Search NASA⌕ Search

SEARCH · Search NASA

Results for “adjust”

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 361 records · Page 20
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Adaptive Online Multivariate Signal Extraction With Locally Weighted Robust Polynomial Regression

High-frequency, multivariate data collected in real-time and used to control or make decisions regarding a process’ operation often contain some noise and outliers. Thus, a method to extract the signal is needed in order to reduce the number and magnitude of control-based adjustments that are implemented. Such a method must be (i) online, depending only on past and current observations; (ii) fast, producing a smooth value more quickly than the measurement frequency; (iii) robust, ignoring brief bursts of erroneously measured values; (iv) multivariate, ignoring observations that are jointly unusual; (v) adaptive, adjusting to periods of rapid fluctuation in the signal versus periods of stability; and (vi) purely data-driven, not incorporating any information about the process from which the data are collected. Most existing methods are only able to address a subset of these six features. Furthermore, we also require the method to be nonlinear, providing a local nonlinear estimate of the signal. In this work, we propose a novel, real-time signal extraction method based on a local, robust polynomial fit. We demonstrate the performance of our method compared to a state-of-the-art competitor through simulation. For illustration, the methodology is applied to data collected from a reverse osmosis water treatment process.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Selection of a Pair of Experiments to Optimally Reduce Uncertainty in Targeted Nuclear Data

We propose a novel process to select a pair of differential and integral experiments that best reduce uncertainties in targeted 239 ⁢Pu nuclear data while compressing the current nuclear data pipeline from 20 to 3 years. 239⁢ Pu nuclear data are poorly understood for neutrons in the intermediate energy range due to sparsity and uncertainty in historical experiments. New experiments targeting this range will enable better understanding of these nuclear data, but choosing the ideal experiments to conduct is challenging. Beginning with a prior distribution represented by samples of nuclear data generated from theory, generalized least squares adjustments are made to incorporate data from historical experiments. To quantify potential uncertainty reduction obtainable from a pair of candidate experiments, we compute the D-optimality criterion of the posterior covariance of intermediate energy range nuclear data compared to the equivalent covariance after additional adjustment to the pair of candidate experiments. Repeating the process for each of many candidate pairs facilitates the final selection. Results support 63⁢ Cu total cross section measurements for differential experiments and alumina and alumina/graphite configurations for integral experiments. This analysis enables choosing differential and integral experiments to be executed concurrently while shortening decision times relative to the current nuclear data pipeline.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2020

Tailoring SnO 2 , (Mg,Zn)O, and Ga:(Mg,Zn)O electro-optical properties and stability for solar cells

The electron density, mobility, bandgap, and band alignment of transparent conducting oxides (TCOs) can be tailored by adjusting composition and stoichiometry, thereby enabling interface engineering for diverse semiconductor applications. For example, solar cell efficiency can change enormously by adjusting TCO properties. At the same time, these TCO properties can shift during the deposition of other layers, anneals, and device operation. An ideal TCO should have tunable but stable electro-optical properties. Here, we deposit SnO 2 , (Mg,Zn)O (MZO), and Ga:(Mg,Zn)O (GMZO) films on glass and measure electro-optical characteristics before and after reducing, inert, oxidizing, and CdCl2 anneals over a range of temperatures. Electron density generally increases in the progression from oxidizing to inert and reducing ambients. SnO 2 is relatively stable compared to MZO but has less flexibility for interface engineering. We investigate GMZO as a similar but more stable alternative to MZO.The addition of Ga to MZO has significant effects on electron density and improves electro-optical stability, which can be advantageous for semiconductor applications. Furthermore, we demonstrate that GMZO can be readily incorporated into solar cells.

14 SOLAR ENERGY↗
DOE OSTI2023

Controlling the charge of dust particles in a plasma afterglow by timed switching of an electrode voltage

Abstract A method is demonstrated for controlling the charge of a dust particle in a plasma afterglow, allowing a wider range of outcomes than an earlier method. As in the earlier method, the dust particles are located near an electrode that has a DC voltage during the afterglow. Here, that DC voltage is switched to a positive value at a specified delay time, instead of maintaining a constant negative voltage as in the earlier method. Adjusting the timing of this switching allows one to control the residual charge gradually over a wide range that includes both negative and positive values of charge. For comparison, only positive residual charges were attained in the earlier method. We were able to adjust the residual charge from about −2000 e to +10 000 e , for our experimental parameters (8.35 µ m particles, 8 mTorr argon pressure, and a DC voltage that was switched from −150 V to +125 V within the first two milliseconds of the afterglow). The plasma conditions near the dust particles changed from ion-rich to electron-rich, when the electrode was switched from cathodic to anodic. Making this change at a specified time, as the electrons and ions decay in the afterglow, provides this control capability. These results also give insight into the time development of a dust particle’s charge in the afterglow, on a sub-millisecond time scale.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2021

Isotope effects on transport in LHD

Abstract Isotope effects are one of the most important issues for predicting future reactor operations. Large helical device (LHD) is the presently working largest stellarator/helical device using super conducting helical coils. In LHD, deuterium experiments started in 2017. Extensive studies regarding isotope effects on transport have been carried out. In this paper, the results of isotope effect studies in LHD are reported. The systematic studies were performed adjusting operational parameters and nondimensional parameters. In L mode like normal confinement plasma, where internal and edge transport barriers are not formed, the scaling of global energy confinement time ( τ E ) with operational parameters shows positive mass dependence ( M 0.27 ; where M is effective ion mass) in electron cyclotron heating plasma and no mass dependence ( M 0.0 ) in neutral beam injection heating plasma. The non-negative ion mass dependence is anti-gyro-Bohm scaling. The role of the turbulence in isotope effects was also found by turbulence measurements and gyrokinetic simulation. Better accessibility to electron and ion internal transport barrier (ITB) plasma is found in deuterium (D) plasma than in hydrogen (H). Gyro kinetic non-linear simulation shows reduced ion heat flux due to the larger generation of zonal flow in deuterium plasma. Peaked carbon density profile plays a prominent role in reducing ion energy transport in ITB plasma. This is evident only in plasma with deuterium ions. New findings on the mixing and non-mixing states of D and H particle transports are reported. In the mixing state, ion particle diffusivities are higher than electron particle diffusivities and D and H ion density profiles are almost identical. In the non-mixing state, ion particle diffusivity is much lower than electron diffusivity. Deuterium and hydrogen ion profiles are clearly different. Different turbulence structures were found in the mixing and non-mixing states suggesting different turbulence modes play a role.

Physics↗
DOE OSTI2022

Stochastic and a posteriori optimization to mitigate coil manufacturing errors in stellarator design

Abstract It was recently shown in Wechsung et al (2022 Proc. Natl Acad. Sci. USA 119 e2202084119) that there exist electromagnetic coils that generate magnetic fields, which are excellent approximations to quasi-symmetric fields and have very good particle confinement properties. Using a Gaussian process-based model for coil perturbations, we investigate the impact of manufacturing errors on the performance of these coils. We show that even fairly small errors result in noticeable performance degradation. While stochastic optimization yields minor improvements, it is not possible to mitigate these errors significantly. As an alternative to stochastic optimization, we then formulate a new optimization problem for computing optimal adjustments of the coil positions and currents without changing the shapes of the coil. These a-posteriori adjustments are able to reduce the impact of coil errors by an order of magnitude, providing a new perspective for dealing with manufacturing tolerances in stellarator design.

Physics↗
DOE OSTI2024

Characterizing the negative triangularity reactor core operating space with integrated modeling

Negative triangularity (NT) has received renewed interest as a fusion reactor regime due to its beneficial power-handling properties, including low scrape-off layer power and a larger divertor wetted area that facilitates simple divertor integration. NT experiments have also demonstrated core performance on par with positive triangularity (PT) high confinement mode (H-mode) without edge-localized modes (ELMs), encouraging further study of an NT reactor core. In this work, we use integrated modeling to scope the operating space around two NT reactor strategies. The first is the high-field, compact fusion pilot plant concept Modular, Adjustable, NT ARC (MANTA) and the second is a low field, high aspect ratio concept based on work by Medvedev et al. By integrating equilibrium, core transport, and edge ballooning instability models, we establish a range of operating points with less than 50 MW scrape-off layer power and fusion power comparable to PT H-mode reactor concepts. Heating and seeded impurities are leveraged to accomplish the same fusion performance and scrape-off layer exhaust power for various pressure edge boundary conditions. Scans over these pressure edge conditions accommodate any current uncertainty of the properties of the NT edge and show that the performance of an NT reactor will be extremely dependent on the edge pressure. The high-field case is found to enable lower scrape-off layer power because it is capable of reaching high fusion powers at a relatively compact size, which allows increased separatrix density without exceeding the Greenwald density limit. Adjustments in NT shaping exhibit small changes in fusion power, with an increase in fusion power density seen at weaker NT. Infinite-n ballooning instability models indicate that an NT reactor core can reach fusion powers comparable to leading PT H-mode reactor concepts while remaining ballooning-stable. Seeded krypton is leveraged to further lower scrape-off layer power since NT does not have a requirement to remain in H-mode while still maintaining high confinement. We contextualize the NT reactor operating space by comparing to popular PT H-mode reactor concepts, and find that NT exhibits competitive ELM-free performance with these concepts for a variety of edge conditions while maintaining relatively low scrape-off layer power.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2022

A new calibration method for charm jet identification validated with proton-proton collision events at $\sqrt{s}$ =13 TeV

Many measurements at the LHC require efficient identification of heavy-flavour jets, i.e. jets originating from bottom (b) or charm (c) quarks. An overview of the algorithms used to identify c jets is described and a novel method to calibrate them is presented. This new method adjusts the entire distributions of the outputs obtained when the algorithms are applied to jets of different flavours. It is based on an iterative approach exploiting three distinct control regions that are enriched with either b jets, c jets, or light-flavour and gluon jets. Results are presented in the form of correction factors evaluated using proton-proton collision data with an integrated luminosity of 41.5 fb$^{-1}$ at √s = 13 TeV, collected by the CMS experiment in 2017. The closure of the method is tested by applying the measured correction factors on simulated data sets and checking the agreement between the adjusted simulation and collision data. Furthermore, a validation is performed by testing the method on pseudodata, which emulate various mismodelling conditions. The calibrated results enable the use of the full distributions of heavy-flavour identification algorithm outputs, e.g. as inputs to machine-learning models. Thus, they are expected to increase the sensitivity of future physics analyses.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2023

A high-resolution clock phase shifter circuitry for ALTIROC

Abstract A high-resolution clock phase shifter is implemented to adjust the phase of multiple clocks at 40 MHz, 80 MHz, or 640 MHz in the ALTIROC chip. The phase shifter has a coarse-phase shifter and a fine-phase shifter to achieve a step size of 97.7 ps and an adjustable range of 25 ns. The fine delay unit is based on a Delay Locked Loop (DLL) operating at 640 MHz. The phase shifter is fabricated in a 130 nm CMOS process. The area of the phase shifter is 725 µm × 248 µm. The Differential Non-Linearity (DNL) and the Integral Non-Linearity (INL) are ±0.6 LSB and ±0.75 LSB, respectively. The jitter from −25 °C to 20 °C is less than 15.5 ps (RMS), including the contributions from the FPGA clock source and the PLL. The power consumption is 11.2 mW.

Instruments & Instrumentation↗
DOE OSTI2025

Improved heavy-ion PID using scintillation light detector with neural network analysis: a Monte Carlo simulation study

The photon collection efficiency of gaseous scintillator detectors varies according to the position of the impinging charged particles in the medium that generates scintillation light. Thus, when impinging particles are distributed over a large area, the intrinsic photon-number resolution of the system is affected by a large variation. This work presents and discusses a method for adjusting the total number of detected photons to account for variation in the photon collection efficiency as a function of the position of the light source within the scintillating medium. The method was developed and validated by processing data from systematic simulation studies based on GEANT4 that model the response of the Energy Loss Optical Scintillation System (ELOSS) detector. The position of the charged particle is calculated using a deep neural network algorithm. This is accomplished by analyzing the distribution of scintillation light recorded by the array of photosensors. The estimated particle position is then used to calculate the correction factor and adjust the amount of captured light to account for variations in the photon collection efficiency. The neural network algorithm provides excellent tracking capabilities, achieving sub-millimeter position resolution and an angular resolution of 12 mrad, approaching the performance of traditional tracking detectors (e.g., drift chambers). The present method can be generalized to any optical scintillation system where the photon collection efficiency depends on the position of the impinging particle.

Heavy-ion detectors↗
DOE OSTI2023

Ion microscopy with evolutionary-algorithm-based autofocusing

Abstract Ion microscopy is an established technique for laser focus diagnostics and the accurate, intensity-resolved measurement of laser ionization processes. In the present feasibility study, we discuss a new ion microscope design, which improves its resolution across a large range of magnifications and simplifies its operation. Instead of the common two einzel lens configuration, which is usually optimized for a fixed magnification, we propose a generic design consisting of an array of equally spaced ring electrodes, whose individually adjustable voltages are controlled by an evolutionary algorithm. In this way, we can realize aberration minimized magnifications between 25 and 100. Moreover, the algorithm can adjust the voltage settings under changing experimental conditions and facilitates autofocusing for user-defined magnification.

Haniel, Franz E.↗
DOE OSTI2022

Diabetes and hypertension are related to amyloid-beta burden in the population-based Rotterdam Study

Abstract Higher vascular disease burden increases the likelihood of developing dementia, including Alzheimer’s disease. Better understanding the association between vascular risk factors and Alzheimer’s disease pathology at the predementia stage is critical for developing effective strategies to delay cognitive decline. In this work, we estimated the impact of six vascular risk factors on the presence and severity of in vivo measured brain amyloid-beta (Aβ) plaques in participants from the population-based Rotterdam Study. Vascular risk factors (hypertension, hypercholesterolaemia, diabetes, obesity, physical inactivity and smoking) were assessed 13 (2004–2008) and 7 years (2009–2014) prior to 18F-florbetaben PET (2018–2021) in 635 dementia-free participants. Vascular risk factors were associated with binary amyloid PET status or continuous PET readouts (standard uptake value ratios, SUVrs) using logistic and linear regression models, respectively, adjusted for age, sex, education, APOE4 risk allele count and time between vascular risk and PET assessment. Participants’ mean age at time of amyloid PET was 69 years (range: 60–90), 325 (51.2%) were women and 190 (29.9%) carried at least one APOE4 risk allele. The adjusted prevalence estimates of an amyloid-positive PET status markedly increased with age [12.8% (95% CI 11.6; 14) in 60–69 years versus 35% (36; 40.8) in 80–89 years age groups] and APOE4 allele count [9.7% (8.8; 10.6) in non-carriers versus 38.4% (36; 40.8) to 60.4% (54; 66.8) in carriers of one or two risk allele(s)]. Diabetes 7 years prior to PET assessment was associated with a higher risk of a positive amyloid status [odds ratio (95% CI) = 3.68 (1.76; 7.61), P < 0.001] and higher standard uptake value ratios, indicating more severe Aβ pathology [standardized beta = 0.40 (0.17; 0.64), P = 0.001]. Hypertension was associated with higher SUVr values in APOE4 carriers (mean SUVr difference of 0.09), but not in non-carriers (mean SUVr difference 0.02; P = 0.005). In contrast, hypercholesterolaemia was related to lower SUVr values in APOE4 carriers (mean SUVr difference −0.06), but not in non-carriers (mean SUVr difference 0.02). Obesity, physical inactivity and smoking were not related to amyloid PET measures. The current findings suggest a contribution of diabetes, hypertension and hypercholesterolaemia to the pathophysiology of Alzheimer’s disease in a general population of older non-demented adults. As these conditions respond well to lifestyle modification and drug treatment, further research should focus on the preventative effect of early risk management on the development of Alzheimer’s disease neuropathology.

van Arendonk, Joyce (ORCID:0000000172941633)↗
DOE OSTI2020

Benefits of colorectal cancer screening using fecal immunochemical testing with varying positivity thresholds by age and sex

Background Fecal immunochemical test (FIT) performance for colorectal cancer screening varies by age and sex, yet most FIT-based screening programs use uniform positivity thresholds. This study assessed the potential benefits of stratifying FIT thresholds based on age and sex. Methods We conducted a meta-analysis of FIT sensitivity and specificity at various positivity thresholds by age and sex. We then used these estimates in 2 microsimulation models of colorectal cancer and projected lifetime clinical outcomes, incremental costs, and quality-adjusted life-years (QALYs) gained from age- and sex-stratified FIT strategies. FIT thresholds ranged from 10 to 50 µg hemoglobin per gram of feces. Results For current uniform FIT screening (20 µg hemoglobin/gram of feces), models projected 85.67 to 122.15 QALYs gained at incremental costs of ‒$982 to $504 per 1000 individuals compared with no screening. At equivalent costs to current uniform screening, only 1 model found stratified FIT approaches cost-effective, yielding a marginal increase of 1.04 and 1.10 QALYs gained/1000 female and male individuals, respectively. At a willingness-to-pay threshold of $100 000/QALYs gained, both models found stratified FIT cutoffs to be the best strategy, with cutoffs being equal to or higher for males and lowest at older ages (70-75 years). Uniform strategies showed comparable effectiveness, falling within 1 quality-adjusted life-day per person of efficient strategies at up to $112 more per person. Results were sensitive to FIT test performance characteristics and 1-time setup costs. Conclusion Stratifying FIT thresholds by age and sex may be cost-effective compared to current screening. The gain in expected health benefits with stratified FIT screening, however, is likely small.

Harlass, Matthias↗
DOE OSTI2022

Interannual consistency of migration phenology is season- and breeding region-specific in North American Golden Eagles

Abstract Interannual consistency (an indicator of the strength of adjustments) in migration phenology of Golden Eagles (Aquila chrysaetos) in North America is most strongly associated with the breeding region, the season, and with late-season temperature on breeding and wintering grounds. Consistency was greatest in boreal spring migration and the breeding regions of eastern Canada. Using multi-year GPS tracks of 83 adults breeding in 3 spatially distant regions (Alaska, northeast Canada, and southeast Canada), we quantified the interannual consistency of migration phenology and wintering latitude within and among individuals tracked across multiple years and the repeatability (r) by breeding regions and seasons. By comparing regions and seasons, we found that consistency was highest (r > 0.85) for boreal spring migration in eastern Canada while Alaska had the lowest value (r < 0.15). Because seasonal consistency of migration phenology was only detected in eastern Canada, we conclude that seasonal features are not a primary constraint. While regional differences in consistency were not related to differences in migratory distances, they could be the result of genetic or habitat differences. We also found that temperatures warmer than the decadal average at the region of departure delayed the start of boreal spring migration by ~10 days and advanced boreal autumn migration by ~20 days. These results suggest that warmer temperatures would reduce residence time on breeding grounds, contrary to expectations and trends found in other studies. Wide variations in migratory strategies across a species distribution can add to the list of challenges for conservation but may give migrants the capacity to acclimate to environmental changes.

Zoology↗
DOE OSTI2022

Tonoplast Sucrose Trafficking Modulates Starch Utilization and Water Deficit Behavior in Poplar Leaves

Leaf osmotic adjustment by the active accrual of compatible organic solutes (e.g. sucrose) contributes to drought tolerance throughout the plant kingdom. In Populus tremula x alba, PtaSUT4 encodes a tonoplast sucrose–proton symporter, whose downregulation by chronic mild drought or transgenic manipulation is known to increase leaf sucrose and turgor. While this may constitute a single drought tolerance mechanism, we now report that other adjustments which can occur during a worsening water deficit are damped when PtaSUT4 is constitutively downregulated. Specifically, we report that starch use and leaf relative water content (RWC) dynamics were compromised when plants with constitutively downregulated PtaSUT4 were subjected to a water deficit. Leaf RWC decreased more in wild-type and vector control lines than in transgenic PtaSUT4-RNAi (RNA-interference) or CRISPR (clustered regularly interspersed short palindromic repeats) knockout (KO) lines. The control line RWC decrease was accompanied by increased PtaSUT4 transcript levels and a mobilization of sucrose from the mesophyll-enriched leaf lamina into the midvein. The findings suggest that changes in SUT4 expression can increase turgor or decrease RWC as different tolerance mechanisms to reduced water availability. Evidence is presented that PtaSUT4-mediated sucrose partitioning between the vacuole and the cytosol is important not only for overall sucrose abundance and turgor, but also for reactive oxygen species (ROS) and antioxidant dynamics. Interestingly, the reduced capacity for accelerated starch breakdown under worsening water-deficit conditions was correlated with reduced ROS in the RNAi and KO lines. A role for PtaSUT4 in the orchestration of ROS, antioxidant, starch utilization and RWC dynamics during water stress and its importance in trees especially, with their high hydraulic resistances, is considered.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Evaluating the impact of anatomical and physiological variability on human equivalent doses using PBPK models

Abstract Addressing human anatomical and physiological variability is a crucial component of human health risk assessment of chemicals. Experts have recommended probabilistic chemical risk assessment paradigms in which distributional adjustment factors are used to account for various sources of uncertainty and variability, including variability in the pharmacokinetic behavior of a given substance in different humans. In practice, convenient assumptions about the distribution forms of adjustment factors and human equivalent doses (HEDs) are often used. Parameters such as tissue volumes and blood flows are likewise often assumed to be lognormally or normally distributed without evaluating empirical data for consistency with these forms. In this work, we performed dosimetric extrapolations using physiologically based pharmacokinetic (PBPK) models for dichloromethane (DCM) and chloroform that incorporate uncertainty and variability to determine if the HEDs associated with such extrapolations are approximately lognormal and how they depend on the underlying distribution shapes chosen to represent model parameters. We accounted for uncertainty and variability in PBPK model parameters by randomly drawing their values from a variety of distribution types. We then performed reverse dosimetry to calculate HEDs based on animal points of departure for each set of sampled parameters. Corresponding samples of HEDs were tested to determine the impact of input parameter distributions on their central tendencies, extreme percentiles, and degree of conformance to lognormality. This work demonstrates that the measurable attributes of human variability should be considered more carefully and that generalized assumptions about parameter distribution shapes may lead to inaccurate estimates of extreme percentiles of HEDs.

Toxicology↗
DOE OSTI2025

Closing the Accuracy Gap in Tandem Photovoltaic Testing: An Accessible and Efficient Spectral Tuning Method Using LED-Based Simulators for Research Laboratories and Industry

Accurate performance calibration of multijunction (MJ) solar cells is critical for advancing this technology toward large-scale terrestrial application, yet existing testing methods developed by photovoltaics calibration laboratories remain prohibitively complex and resource intensive for most research laboratories. Current approaches rely on expensive multisource simulators and/or intricate spectral tuning algorithms, limiting accessibility and hindering standardized comparisons of emerging MJ technologies such as perovskite-based tandems. This paper introduces a streamlined spectral irradiance adjustment method for light-emitting diode (LED)-based solar simulators, which are increasingly adopted in the photovoltaics community due to their versatility and cost-effectiveness. The method we present bins LED channels into groups based on the number of junctions of the MJ photovoltaic device under test (DUT) and their corresponding band gaps and incorporates an automated tuning algorithm that eliminates the need to adjust each channel’s irradiance individually. This tuning algorithm requires the relative spectral responsivities of both the DUT and a broadband reference cell, as well as a calibrated spectroradiometer. Measurement validation across various MJ devices, including III-V and perovskite/silicon tandems, demonstrates excellent agreement within 1% with well-established xenon-tungsten multisource simulators and associated tuning algorithms. By enabling precise spectral tuning with readily available equipment and a simple tuning algorithm, our approach bridges the measurement accuracy gap between research laboratories and accredited testing facilities, fostering more reliable device comparisons and accelerating the translation of MJ technologies into real-world energy systems.

14 SOLAR ENERGY↗
DOE OSTI2022

Demonstration of Decentralized Physics-Driven Learning

In typical artificial neural networks, neurons adjust according to global calculations of a central processor, but in the brain, neurons and synapses self-adjust based on local information. Contrastive learning algorithms have recently been proposed to train physical systems, such as fluidic, mechanical, or electrical networks, to perform machine-learning tasks from local evolution rules. However, to date, such systems have only been implemented in silico due to the engineering challenge of creating elements that autonomously evolve based on their own response to two sets of global boundary conditions. Here, we introduce and implement a physics-driven contrastive learning scheme for a network of variable resistors, using circuitry to locally compare the response of two identical networks subjected to the two different sets of boundary conditions. Using this method, our system effectively trains itself, optimizing its resistance values without the use of a central processor or external information storage. Once the system is trained for a specified allostery, regression, or classification task, the task is subsequently performed rapidly and automatically by the physical imperative to minimize power dissipation in response to the given voltage inputs. Here, we demonstrate that, unlike typical computers, such learning systems are robust to extreme damage (and thus manufacturing defects) due to their decentralized learning. Our twin-network approach is therefore readily scalable to extremely large or nonlinear networks, where its distributed nature will be an enormous advantage; a laboratory network of only 500 edges will already outpace its in silico counterpart.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗