SEARCH · Search NASA

Results for “sensitivity estimation”

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 343 records · Page 19

All sources NASA NTRS DOE OSTI

DOE OSTI2026

Meta-analysis of North American Arctic and boreal aboveground biomass datasets: assessing accuracy, dynamics, and similarities

The North American arctic and boreal regions (ABRs) are rapidly warming and experiencing intensifying disturbances. Accurately quantifying aboveground biomass (AGB) is critical for understanding the impacts of these changes on the carbon cycle and for designing climate change mitigation strategies. Several AGB maps have been developed for the North American ABRs, including recent contributions from National Aeronautics and Space Administration’s Arctic-Boreal Vulnerability Experiment (ABoVE) campaign. However, these maps differ widely in training data, methodology, and resulting AGB density estimates. Presently, a comprehensive comparative evaluation is lacking, making it difficult for users to select datasets suited to their research or management needs. Here, in this study, we conducted a comparative analysis of nine AGB density datasets across North American ABRs, specifically for Alaska and Canada. We (1) summarized AGB by ecoregion and Canadian provinces, (2) evaluated their accuracy against field-based measurements, (3) analyzed spatial and temporal similarities among datasets, and (4) assessed their ability to capture disturbance (fire and harvest) impacts on AGB. We found substantial variation in regional and local AGB estimates across datasets, with overall accuracy ranging from R 2 = 0.25–0.62 and Bias% from −47.8% to 69.9% when validated against field plots. Despite these differences, most datasets have comparatively consistent spatial patterns in AGB (r > 0.8 for most cases). In contrast, agreement on the temporal patterns of AGB change is generally low. We found datasets with spatial resolutions ⩽300 m are capable of capturing disturbance impacts on AGB dynamics, though sensitivity varies across products. Our findings and dataset summary provide guidance for selecting appropriate AGB datasets for different applications within our study area. Our analysis also highlights the need to decrease map bias and increase capability to detect temporal change to decrease uncertainty of AGB datasets potentially by using training data which is representative of major plant functional types within the mapped area.

ABoVE↗

DOE OSTI2026

Regression Convolutional Neural Network for Energy Estimation in NOvA

Regression Convolutional Neural Network for Energy Estimation in NOvA" Abstract: "NOvA (NuMI Off-Axis $\nu_e$ Appearance) is a long baseline neutrino experiment designed to measure neutrino oscillations over a distance of 810 km. NOvA employs a near and far detector to observe $\nu_\mu$ disappearance and $\nu_e$ appearance of neutrinos produced by the NuMI beam at Fermilab. Energy reconstruction is critical for precise measurements of neutrino oscillation parameters and cross sections, which are functions of neutrino energy. Energy estimation remains difficult due to the complexity of detector response and final state particle kinematics. We present a regression-based convolutional neural network (CNN) method that reconstructs neutrino and lepton energies based on raw pixel inputs for NOvA. The trained model is able to reconstruct event energy for different interaction modes and complex final states containing leptons and hadrons. Studies of regression CNN networks show improved energy resolution and reduced sensitivity to calibration scale uncertainties relative to traditional kinematics-based energy reconstruction techniques. The results demonstrate the potential of the regression CNN method for neutrino physics analyses by improving on standard kinematics-based reconstruction.

Zhao, Larry [UC, Irvine (main)]↗

DOE OSTI2024

A multiscale model to understand the interface chemistry, contacts, and dynamics during lithium stripping

A reversible Li-metal electrode, paired with a solid electrolyte, is critical for attaining higher energy density and safer batteries beyond the current lithium-ion cells. A stable stripping process may be even harder to attain as the stripping process will remove Li-atoms from the surface, and naturally reduce surface contact area, if not self-corrected by other mechanisms, such as diffusion and plastic deformation under an applied external stack pressure. Here, we capture these mechanisms occurring at multiple length- and time- scales, i.e., interface interactions, vacancy hopping, and plastic deformation, by integrating density functional theory (DFT) simulations, kinetic Monte Carlo (KMC), and continuum finite element method (FEM). By assuming the self-affine nature of multiscale contacts, we predict the steady-state contact area as a function of stripping current density, interface wettability, and stack pressure. We further estimate the exponential increase of overpotential due to contact area loss to maintain the same stripping current density. We demonstrate that a lithiophilic interface requires less stack pressure to reach the same steady-state contact area fraction than a lithiophobic interface. A “tolerable steady-state” contact area loss for maintaining stable stripping is estimated at 20 %, corresponding to a 10 % increase in overpotential. To constrain contact loss within the tolerance, the required stack pressure is 0.1, 0.5, and 2 times the yield strength of lithium metal for three distinct interfaces, lithiophilic Li/lithium oxide(Li2O), Li/lithium lanthanum zirconium oxide(LLZO), and lithiophoblic Li/lithium fluoride(LiF), respectively. The modeling results agree with experiments on the impact of the stack pressure quantitatively, while the discrepancy in stripping rate sensitivity is attributed to the simplifying interface interaction in our simulations. Overall, this multiscale simulation framework demonstrates the importance of electrochemical-mechanical coupling in understanding the dynamics of the Li/SE interface during stripping.

Feng, Min↗

DOE OSTI2026

Overview of the Nephele Perturbed Parameter Ensemble for Aerosol‐Cloud Interactions in E3SMv3

Aerosol-cloud interactions (aci) are the leading source of uncertainty in inferring climate sensitivity from the historical record. Earth system models (ESMs) struggle to represent aci because the processes responsible for these phenomena occur at much finer time and space scales than can be resolved by any ESM. Observational constraints provide key benchmarks to test ESMs, but cannot be used alone to fully understand aci processes except in very specific cases where causality is controlled; some degree of modeling is required to infer aci and estimate radiative forcing. Here, we generate and characterize a perturbed parameter ensemble (PPE) in version 3 of the Energy Exascale ESM (E3SMv3). We perturb 25 parameters that govern aci processes over 250 members and integrate the model over present-day and preindustrial aerosol emissions. We find that the process representation in E3SMv3 is flexible and can generate global-mean effective radiative forcings due to aci (ERFaci) ranging from −3.0 to +0.9 W m −2 . The positive ERFaci values simulated by a portion of the PPE are implausible and result from parameter combinations that produce unrealistic top-of-atmosphere energy fluxes. While global-mean cloud droplet number concentration always increases in response to anthropogenic aerosol, cloud liquid water path can both increase and decrease, suggesting that precipitation suppression is not the only aerosol-cloud adjustment represented by E3SMv3. Analysis of which processes control liquid cloud adjustment in the PPE points toward stratiform precipitation processes and aerosol activation, which is consistent with many previous ESMs, as well as the new two-moment convective cloud microphysics in E3SMv3.

Nugent, Jacqueline M. [Univ. of Wyoming, Laramie, ↗

DOE OSTI2025

Examination of Factors Affecting the Cost and Performance of a Natural Gas Combined Cycle Equipped with Carbon Dioxide Capture

The purpose of this Technical Note is to report the findings of an examination of the effect of plausible deviations in select study assumptions on the reported cost and performance estimates for a power plant case drawn from NETL’s “Cost and Performance Baseline for Fossil Energy Plants Volume 1: Bituminous Coal and Natural Gas to Electricity” (known as the Fossil Energy Baseline). An F-Class NGCC power plant equipped with state-of-the-art, solvent-based, post-combustion carbon dioxide (CO2) capture (95 percent carbon capture rate)—designated as Case B31B.95—was selected for this work. This sensitivity analysis provides insight into the effects of parameter variations within and across selected categories—ambient conditions, construction cost, natural gas (NG) price, capacity factor, and finance—on the plant performance and capital and operating and maintenance (O&M) costs, and the subsequent impact on common figures of merit.

20 FOSSIL-FUELED POWER PLANTS↗

DOE OSTI2025

Demonstration of Optimal Benchmark Selection Website and Validation of the q c Coverage Metric Using HEU-SOL-THERM-013-003 Experiment

In the work documented in this interim report, the experiment selection toolkit web site was demonstrated and q C coverage metric methodology was validated for IEU-MET-FAST-002-001, MIX-COMP-THERM 004-004, and HEU-SOL-THERM-013-003 experiments. 𝑞 𝐶 is an information-theoretic measure based on mutual information that quantifies the ability of candidate benchmark experiments to reduce the bias and uncertainty of a target criticality safety application. The metric and an accompanying open-source Python toolkit with a web-based interface were tested against a benchmark set of 425 experiments drawn from the International Criticality Safety Benchmark Evaluation Project Handbook. The interface is hosted at https://edim.covdef.com. It accepts sensitivity data files produced by the TSUNAMI-IP module of the SCALE code system and supports both (i) deterministic analysis using the ENDF/B-VII.0 covariance library and (ii) stochastic analysis based on user-supplied keff samples. Demonstrations on representative applications across a range of material composition, spectrum, and form show that q C -guided benchmark selection achieves greater uncertainty reduction with fewer experiments and yields more stable posterior bias and uncertainty estimates than traditional similarity coefficient ( c k )–based selection, while also capturing valuable low-ck experiments that one-to-one metrics overlook.

Abdel-khalik, Hany S. [Indiana Univ.-Purdue Univ. ↗

DOE OSTI2025

Evaluation of cloud height, optical thickness, and phase retrievals from the CHROMA algorithm applied to Sentinel-3 OLCI data

We previously developed the Cloud Height Retrieval from O 2 Molecular Absorption (CHROMA) algorithm for the Ocean Color Instrument (OCI) on the new NASA Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) mission. Here, we apply CHROMA to observations from the Ocean Land Colour Instrument (OLCI) to guide expectations for PACE, as it will take some time to obtain large-scale validation data for OCI. We use cloud top height (CTH), phase, and (for liquid clouds) cloud optical thickness (COT) data from the ground-based Atmospheric Radiation Measurement (ARM) network to evaluate the OLCI retrievals. We found that OLCI and Moderate Resolution Imaging Spectroradiometer (MODIS) CTH compare similarly well to the ARM reference. OLCI has a tendency to underestimate CTH as CTH increases, and algorithm assumptions about cloud geometric thickness may contribute to this. ARM COT from multifilter shadowband radiometers (MFRSR) and Sun photometers are well-correlated with one another, albeit with a roughly 30 % offset on average; OLCI and MODIS COT agree more closely with the MFRSR data. OLCI retrieval uncertainty estimates show skill at telling low-uncertainty cases from high-uncertainty ones, although CTH uncertainties are underestimated. Additionally, we compare the OLCI data to satellite retrievals based on thermal infrared measurements from MODIS and Sea and Land Surface Temperature Radiometer (SLSTR) data. Differences are broadly consistent with physical expectations based on the A-band vs. thermal techniques, although one key challenge in such aggregated comparisons is different cloud masking sensitivities and algorithm failure rates meaning additional sampling differences are introduced. We conclude by discussing the transition to and possible enhancements for PACE OCI.

Sayer, Andrew M. [Univ. of Maryland Baltimore Coun↗

DOE OSTI2025

Connecting GRBs from binary neutron star mergers to nuclear properties of neutron stars

The fate of the binary neutron star (NS) merger remnants hinges sensitively upon the NS equation of state and the threshold mass, M ls , that separates a long-lived from a short-lived NS remnant. The nature of the electromagnetic counterparts is also influenced by the remnant type, particularly in determining whether a gamma-ray burst from a compact binary merger (cbGRB) is of short or long duration. We propose a novel approach to probe Mls by linking it to the estimated observed ratio of long to short cbGRBs. We find that current observations broadly favor a relatively high value for this transition, M ls ≃ 1.3M TOV , for which M TOV ≲ 2.6M ⊙ , consistent with numerical simulations, as also shown here. Furthermore, our results disfavor nuclear physics scenarios that would lead to catastrophic pressure loss at a few times nuclear density and temperatures of tens of MeV, leading to a rapid gravitational collapse of binaries with total mass M ≲ 1.3M TOV . Future individual gravitational wave events with on-axis cbGRBs can further bound Mls.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2026

Elemental analysis of air-sensitive frozen molten salt samples using an inert transfer chamber for LIBS/LA-ICP-TOF-MS analysis

A novel inert sample transfer system was developed and employed to enable, for the first time, the analysis of air-sensitive salt samples in a two-volume ablation cell using simultaneous laser-induced breakdown spectroscopy (LIBS) and laser ablation (LA)-inductively coupled plasma (ICP)-time-of-flight (TOF)-mass spectrometry (MS) analysis. Molten salts are of growing interest as a medium for advanced nuclear reactors and nuclear fuel reprocessing technologies continue to be developed around their use. However, compositional analysis of molten salt samples can be challenging because of their air-sensitive nature and varying solubilities leading to inaccurate measurements when digested. LA-based analysis provides an alternate method to digestion and can provide rapid elemental information with little sample preparation. In this study, LIBS and LA-ICP-TOF-MS were used to analyze the Ce content in frozen salt samples taken from a series of electrochemical experiments. Calibrations were built for each technique, and the resulting limits of detection for Ce were estimated to be 107 and 58 µg g −1 for LIBS and LA-ICP-TOF-MS, respectively. Test samples from the electrochemical experiments were analyzed using these calibrations. The results matched bulk digestion-based ICP-optical emission spectroscopy values, and daily trends in Ce concentration changes were identified. Additionally, the LIBS and LA-ICP-TOF-MS analysis was demonstrated for identifying microgram per gram levels of components and detecting trace contaminants. The impurities detected by LIBS included Al, Mg, Ca, and Na. The impurities detected by LA-ICP-TOF-MS included W, Ag, Al, Fe, Ni, Mo, Nd, Sm, Th, and U.

Andrews, Hunter B. [Oak Ridge National Laboratory ↗

DOE OSTI2026

Towards modeling phage therapy

Patients infected with life-threatening multi-drug resistant (MDR) bacteria have been treated with cocktails of bacteriophages. This is a complicated form of personalized medicine as the phages given to a patient have to be selected beforehand on the basis of their lytic capacity of the infecting bacteria. Because bacteria rapidly become resistant, the evolution of resistance to a diverse cocktail of phages is a complicated dynamical process, during which competing bacterial strains replace one another by accumulating several resistance mechanisms, each of which may involve a fitness cost. As a consequence, it is typically not known why a particular phage therapy succeeded or failed, and how one can optimize the composition of the cocktails to maximize the rate of success. To improve upon this, we extend an existing in vivo -calibrated mouse model into a novel mathematical model for the human situation, and include multiple phages infecting multiple bacterial strains, differing in their resistance to each of the phages. We adjust several parameter estimates of the bacterial model to the human situation, and use the model to describe a successful case of phage therapy involving several cocktails, each containing several phages. In the model, treatment success crucially depended on pretreatment resistance levels, and on the diversity and the timing of the cocktails. Once an appropriate cocktail is found, it is less important to further optimize the infection rates of the phages. Resistant bacterial strains expand rapidly when sensitive strains decline, and the higher the infectivity of the phages, the faster resistant strains expand. Because resistance evolves rapidly, it is best to provide a diverse set of phages right from the start of therapy, i.e., to hit hard and early, and create a high genetic barrier to bacterial resistance.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2024

Optimal Stopping Ages for Colorectal Cancer Screening

Importance Prior studies have shown that the benefits, harms, and costs of colorectal cancer (CRC) screening at older ages are associated with a patient’s sex, health, and screening history. However, these studies were hypothetical exercises and not directly informed by data on CRC risk. Objective To identify the optimal stopping ages for CRC screening by sex, comorbidity, and screening history from a cost-effectiveness perspective. Design, Setting, and Participants This economic evaluation first validated the MISCAN-Colon (Microsimulation Screening Analysis–Colon) model against community-based CRC incidence and mortality rates for 2 subcohorts of the PRECISE (Optimizing Colorectal Cancer Screening Precision and Outcomes in Community-Based Populations) cohort. Subsequently, different CRC screening scenarios were simulated in older individuals. Cohorts of US adults aged 76 to 90 years varied by sex and comorbidity status (none, low, moderate, or severe). Statistical and sensitivity analyses were performed from March 2023 to May 2024. Exposures CRC screening histories including fecal immunochemical test (FIT) or colonoscopy, such as a negative colonoscopy result from 10, 15, 20, 25, or 30 years before the index age; 1 to 5 negative FIT results within 5 years of the index age, with different patterns of recency; or a combination of negative colonoscopy and negative FIT results. Main Outcomes and Measures The main outcomes included estimated lifetime clinical outcomes, incremental costs, and quality-adjusted life-years gained (QALYG) associated with 1 additional FIT or colonoscopy. Optimal stopping age for screening, defined as the oldest age for which the incremental cost-effectiveness ratio was still below the willingness-to-pay threshold of $\$$100 000 per QALYG, was evaluated. Results The first of the 2 PRECISE subcohorts used in validating the simulation model included 25 974 adults (15 060 females [58.0%]; 54.7% aged 76 to 80 years) with a negative colonoscopy result 10 years before the index date. The second subcohort consisted of 118 269 adults (67 058 females [56.7%]; 90.5% aged 76 to 80 years) with a negative FIT result 1 year before the index date. Older age, male sex, higher comorbidity levels, and recent CRC screenings were associated with reduced incremental benefit and cost-effectiveness of additional screening. For the reference cohort of 76-year-old females without comorbidities and a negative colonoscopy result 10 years before the index age, 1 additional colonoscopy cost $\$$38 226 per QALYG. For cohorts with otherwise equivalent characteristics, associated costs increased to $\$$1 689 945 per QALYG for females at age 90 years without comorbidities and a negative colonoscopy results 10 years before the index age, $\$$51 604 per QALYG for males at age 76 years without comorbidities and a negative colonoscopy result 10 years before the index age, and $\$$108 480 per QALYG for females at age 76 years with severe comorbidities and a negative colonoscopy result 10 years before the index age and decreased to $\$$16 870 per QALYG for females without comorbidities and a negative colonoscopy result 30 years before the index age. The optimal stopping ages across different cohorts ranged from younger than 76 to 86 years for colonoscopy and younger than 76 to 88 years for FIT. Conclusions and Relevance In this economic evaluation, age, sex, screening history, comorbidity, and future screening modality were associated with the clinical outcomes, cost-effectiveness, and optimal stopping age for CRC screening. These results can inform guideline development and patient-directed informed decision-making.

Harlass, Matthias [Erasmus Erasmus University Medi↗

DOE OSTI2026

Agentic Diagrammatica: Towards Autonomous Symbolic Computation in High Energy Physics

We present Diagrammatica, a symbolic computation extension to the HEPTAPOD agentic framework, which enables LLM agents to plan and execute multi-step theoretical calculations. Symbolic computation poses a distinctive reliability challenge for LLM agents, as correctness is governed by implicit mathematical conventions that are not encoded in a form that can be easily checked in the computational backend. We identify two complementary remedies, tool-constrained computation and targeted knowledge grounding, and pursue the first as the primary architecture. Concretely, we concentrate the agent's action distribution onto tool calls with convention-fixing semantics, in which the agent specifies a compact, human-auditable diagram specification and a trusted backend performs the symbolic or numerical manipulations exactly. The toolkit provides two complementary calculation paths consuming a shared diagram specification: Naive Dimensional Analysis (NDA) for order-of-magnitude rate estimates and Exact Diagrammatic Analysis (EDA) for tree-level symbolic calculations via automatic FeynCalc code generation, both supplemented by automatic Feynman diagram enumeration and a navigable theory knowledge base. The architecture is validated on two benchmarks: (1) an exhaustive catalog of all tree-level, single-vertex $1\to 2$ partial decay widths across scalar, fermion, and vector parents, with complete massless and threshold limits and Standard Model validation; and (2) an NDA sensitivity study of the muon decay multiplicity $μ^+ \to ν_μ\barν_e + n(e^+e^-) + e^-$, determining the maximum observable $n$ at current and planned muon experiments.

Menzo, Tony [Alabama U.; Fermilab] (ORCID:00000002↗

DOE OSTI2024

Search for axion-like particles through nuclear Primakoff production using the GlueX detector

We report on the results of the first search for the production of axion-like particles (ALPs) via Primakoff production on nuclear targets, γA → aA, in the “SRC-CT” experiment using the GlueX detector at Jefferson Lab. This search uses an integrated luminosity of 100 pb -1 nucleon on a 12 C target with a real photon beam of energies 6 < E γ < 10.8 GeV, and explores the mass region of 200 < m a < 450 MeV via the decay a → γγ. This mass range is between the π 0 and η meson masses, which enables the use of the measured η meson production rate to obtain absolute bounds on the ALP production with reduced sensitivity to experimental luminosity and detection efficiency. We find no evidence for an ALP, consistent with previous searches in the quoted mass range, and present limits on the effective photon coupling scale of $\mathscr{O}$(1 TeV -1 ). We further find that the ALP production limit we obtain is hindered by the peaking structure of the non-target-related dominant background the in GlueX spectrometer, which we treat by using data on 4 He to estimate and subtract it. We comment on how this search can be improved in a future higher-statistics dedicated measurement.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

DOE OSTI2024

Search For Low-Mass Quark-Antiquark Resonances Produced With an Initial State Photon at 13 TeV Using the CMS Detector

We present a search for low-mass narrow qq̅ resonances. This search uses data from LHC pp collisions at a center of mass of 13 TeV in Run 2, and corresponds to an integrated luminosity of 137 fb^{-1}, currently using 10\% of data. Utilizing full Run 2 data allows the use of a lower photon pT threshold trigger than a previous analysis performed with only 2016 data, allowing this analysis to be more sensitive to resonances in the low mass region. We require an initial state photon recoiling against the narrow resonance, leading to the resonance having a high transverse momentum. The high pT decay products of the resonance collimate and are reconstructed as a single large jet with an internal two-pronged substructure. A two-pronged dijet score based on the ParticleNet tagger is used to select jets with two-pronged substructure. The background is estimated via a data-driven method using a transfer factor between the distributions which fail and pass the two-pronged substructure requirement. The new physics signal is searched for as a narrow peak excess above the Standard Model backgrounds in the jet mass spectrum.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2025

A Measurement of the Largest-scale CMB E -mode Polarization with CLASS

We present measurements of large-scale cosmic microwave background E-mode polarization from the Cosmology Large Angular Scale Surveyor 90 GHz data. Using 115 det-yr of observations collected through 2024 with a variable-delay polarization modulator, we achieved a polarization sensitivity of 82 μK arcimin, comparable to Planck at similar frequencies (100 and 143 GHz ). The analysis demonstrates effective mitigation of systematic errors and addresses challenges to large-angular-scale power recovery posed by time-domain filtering in maximum-likelihood map-making. A novel implementation of the pixel-space transfer matrix is introduced, which enables efficient filtering simulations and bias correction in the power spectrum using the quadratic cross-spectrum estimator. Overall, we achieved an unbiased time-domain filtering correction to recover the largest angular scale polarization, with the only power deficit, arising from map-making nonlinearity, being characterized as <3%. Through cross-correlation with Planck, we detected the cosmic reionization at 99.4% significance and measured the reionization optical depth τ = $0.053^{+0.018}_{-0.019}$, marking the first ground-based attempt at such a measurement. At intermediate angular scales (ℓ > 30), our results, both independently and in cross-correlation with Planck, remain fully consistent with Planck’s measurements.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2024

Search For Low-Mass Quark-Antiquark Resonances Produced With an Initial State Photon at 13 TeV Using the CMS Detector

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2025

Estimating the energy threshold of phonon-mediated superconducting qubit detectors operated in an energy-relaxation sensing scheme

In recent years, the lack of a conclusive detection of weakly interacting massive particle dark matter at the 10 GeV/c 2 mass scale and above has encouraged development of low-threshold detector technology aimed at probing lighter dark matter candidates. Detectors based on Cooper-pair-breaking sensors have emerged as a promising avenue for this detection due to the low (meV-scale) energy required for breaking a Cooper pair in most superconductors. Among them, devices based on superconducting qubits are interesting candidates for sensing due to their observed sensitivity to broken Cooper pairs. Here, we have developed an end-to-end G 4 CMP -based simulation framework and have used it to evaluate performance metrics of qubit-based devices operating in a gate-based “energy relaxation” readout scheme, akin to those used in recent studies of qubit sensitivity to ionizing radiation. We find that for this readout scheme, the qubit acts as a phonon sensor with an energy threshold ranging down to ≃ 0.4 eV for near-term performance parameters.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

DOE OSTI2024

Extreme Risk Mitigation in Reinforcement Learning using Extreme Value Theory

Risk-sensitive reinforcement learning (RL) has garnered significant attention in recent years due to the growing interest in deploying RL agents in real-world scenarios. A critical aspect of risk awareness involves modelling highly rare risk events (rewards) that could potentially lead to catastrophic outcomes. These infrequent occurrences present a formidable challenge for data-driven methods aiming to capture such risky events accurately. While risk-aware RL techniques do exist, they suffer from high variance estimation due to the inherent data scarcity. Our work proposes to enhance the resilience of RL agents when faced with very rare and risky events by focusing on refining the predictions of the extreme values predicted by the state-action value distribution. To achieve this, we formulate the extreme values of the state-action value function distribution as parameterized distributions, drawing inspiration from the principles of extreme value theory (EVT). We propose an extreme value theory based actor-critic approach, namely, Extreme Valued Actor-Critic (EVAC) which effectively addresses the issue of infrequent occurrence by leveraging EVT-based parameterization. Importantly, we theoretically demonstrate the advantages of employing these parameterized distributions in contrast to other risk-averse algorithms. Our evaluations show that the proposed method outperforms other risk averse RL algorithms on a diverse range of benchmark tasks, each encompassing distinct risk scenarios.

Wang, Yu↗