SEARCH · Engineering Papers

Results for “task tuning”

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 127 records · Page 7

All sources NASA NTRS DOE OSTI

DOE OSTI2026

Performance-Aligned LLMs for Generating Fast HPC Code

Optimizing scientific software is a difficult task because codebases are often large and complex, and performance can depend upon several factors including the algorithm, its implementation, and hardware among others. Causes of poor performance can originate from disparate sources and be difficult to diagnose. Recent years have seen a multitude of work that use large language models (LLMs) to assist in software development tasks. However, these tools are trained to model the distribution of code as text, and are not specifically designed to understand performance aspects of code. In this work, we introduce a reinforcement learning based methodology to align the outputs of code LLMs with performance. This allows us to build upon the current code modeling capabilities of LLMs and extend them to generate better performing code. Here, we demonstrate that our fine-tuned model improves the expected speedup of generated code over base models for a set of benchmark tasks from 0.9 to 1.6 for serial code and 1.9 to 4.5 for OpenMP parallel code.

Computer science↗

DOE OSTI2022

Automatic Tuner for the Step Sizes of Gradient-Based RT-OPF DERMS Control Systems [SWR-22-50]

Gradient-based controllers have parameters called step sizes that determine how sensitive its control actions are to received inputs. Finding a practical setting for a step size is important. If it is too small, the controller is ineffective by not reacting in a significant manor to its received inputs. On the other hand, if it is too large, the actions implemented by the controller may be excessive and cause the controlled system to become unstable. The main objective of the Automatic Tuner for the Step Sizes of Gradient-Based RT-OPF DERMS Control Systems is to speed up the implementation of gradient-based control schemes by automatically tuning its step sizes instead of having a practitioner spend considerable time doing the task manually. Only simple adjustments need to be made to the code of each controller with a step size to allow the addition of an individual automatic tuner. This can be applied to any RT-OPF based control for DER management.

Comden, Joshua↗

NASA NTRS2010

What's New is What's Old: Use of Bode's Integral Theorem (circa 1945) to Provide Insight for 21st Century Spacecraft Attitude Control System Design Tuning

This paper revisits the Bode integral theorem, first described in 1945 for feedback amplifier design, in the context of modern satellite Attitude Control System (ACS) design tasks. Use of Bode's Integral clarifies in an elegant way the connection between open-loop stability margins and closed-loop bandwidth. More importantly it shows that there is a very strong tradeoff between disturbance rejection below the satellite controller design bandwidth, and disturbance amplification in the 'penalty region' just above the design bandwidth. This information has been successfully used to re-tune the control designs for several NASA science-mission satellites. The Appendix of this paper contains a complete summary of the relevant integral conservation theorems for stable, unstable, and non-minimum- phase plants.

Ruth, Mike↗

NASA NTRS2000

Primate translational vestibuloocular reflexes. II. Version and vergence responses to fore-aft motion

To maintain binocular fixation on near targets during fore-aft translational disturbances, largely disjunctive eye movements are elicited the amplitude and direction of which should be tuned to the horizontal and vertical eccentricities of the target. The eye movements generated during this task have been investigated here as trained rhesus monkeys fixated isovergence targets at different horizontal and vertical eccentricities during 10 Hz fore-aft oscillations. The elicited eye movements complied with the geometric requirements for binocular fixation, although not ideally. First, the corresponding vergence angle for which the movement of each eye would be compensatory was consistently less than that dictated by the actual fixation parameters. Second, the eye position with zero sensitivity to translation was not straight ahead, as geometrically required, but rather exhibited a systematic dependence on viewing distance and vergence angle. Third, responses were asymmetric, with gains being larger for abducting and downward compared with adducting and upward gaze directions, respectively. As frequency was varied between 4 and 12 Hz, responses exhibited high-pass filter properties with significant differences between abduction and adduction responses. As a result of these differences, vergence sensitivity increased as a function of frequency with a steeper slope than that of version. Despite largely undercompensatory version responses, vergence sensitivity was closer to ideal. Moreover, the observed dependence of vergence sensitivity on vergence angle, which was varied between 2.5 and 10 MA, was largely linear rather than quadratic (as geometrically predicted). We conclude that the spatial tuning of eye velocity sensitivity as a function of gaze and viewing distance follows the general geometric dependencies required for the maintenance of foveal visual acuity. However, systematic deviations from ideal behavior exist that might reflect asymmetric processing of abduction/adduction responses perhaps because of different functional dependencies of version and vergence eye movement components during translation.

NASA Program Biomedical Research and Countermeasur↗

DOE OSTI2025

A kinetic-based regularization method for data science applications

We propose a physics-based regularization technique for function learning, inspired by statistical mechanics. By drawing an analogy between optimizing the parameters of an interpolator and minimizing the energy of a system, we introduce corrections that impose constraints on the lower-order moments of the data distribution. This minimizes the discrepancy between the discrete and continuum representations of the data, in turn allowing to access more favorable energy landscapes, thus improving the accuracy of the interpolator. Our approach improves performance in both interpolation and regression tasks, even in high-dimensional spaces. Unlike traditional methods, it does not require empirical parameter tuning, making it particularly effective for handling noisy data. We also show that thanks to its local nature, the method offers computational and memory efficiency advantages over Radial Basis Function interpolators, especially for large datasets.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Domain-specific text embedding model for accelerator physics

Accelerator physics presents unique challenges for natural language processing (NLP) due to its specialized terminology and complex concepts. A key component in overcoming these challenges is the development of robust text embedding models that transform textual data into dense vector representations, facilitating efficient information retrieval and semantic understanding. In this work, we introduce AccPhysBERT, a sentence embedding model fine-tuned specifically for accelerator physics. Our model demonstrates superior performance across a range of downstream NLP tasks, surpassing existing models in capturing the domain-specific nuances of the field. We further showcase its practical applications, including semantic paper-reviewer matching and integration into retrieval-augmented generation systems, highlighting its potential to enhance information retrieval and knowledge discovery in accelerator physics. Published by the American Physical Society 2025

Hellert, Thorsten (ORCID:0000000227970926)↗

NASA NTRS2010

Reduced Uncertainties in the Flutter Analysis of the Aerostructures Test Wing

Tuning the finite element model using measured data to minimize the model uncertainties is a challenging task in the area of structural dynamics. A test validated finite element model can provide a reliable flutter analysis to define the flutter placard speed to which the aircraft can be flown prior to flight flutter testing. Minimizing the difference between numerical and experimental results is a type of optimization problem. Through the use of the National Aeronautics and Space Administration Dryden Flight Research Center s (Edwards, California, USA) multidisciplinary design, analysis, and optimization tool to optimize the objective function and constraints; the mass properties, the natural frequencies, and the mode shapes are matched to the target data and the mass matrix orthogonality is retained. The approach in this study has been applied to minimize the model uncertainties for the structural dynamic model of the aerostructures test wing, which was designed, built, and tested at the National Aeronautics and Space Administration Dryden Flight Research Center. A 25-percent change in flutter speed has been shown after reducing the uncertainties

Pak, Chan-gi↗

NASA NTRS2011

Reduced Uncertainties in the Flutter Analysis of the Aerostructures Test Wing

Tuning the finite element model using measured data to minimize the model uncertainties is a challenging task in the area of structural dynamics. A test validated finite element model can provide a reliable flutter analysis to define the flutter placard speed to which the aircraft can be flown prior to flight flutter testing. Minimizing the difference between numerical and experimental results is a type of optimization problem. Through the use of the National Aeronautics and Space Administration Dryden Flight Research Center's (Edwards, California) multidisciplinary design, analysis, and optimization tool to optimize the objective function and constraints; the mass properties, the natural frequencies, and the mode shapes are matched to the target data, and the mass matrix orthogonality is retained. The approach in this study has been applied to minimize the model uncertainties for the structural dynamic model of the aerostructures test wing, which was designed, built, and tested at the National Aeronautics and Space Administration Dryden Flight Research Center. A 25 percent change in flutter speed has been shown after reducing the uncertainties.

Pak, Chan-Gi↗

DOE OSTI2020

The effect of lighting environment on task performance in buildings – A review

The effects of indoor environmental conditions on human health, satisfaction, and performance have been the focal point of research for decades. This paper reviews and summarizes the impact of lighting environment on task performance, specifically for the built environment audience. Existing studies included a variety of performance tests on cognitive performance and perception, visual acuity and reaction, memory, reasoning, and labor productivity. Illuminance, luminance ratio and correlated color temperature were found to affect performance in different ways, reflecting the impact of experimental techniques, conditions, performance evaluation methods used and data analysis methods. These were reviewed and categorized, with discussion on limitations related to sample size, modeling approach, carryover effects and other factors affecting individual differences in performance, with recommendations for future improvement. Although no universal conclusions can be made, in general, task performance seems to improve with higher illuminances, contrast ratios in the range of 7–11:1 (while always making sure that glare will not occur in the space) and higher correlated color temperature, while spectral tuning in the red or blue wavelengths has also shown positive effects. To obtain more generic evidence, future studies should be more consistent in terms of experimental procedures and overall light conditions, and also consider the effects of vertical illuminance, daylight provision/control, and outside views on task performance. Finally, studying performance with multi-factorial designs in a human-centered optimized manner (such as deploying variable lighting scenarios optimized for various tasks) can lead to deeper understanding of lighting effects on task performance, and ultimately to improved lighting design and operation in buildings overall.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2025

Data Agnostic Feature-Target Analysis & Ranking Machine Learning Pipeline (DAFTAR-ML) v0.1.0

DAFTAR-ML is a specialized machine-learning pipeline that identifies relevant features based on their relationship to a target variable. Many ML pipelines focus solely on prediction, and feature ranking is often absent or lacks robust statistical methods. DAFTAR-ML performs its tasks with this outcome in mind. Model training is robust, using nested cross-validation and hyperparameter tuning. Instead of relying on native feature-importance scores, it employs SHAP (SHapley Additive exPlanations) to quantify feature importance. The pipeline also produces comprehensive results, including publication-quality visualizations.

Melie, Tina [Lawrence Berkeley National Laboratory↗

NASA NTRS1994

Mobile robot exploration and navigation of indoor spaces using sonar and vision

Integration of skills into an autonomous robot that performs a complex task is described. Time constraints prevented complete integration of all the described skills. The biggest problem was tuning the sensor-based region-finding algorithm to the environment involved. Since localization depended on matching regions found with the a priori map, the robot became lost very quickly. If the low level sensing of the world is not working, then high level reasoning or map making will be unsuccessful.

Kortenkamp, David↗

NASA NTRS2001

Semiconductor Reference Oscillator Development for Coherent Detection Optical Remote Sensing Applications

The NASA Earth Science Enterprise Advanced Technology Initiatives Program is supporting a program for the development of semiconductor laser reference oscillators for application to coherent optical remote sensing from Earth orbit. Local oscillators provide the frequency reference required for active spaceborne optical remote sensing concepts that involve heterodyne (coherent) detection. Two recent examples of such schemes are Doppler wind lidar and tropospheric carbon dioxide measurement by laser absorption spectrometry, both of which are being proposed at a wavelength of 2.05 microns. Frequency-agile local oscillator technology is important to such applications because of the need to compensate for large platform-induced Doppler components that would otherwise interfere with data interpretation. Development of frequency-agile local oscillator approaches has heretofore utilized the same laser material as the transmitter laser (Tm,Ho:YLF in the case of the 2.05-micron wavelength mentioned above). However, a semiconductor laser-based frequency-agile local oscillator offers considerable scope for reduced mechanical complexity and improved frequency agility over equivalent crystal laser devices, while their potentially faster tuning capability suggest the potential for greater scanning versatility. The program we report on here is specifically tasked with the development of prototype novel architecture semiconductor lasers with the power, tunability, and spectral characteristics required for coherent Doppler lidar. The baseline approach for this work is the distributed feedback (DFB) laser, in which gratings are etched into the semiconductor waveguide structures along the entire length of the laser cavity. However, typical DFB lasers at the wavelength of interest have linewidths that exhibit unacceptable growth when driven at the high currents and powers that are required for the Doppler lidar application. Suppression of this behavior by means of corrugation pitch-modulation (using a detuned central section to prevent intensity peaking in the center of the cavity) is currently under investigation to achieve the required performance goals.

distributed feedback (DFB) laser↗

NASA NTRS1977

Low-noise receivers: Microwave maser development

An S-Band traveling-wave maser with a 2260- to 2400-MHz tuning capability has been built and installed on the 26-m antenna. The traveling-wave maser closed cycle refrigerator package was assembled within a limited time schedule and at low cost to meet the requirements of the very long baseline interferometry validation task. The maser uses a superconducting magnet with a field staggering coil for gain/bandwidth adjustment. The maser pump source is a Gunn effect oscillator with the capability of continuous tuning throughout the entire maser tuning range. The package was assembled from surplus components that have been used previously in the Deep Space Network in a variety of applications.

Quinn, R.↗

DOE OSTI2025

Can Large Language Models Understand Intermediate Representations?

Intermediate Representations (IRs) are essential in compiler design and program analysis, yet their comprehension by Large Language Models (LLMs) remains underexplored. This paper presents a pioneering empirical study to investigate the capabilities of LLMs, including GPT-4, GPT-3, Gemma 2, LLaMA 3.1, and Code Llama, in understanding IRs. We analyze their performance across four tasks: Control Flow Graph (CFG) reconstruction, decompilation, code summarization, and execution reasoning. Our results indicate that while LLMs demonstrate competence in parsing IR syntax and recognizing high-level structures, they struggle with control flow reasoning, execution semantics, and loop handling. Specifically, they often misinterpret branching instructions, omit critical IR operations, and rely on heuristic-based reasoning, leading to errors in CFG reconstruction, IR decompilation, and execution reasoning. The study underscores the necessity for IR-specific enhancements in LLMs, recommending fine-tuning on structured IR datasets and integration of explicit control flow models to augment their comprehension and handling of IR-related tasks.

Jiang, Hailong↗

DOE OSTI2025

Swap Path Network for Robust Person Search Pre-training

This code corresponds to the WACV25 conference paper, "Swap Path Network for Robust Person Search Pre-training". In that paper, we introduce a new model for the person search task called the Swap Path Net (SPNet). The person search task is a problem in computer vision, where we locate and rank matches to an image of a query person in a set of other images where we want to find them. We also introduce a novel pre-training algorithm specific to the Swap Path Net architecture. The code implements pre-training and fine-tuning of the Swap Path Net (SPNet). This includes ingesting image datasets and updating the weights of the SPNet neural network to train it for the person search task. The repository contains code, configs, and instructions to reproduce all results from the paper.

Jaffe, LucasW [Lawrence Livermore National Laborat↗

DOE OSTI2025

A Deep State Space Model for Rainfall‐Runoff Simulations

The classical way of studying the rainfall‐runoff processes in the water cycle relies on conceptual or physically‐based hydrologic models. Deep learning (DL) has recently emerged as an alternative and blossomed in the hydrology community for rainfall‐runoff simulations. However, the decades‐old Long Short‐Term Memory (LSTM) network remains the benchmark for this task, outperforming newer architectures like Transformers. In this work, we propose a State Space Model (SSM), specifically the Frequency Tuned Diagonal State Space Sequence (S4D‐FT) model, for rainfall‐runoff simulations. The proposed S4D‐FT is benchmarked against the established LSTM and a physically‐based Sacramento Soil Moisture Accounting model under in‐sample and out‐of‐sample simulation setups across 531 watersheds in the contiguous United States (CONUS). Results show that S4D‐FT is able to outperform the LSTM model across diverse regions under both simulation setups, especially for regions that feature snowmelt‐driven or intermittent flow regimes. In contrast, S4D‐FT tends to underperform in flashier, high‐magnitude flow regimes, likely due to its global state‐space convolution computation that emphasizes slow, storage‐driven dynamics, which makes it less effective at picking up short bursts and noisy spikes in the data. In summary, our pioneering introduction of the S4D‐FT for rainfall‐runoff simulations challenges the dominance of LSTM in the hydrology community and expands the arsenal of DL tools available for hydrological modeling.

Wang, Yihan [Univ. of Oklahoma, Norman, OK (United↗

DOE OSTI2025

Design of Hopfield Networks Based on Superconducting Coupled Oscillators

The global energy shortage has driven the development of many energy-efficient computational platforms beyond Moore's law, among which brain-inspired neuromorphic computing is one of the promising solutions. Associative memory and pattern recognition are important computations solved by brain-inspired Hopfield networks. Classical Hopfield networks store memories via fixed point attractors of their dynamics. In oscillatory Hopfield networks, these attractors are replaced by periodic orbits. Here, we design an oscillatory Hopfield network based on coupled superconducting oscillators. We first employ a mathematical phase reduction approach to map networks of coupled superconducting rapid single flux quantum (RSFQ) ring oscillators to coupled Kuramoto phase-oscillator networks. We use this theory to numerically optimize the hardware's mutual inductances in order to directly match the phase-reduced superconducting oscillators to a model of phase-oscillator-based Hopfield networks. The resulting network can store multiple oscillatory phase-locked memory patterns and recover the patterns based on the initial phase conditions. As different pattern recognition tasks, or learning, require tunable connectivity strengths between the oscillatory nodes, we further employ a coupler circuit that enables tuning the coupling strength between two oscillators by applying an external flux. We demonstrate the functionality of our design through numerical simulations of a small example network with oscillators operating at 86 GHz and recognizing patterns within 10 ns. Our approach enables the learning and retrieval of dynamical memory patterns with a wide range of applications where rhythmic dynamic output is beneficial.

Cheng, Ran↗

NASA NTRS2000

Acousto-Optic Imaging Spectrometers for Mars Surface Science

NASA's long term plan for Mars sample collection and return requires a highly streamlined approach for spectrally characterizing a landing site, documenting the mineralogical make-up of the site and guiding the collections of samples which represent the diversity of the site. Ideally, image data should be acquired at hundreds of VIS and IR wavelengths, in order to separately distinguish numerous anticipated species, using principal component analysis and linear unmixing. Cameras with bore-sighted point spectrometers can acquire spectra of isolated scene elements, but it requires 10(exp 2) to 10(exp 2) successive motions and precise relative pointing knowledge in order to create a single data cube which qualifies as a spectral map. These and other competing science objectives have to be accomplished within very short lander/rover operational lifetime (a few sols). True, 2-D imaging spectroscopy greatly speeds up the data acquisition process, since the spectra of all pixels in the scene are collected at once. This task can be accomplished with cameras that use electronically tunable acousto-optic tunable filters (AOTFs) as the optical tuning element. AOTFs made from TeO2 are now a mature technology, and operate at wavelengths from near-UV to about 5 microns. Because of incremental improvements in the last few years, present generation devices are rugged, radiation-hard and operate at temperatures down to at least 150K so they can be safely integrated into the ambient temperature optics of in-situ instruments such as planetary or small-body landers. They have been used for ground-based astronomy, and were also baselined for the ST-4 Champollion IR comet lander experiment (CIRCLE), prior to cancellation of the ST-4 mission last year. AIMS (for Acousto-optic Imaging spectrometer), is a prototype lander instrument which is being built at GSFC with support by the NASA OSS Advanced Technologies and Mission Studies, Mars Instrument Definition and Development Program (MIDP). AIMS is capable of tunable spectroscopic imaging of surface mineralogy, ices and dust between 0.5 and 2.4 microns, at a resolving power (lambda/delta lambda) which is typically several hundred. The design spatial resolution, similar to IMP and SSI, will allow mapping at scales down to about 1 cm.

Glenar, D. A.↗