suggesting that less congested conditions reduce NO 2 emissions. This study highlights the robustness of Bayesian methods for generating reliable air quality insights even under data-constrained conditions. The findings underscore the importance of traffic flow management (e.g., reducing congestion) for mitigating near-road NO 2 exposure and provide a basis for developing targeted public health strategies.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

Binder jetting additive manufacturing of silicon carbide ceramics: Development of bimodal powder feedstocks by modeling and experimental methods

A limitation of binder jetting additive manufacturing is the low density of fabricated parts. Mixing powders with different sizes is a promising approach to increase powder bed packing density and, hence, printed part density. However, in previous studies mixed powder feedstock was prepared by trial and error method. In this research, both modeling and experimental methods were used to prepare the bimodal powder feedstocks. Analytical packing model was introduced for irregular powders. A bimodal powder was prepared by mixing two different-sized silicon carbide powders (i.e. coarse and fine) using ball mill, and their tap densities were measured. Silicon carbide plates were printed using the coarse and bimodal powders by a commercial binder jetting system. Results showed that the modeling method could predict the tap density of bimodal powders with high accuracy. The printed parts from bimodal powder achieved higher green densities than those from the unimodal powder.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Convergent Concordant Mode Approach for Molecular Vibrations: CMA-2

The concordant mode approach (CMA) is a promising new scheme for dramatically increasing the system size and level of theory achievable in quantum chemical computations of molecular vibrational frequencies. Here, we achieve advances in the CMA hierarchy by computations targeting CCSD(T)/cc-pVTZ (coupled cluster singles and doubles with perturbative triples using a correlation-consistent polarized-valence triple-ζ basis set) benchmarks within the G2 molecular test set, executing a statistical analysis for 1501 frequencies from 111 compounds and then separately solving the refractory case of pyridine. First, MP2/cc-pVTZ (second-order Møller–Plesset perturbation theory with the same basis set) proves to be an excellent and preferred choice for generating the underlying (Level B) normal modes of the CMA scheme. Utilizing this Level B within the CMA-0A method reproduces the 1501 benchmark frequencies with a mean absolute error (MAE) of only 0.11 cm –1 and an attendant standard deviation of 0.49 cm –1 . Second, a convergent CMA-2 method is constituted that allows efficient computation of higher level (Level A) frequencies to any reasonable accuracy threshold by using only Hartree–Fock (HF) and MP2 or density functional theory (DFT) data to generate ξ parameters, which select the sparse off-diagonal force field elements for explicit evaluation at Level A. When Level B = MP2/cc-pVTZ, a cutoff of ξ = 0.02 provides an average maximum absolute error per molecule of only 0.17 cm –1 by incurring merely a 33% increase in average cost over CMA-0A. This CMA-2 method also eradicates the 4 problematic CMA-0A outliers of pyridine with even less effort (ξ = 0.04, 22% increase). Finally, the newly developed CMA procedures are shown to be highly successful when applied to 1-(1H-pyrrol-3-yl)ethanol, a new test molecule with diverse types of vibration.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Active Thermochemical Tables: Enthalpies of Formation of Bromo- and Iodo-Methanes, Ethenes and Ethynes

Here, the thermochemistry of halocarbon species containing iodine and bromine is examined through an extensive interplay between new Feller–Peterson–Dixon (FPD) style composite methods and a detailed analysis of all available experimental and theoretical determinations using the thermochemical network that underlies the Active Thermochemical Tables (ATcT). From the computational viewpoint, a slower convergence of the components of composite thermochemistry methods is observed relative to species that solely contain first row elements, leading to a higher computational expense for achieving comparable levels of accuracy. Potential systematic sources of computational uncertainty are investigated, and, not surprisingly, spin-orbit coupling is found to be a critical component, particularly for iodine containing molecular species. The ATcT analysis of available experimental and theoretical determinations indicates that prior theoretical determinations have significantly larger uncertainties than originally reported, particularly in cases where molecular spin-orbit effects were ignored. Accurate and reliable heats of formation are reported for 38 halogen containing systems, based on combining the current computations with previous experimental and theoretical work via the ATcT approach.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Quantum mechanical dataset of 836k neutral closed-shell molecules with up to 5 heavy atoms from C, N, O, F, Si, P, S, Cl, Br

Abstract We introduce the Vector-QM24 (VQM24) dataset comprehensively covering all possible neutral closed-shell small organic and inorganic molecules with up to five heavy (p-block) atoms: C, N, O, F, Si, P, S, Cl, Br. All valid stoichiometries, Lewis-rule-consistent graphs, and stable conformers (identified via GFN2-xTB) were enumerated combinatorially, yielding 577k conformational isomers spanning 258k constitutional isomers and 5,599 unique stoichiometries. DFT (ωB97X-D3/cc-pVDZ) optimizations were performed for all, and diffusion quantum Monte Carlo (DMC@PBE0(ccECP/cc-pVQZ)) energies are provided for 10,793 lowest-energy conformers with up to 4 heavy atoms. VQM24 includes structures, vibrational modes, rotational constants, thermodynamic properties (Gibbs free energies, enthalpies, ZPVEs, entropies, heat capacities), and electronic properties such as atomization, electron interaction, exchange-correlation, dispersion energies, multipole moments (dipole to hexadecapole), alchemical potentials, Mulliken charges, and wavefunctions. Machine learning models of atomization energies on this dataset reveal significantly higher complexity than QM9, with none achieving chemical accuracy. VQM24 offers a rigorous, high-fidelity benchmark for evaluating quantum machine learning models.

Science & Technology - Other Topics↗
DOE OSTI2021

Pushing automated morphological classifications to their limits with the Dark Energy Survey

ABSTRACT We present morphological classifications of ∼27 million galaxies from the Dark Energy Survey (DES) Data Release 1 (DR1) using a supervised deep learning algorithm. The classification scheme separates: (a) early-type galaxies (ETGs) from late-type galaxies (LTGs); and (b) face-on galaxies from edge-on. Our convolutional neural networks (CNNs) are trained on a small subset of DES objects with previously known classifications. These typically have mr ≲ 17.7 mag; we model fainter objects to mr < 21.5 mag by simulating what the brighter objects with well-determined classifications would look like if they were at higher redshifts. The CNNs reach 97 per cent accuracy to mr < 21.5 on their training sets, suggesting that they are able to recover features more accurately than the human eye. We then used the trained CNNs to classify the vast majority of the other DES images. The final catalogue comprises five independent CNN predictions for each classification scheme, helping to determine if the CNN predictions are robust or not. We obtain secure classifications for ∼87 per cent and 73 per cent of the catalogue for the ETG versus LTG and edge-on versus face-on models, respectively. Combining the two classifications (a) and (b) helps to increase the purity of the ETG sample and to identify edge-on lenticular galaxies (as ETGs with high ellipticity). Where a comparison is possible, our classifications correlate very well with Sérsic index (n), ellipticity (ϵ), and spectral type, even for the fainter galaxies. This is the largest multiband catalogue of automated galaxy morphologies to date.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Variation Tolerant and Energy-Efficient Charge Domain Compute-in-Memory Array with Binary and Multi-Level Cell Ferroelectric FET

Here, in this work, we present a variation-tolerant and energy-efficient charge-domain Ferroelectric FET (FeFET) based Compute-in-Memory (CiM) array design that is compatible with both binary and multi-level cell memory sensing. We demonstrate that: 1) by exploiting FeFET as a nonvolatile switch, its high ON/OFF ratio in the subthreshold region can suppress the error introduced by the inaccurate ON state conductance, thus realizing robust CiM operations, unlike the current-domain CiM design where the computation results is highly sensitive to the device conductance variation; 2) by leveraging a dense dynamic random access memory (DRAM)-like 1FeFET1C cell structure, the proposed design benefits from the existing high density DRAM establishment while also significantly relaxing the capacitor retention and transistor leakage requirement; 3) the charge-domain CiM supports both binary FeFET with minimum overhead and MLC FeFET with tolerable latency for MLC state sensing, whose efficacy is validated experimentally on both cell-level and array-level; 4) the proposed CiM shows much better device variation resilience than conventional current-domain CiM, and also improves inference accuracy. Macro-level evaluation results demonstrate significantly higher energy efficiency and area efficiency compared to prior CiM works.

Duan, Jiahui [University of Notre Dame, IN (United↗