Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Prior distribution”

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 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
NASA NTRS1990

Simple model for lambda-doublet propensities in bimolecular reactions

A simple geometric model is presented to account for lambda-doublet propensities in bimolecular reactions A + BC - AB + C. It applies to reactions in which AB is formed in a pi state, and in which the unpaired molecular orbital responsible for lambda-doubling arises from breaking the B-C bond. The lambda-doublet population ratio is predicted to be 2:1 provided that: (1) the motion of A in the transition state determines the plane of rotation of AB; (2) the unpaired pi orbital lying initially along the B-C bond may be resolved into a projection onto the AB plane of rotation and a projection perpendicular to this plane; (3) there is no preferred geometry for dissociation of ABC. The 2:1 lambda-doublet ratio is the 'unconstrained dynamics prior' lambda-doublet distribution for such reactions.

Bronikowski, Michael J.↗
NASA NTRS1998

Multiwavelength Observations of a Dramatic High-Energy Flare in the Blazar 3C 279

The blazar 3C 279, one of the brightest identified extragalactic objects in the gamma-ray sky, underwent a large (factor of approximately 10 in amplitude) flare in gamma-rays toward the end of a 3 week pointing by Compton Gamma Ray Observatory (CGRO), in 1996 January-February. The flare peak represents the highest gamma-ray intensity ever recorded for this object. During the high state, extremely rapid gamma-ray variability was seen, including an increase of a factor of 2.6 in approximately 8 hr, which strengthens the case for relativistic beaming. Coordinated multifrequency observations were carried out with Rossi X-Ray Timing Explorer (RXTE), Advanced Satellite for Cosmology and Astrophysics (ASCA; or, Astro-D), Roentgen Satellite (ROSAT), and International Ultraviolet Explorer (IUE) and from many ground-based observatories, covering most accessible wavelengths. The well-sampled, simultaneous RXTE light curve shows an outburst of lower amplitude (factor of approximately = 3) well correlated with the gamma-ray flare without any lag larger than the temporal resolution of approximately 1 day. The optical-UV light curves, which are not well sampled during the high-energy flare, exhibit more modest variations (factor of approximately 2) and a lower degree of correlation. The flux at millimetric wavelengths was near a historical maximum during the gamma-ray flare peak, and there is a suggestion of a correlated decay. We present simultaneous spectral energy distributions of 3C 279 prior to and near to the flare peak. The gamma-rays vary by more than the square of the observed IR-optical flux change, which poses some problems for specific blazar emission models. The synchrotron self-Compton (SSC) model would require that the largest synchrotron variability occurred in the mostly unobserved submillimeter/far-infrared region. Alternatively, a large variation in the external photon field could occur over a time-scale of a few days. This occurs naturally in the "mirror" model wherein the flaring region in the jet photoionizes nearby broad emission line clouds, which, in turn, provide soft external photons that are Comptonized to gamma-ray energies.

Wehrle, A. E.↗
NASA NTRS1998

Multiwavelength Observations of a Dramatic High-Energy Flare in the Blazar 3C 279

The blazar 3C 279, one of the brightest identified extragalactic objects in the gamma-ray sky, underwent a large (factor of approx. 10 in amplitude) flare in gamma-rays toward the end of a 3 week pointing by Compton Gamma Ray Observatory (CGRO), in 1996 January-February. The flare peak represents the highest gamma-ray intensity ever recorded for this object. During the high state, extremely rapid gamma-ray variability was seen, including an increase of a factor of 2.6 in approx. 8 hrs., which strengthens the case for relativistic beaming. Coordinated multifrequency observations were carried out with Rossi X-Ray Timing Explorer (RXTE), Advanced Satellite for Cosmology and Astrophysics (ASCA; or, Astro-D), Roentgen Satellite (ROSAT), and International Ultraviolet Explorer (IUE) and from many ground-based observatories, covering most accessible wavelengths. The well-sampled, simultaneous RXTE light curve shows an outburst of lower amplitude (factor of approx. or equal to 3) well correlated with the gamma-ray flare without any lag larger than the temporal resolution of approx. 1 day. The optical-UV light curves, which are not well sampled during the high-energy flare, exhibit more modest variations (factor of approx. 2) and a lower degree of correlation. The flux at millimetric wavelengths was near a historical maximum during the gamma-ray flare peak, and there is a suggestion of a correlated decay. We present simultaneous spectral energy distributions of 3C 279 prior to and near to the flare peak. The gamma-rays vary by more than the square of the observed IR-optical flux change, which poses some problems for specific blazar emission models. The synchrotron self-Compton (SSC) model would require that the largest synchrotron variability occurred in the mostly unobserved submillimeter/far-infrared region. Alternatively, a large variation in the external photon field could occur over a time scale of a few days. This occurs naturally in the "mirror" model wherein the flaring region in the jet photoionizes nearby broad emission line clouds, which, in turn, provide soft external photons that are Comptonized to gamma-ray energies.

Wehrle, A. E.↗
NASA NTRS2001

The AutoBayes Program Synthesis System: System Description

AUTOBAYES is a fully automatic program synthesis system for the statistical data analysis domain. Its input is a concise description of a data analysis problem in the form of a statistical model; its output is optimized and fully documented C/C++ code which can be linked dynamically into the Matlab and Octave environments. AUTOBAYES synthesizes code by a schema-guided deductive process. Schemas (i.e., code templates with associated semantic constraints) are applied to the original problem and recursively to emerging subproblems. AUTOBAYES complements this approach by symbolic computation to derive closed-form solutions whenever possible. In this paper, we concentrate on the interaction between the symbolic computations and the deductive synthesis process. A statistical model specifies for each problem variable (i.e., data or parameter) its properties and dependencies in the form of a probability distribution, A typical data analysis task is to estimate the best possible parameter values from the given observations or measurements. The following example models normal-distributed data but takes prior information (e.g., from previous experiments) on the data's mean value and variance into account.

Fischer, Bernd↗
NASA NTRS2008

Operational Characteristics and Plasma Measurements in a Low-Energy FARAD Thruster

Pulsed inductive plasma accelerators are spacecraft propulsion devices in which energy is stored in a capacitor and then discharged through an inductive coil. The device is electrodeless, inducing a plasma current sheet in propellant located near the face of the coil. The propellant is accelerated and expelled at a high exhaust velocity (order of 10 km/s) through the interaction of the plasma current with an induced magnetic field. The Faraday Accelerator with RF-Assisted Discharge (FARAD) thruster is a type of pulsed inductive plasma accelerator in which the plasma is preionized by a mechanism separate from that used to form the current sheet and accelerate the gas. Employing a separate preionization mechanism in this manner allows for the formation of an inductive current sheet at much lower discharge energies and voltages than those found in previous pulsed inductive accelerators like the Pulsed Inductive Thruster (PIT). In this paper, we present measurements aimed at quantifying the thruster's overall operational characteristics and providing additional insight into the nature of operation. Measurements of the terminal current and voltage characteristics during the pulse help quantify the output of the pulsed power train driving the acceleration coil. A fast ionization gauge is used to measure the evolution of the neutral gas distribution in the accelerator prior to a pulse. The preionization process is diagnosed by monitoring light emission from the gas using a photodiode, and a time-resolved global view of the evolving, accelerating current sheet is obtained using a fast-framing camera. Local plasma and field measurements are obtained using an array of intrusive probes. The local induced magnetic field and azimuthal current density are measured using B-dot probes and mini-Rogowski coils, respectively. Direct probing of the number density and electron temperature is performed using a triple probe.

Polzin, K. A.↗
DOE OSTI2022

Bayesian Monte Carlo Evaluation Framework for Imperfect Data and Models [Abstract]

Nuclear data evaluation methods conventionally make the following assumptions: prior and posterior probability distribution functions (PDFs) of all model parameters and data are normal (Gaussian); the linear approximation is sufficiently accurate for minimization of a cost function (even for non-linear models); and that both the model (of, e.g., neutron cross section) and experimental data (including their covariance data) are perfect. These assumptions are inherent to the well-known generalized linear least squares (GLLS) minimization method commonly used for evaluations of resolved resonance region (RRR) neutron cross sections. However, these assumptions are often not justified due to the presence of non-normal PDFs, non-linear models (e.g. R -matrix formalism), and inherent imperfections in data and models (e.g. discrepant data sets, discrepancies between the previous evaluation and newly measured data, or imperfect covariance data). We remove the said assumptions in a mathematical framework of Bayes’ theorem, and implement it using the Metropolis-Hastings Monte Carlo method. Parameters of a new kind are introduced to parameterize inherent imperfections, e.g. , any discrepancies between the theoretical model and measured data. These new parameters enable evaluators to quantify their expert judgement about any discrepancies or imperfections in a reproducible manner. We demonstrate the framework with an ongoing evaluation of 233 U in the eV region using the ENDF-B/VIII library and transmission data measured by Guber, et al. , and compare the posterior parameters to those obtained by conventional evaluation methods.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
NASA NTRS2017

Distributed Spatiotemporal Motion Planning for Spacecraft Swarms in Cluttered Environments

This paper focuses on trajectory planning for spacecraft swarms in cluttered environments, like debris fields or the asteroid belt. Our objective is to reconfigure the spacecraft swarm to a desired formation in a distributed manner while minimizing fuel and avoiding collisions among themselves and with obstacles. In our prior work we proposed a novel distributed guidance algorithm for spacecraft swarms in static environments. In this paper, we present the Multi-Agent Moving-Obstacles Spherical Expansion and Sequential Convex Programming (MAMO SE-SCP) algorithm that extends our prior work to include spatiotemporal constraints such as time-varying, moving obstacles and desired time-varying terminal positions. In the MAMO SE-SCP algorithm, each agent uses a spherical-expansion-based sampling algorithm to cooperatively explore the time-varying environment, a distributed assignment algorithm to agree on the terminal position for each agent, and a sequential-convex-programming-based optimization step to compute the locally-optimal trajectories from the current location to the assigned time-varying terminal position while avoiding collision with other agents and moving obstacles. Simulation results demonstrate that the proposed distributed algorithm can be used by a spacecraft swarm to achieve a time-varying, desired formation around an object of interest in a dynamic environment with many moving and tumbling obstacles.

Bandyopadhyay, Saptarshi↗
NASA NTRSDate not supplied

Design Concepts to Meet EASA SC-VTOL-01 Single Failure Criteria

The objective of the current work is to discuss European Union Aviation Safety Agency (EASA) SC-VTOL-01 single failure criteria, VTOL.2250(c). Prior studies have developed concept distributed propulsion and flight control (DPFC) system architectures and found they comply with EASA SC-VTOL-01 probabilistic failure criteria, VTOL.2510(a). Prior work developed two all-electric DPFC systems utilized in a quadrotor concept aircraft developed by the National Aeronautics and Space Administration (NASA); one uses interconnecting shafts and gearboxes to interconnect redundant motors with each rotor system and the other uses gearboxes to connect redundant motors locally, near each rotor. Common between the two electric DPFC systems were rotor shafts, epicyclic systems, and motors. The current work explores Category I failures in drive systems, relevant research to support fail-safe design practices for gear systems, research and adjacent industry trends in motor fail-safety and reliability, and proposed design concepts to comply with VTOL.2250(c). Continued research in fail-safe design concepts and design guidance will benefit eVTOL and conventional rotorcraft, alike. Continued research in these areas will benefit eVTOL certification against SC-VTOL-01, and could optimistically translate to more widespread adoption of similar fail-safe design concepts into new rotorcraft designs certified against CS-29.

Rotorcraft↗
NASA NTRS2017

Distributed Spatiotemporal Motion Planning for Spacecraft Swarms in Cluttered Environments

This paper focuses on trajectory planning for spacecraft swarms in cluttered environments, like debris fields or the asteroid belt. Our objective is to reconfigure the spacecraft swarm to a desired formation in a distributed manner while minimizing fuel and avoiding collisions among themselves and with the obstacles. In our prior work we proposed a novel distributed guidance algorithm for spacecraft swarms in static environments.1 In this paper, we present the Multi-Agent Moving-Obstacles Spherical Expansion and Sequential Convex Programming (MAMO SE–SCP) algorithm that extends our prior work to include spatiotemporal constraints such as time-varying, moving obstacles and desired time-varying terminal positions. In the MAMO SE–SCP algorithm, each agent uses a spherical-expansion-based sampling algorithm to cooperatively explore the time-varying environment, a distributed assignment algorithm to agree on the terminal position for each agent, and a sequential-convex-programming-based optimization step to compute the locally-optimal trajectories from the current location to the assigned time-varying terminal position while avoiding collision with other agent and the moving obstacles. Simulations results demonstrate that the proposed distributed algorithm can be used by a spacecraft swarm to achieve a time-varying, desired formation around an object of interest in a dynamic environment with many moving and tumbling obstacles.

Hadaegh, Fred Y.↗
DOE OSTI2026

Adaptive Cybersecurity for Distributed Energy Resources (AdCyDER): Online Reinforcement Learning with Stackelberg-Optimized Defenses — Pipeline Architecture, Evaluation Methodology, and Findings from a Synthetic-Data Evaluation

This report documents the design and evaluation of an integrated online-learning pipeline developed within the AdCyDER project for Distributed Energy Resource (DER) cybersecurity. The pipeline couples a Reinforcement Learning (RL) attack classifier — which produces an attack-type probability distribution — with a Stackelberg game-theoretic (GT) defense selector that consumes those distributions alongside SME-encoded priors over (defense, attack) effectiveness pairings and perdefense costs to choose grid-health-preserving defenses. The objective is not attack classification per se but production of distributions that drive effective defense selection through the Stackelberg layer, learned from delayed grid-health feedback rather than labeled attack data. AdCyDER as a whole is broader than the work presented here; this report covers the specific RL/GT loop integration and its evaluation. We present the integrated pipeline (SCADA telemetry with Fronius inverter physics, Suricata IDS, time-windowed aggregation, per-facility LSTM classifier, Stackelberg optimizer, OpenC2 actuators), an experimental campaign of 28 eight-hour iterations across three baseline modes, and a pipeline-ordered diagnostic protocol. The protocol identifies two distinct failure modes within the loop: paired supervised ceilings on the same features establish that the deployed online RL classifier (macro F1 ≈ 0.07) sits at least 4.7× below a same-architecture supervised LSTM (≈ 0.34) and 10–11× below a linear feature-signal ceiling (≈ 0.70–0.79 depending on per-facility isolation), localizing the dominant failure to the training procedure; and the reward signal driving online updates carries weak directional coupling with classifier correctness in the methodology-expected direction (multi-lens convergent: top-decile P(true) records produce more frequent state changes and slightly larger improvements, top-vs-bot Cohen’s 𝑑 ≈ −0.19), but at effect magnitudes too small to drive gradient-based learning at the campaign sample size. The original learning hypothesis is not supported by the data. The primary contributions are the diagnostic methodology — proposed as a transferable falsification protocol for online RL/GT defense pipelines learning from delayed environmental reward — and the open, reproducible experimental infrastructure. We outline reward reformulation as the highest-priority aspirational next step given the underpowered-but-aligned Q6 reading, with hardware-in-the-loop evaluation as the broadest scope-expansion option.

Blakely, Benjamin [Argonne National Laboratory (AN↗
DOE OSTI2023

Bayesian inference for the seismic moment tensor using regional waveforms and a data-derived distribution of velocity models

The largest source of uncertainty in any source inversion is the velocity model used to construct the transfer function employed in the forward model that relates observed ground motion to the seismic moment tensor. However, standard inverse procedures often does not quantify uncertainty in the seismic moment tensor due to error in the Green’s functions from uncertain event location and Earth structure. We attempt to incorporate this uncertainty into an estimation of the seismic moment tensor using a distribution of velocity models calculated in a prior effort based on different and complementary data sets. The posterior distribution of velocity models is then used to construct Green’s functions for use in Bayesian inference of an unknown seismic moment tensor using regional waveform data. The combined likelihood is estimated using data-specific error models and the posterior of the seismic moment tensor is estimated and can be interpreted in terms of most-probable source-type.

58 GEOSCIENCES↗
DOE OSTI2025

Hopping mediated transport between finite pools of redox proteins

Transport reactions in biology involve the flow of particles—electrons, ions, or molecules—between reservoirs. Here, we explore how electron transport between finite reservoirs depends on the nature of the reservoirs, including their size, occupancy, and interactions. We compare the transport kinetics produced by narrowband and wideband infinite reservoir models (described earlier) with a finite narrowband reservoir model. The transport between finite reservoirs is found to depend on both the initial charge distribution and the number of carriers present. Whether or not a steady-state transport regime is accessed prior to reaching the equilibrium charge distribution depends on these initial conditions.

10 SYNTHETIC FUELS↗
NASA NTRS2010

Sectors and Large-Scale Magnetic Field Strength Fluctuations in the Heliosheath Near 110 AU: Voyager 1,2009

This paper describes observations of daily averages of the magnetic field strength B and the magnetic polarity measured by Voyager 1 (V1) in the heliosheath during 2009 between 108.5 and 112.1 AU and at heliographic latitude 34. . 4. A negative magnetic polarity sector was observed during 2009 DOY 43.255. A positive polarity sector was observed during 2009 DOY 256.365. We offer the hypothesis that the existence of the two sectors is the result of the displacement of the wavy heliospheric current sheet to the position of V1 as a result of northward flow in the heliosheath. The large size of the sectors is caused by the slow radial motion of the flow observed by V1 in the heliosheath. The distribution of B during 2009 was lognormal, in contrast to the Gaussian distributions observed by V1 in the heliosheath prior to 2009. The large-scale fluctuations of B, described by the distribution of increments of daily averages of B, have a Tsallis distribution with q = 1.6. The large-scale fluctuations of B observed by V1 during 2009 have a multifractal spectrum with the same parameters that V1 observed during 2005 close to the termination shock at 94 AU. These results suggest that the large-scale magnetic fluctuations of B are in a metastable equilibrium state in the heliosheath between 94 AU and 112.1 AU.

Burlaga, L. F.↗
NASA NTRS2001

Damage Assessment of Creep Tested and Thermally Aged Metallic Alloys Using Acousto-Ultrasonics

In recent years emphasis has been placed on the early detection of material changes experienced in turbine powerplant components. During the scheduled overhaul of a turbine, the current techniques of examination of various hot section components aim to find flaws such as cracks, wear, and erosion, as well as excessive deformations. Thus far, these localized damage modes have been detected with satisfactory results. However, the techniques used to find these flaws provide no information on life until the flaws are actually detected. Major improvements in damage assessment, safety, as well as more accurate life prediction could be achieved if nondestructive evaluation (NDE) techniques could be utilized to sense material changes that occur prior to the localized defects mentioned. Because of elevated temperatures and excessive stresses, turbine components may experience creep behavior. As a result, it is desirable to monitor and access the current condition of such components. Research at the NASA Glenn Research Center involves developing and utilizing an NDE technique that discloses distributed material changes that occur prior to the localized damage detected by the current methods of inspection. In a recent study, creep processes in a nickel-base alloy were the life-limiting condition of interest, and the NDE technique was acousto-ultrasonics (AU). AU is an NDE technique that utilizes two ultrasonic transducers to interrogate the condition of a test specimen. The sending transducer introduces an ultrasonic pulse at a point on the surface of the specimen while a receiving transducer detects the signal after it has passed through the material. The goal of the method is to correlate certain parameters of the detected waveform to characteristics of the material between the two transducers. Here, the waveform parameter of interest is the attenuation due to internal damping for which information is being garnered from the frequency domain. The parameters utilized to indirectly quantify the attenuation are the ultrasonic decay rate as well as various moments of the frequency power spectrum. A new, user-friendly, graphical interface AU system was developed at NASA Glenn. This system is an all-inclusive, multifunction system that controls the sending and receiving ultrasonic transducers as well as all posttest signal analysis. The system's postprocessing software calculates the multiple parameters used to study the material of interest.

Gyekenyesi, Andrew L.↗
NASA NTRS1980

The use of prior probabilities in maximum likelihood classification of remotely sensed data

Possibilities for the improvement of classification accuracies by the use of prior information about the expected distribution of classes in the maximum likelihood classification of remote sensing data are examined. The modification of the maximum likelihood decision rule to take into account one or several sets of probabilities for the occurrence of classes which probabilities are based on independent knowledge of the area surveyed is demonstrated. It is then shown that the use of prior probabilities is sufficiently versatile so as to allow the prior weighting of output classes based on their anticipated sizes as well as the merging of continuously varying measurements with discrete collateral information data sets and the construction of time-sequential classification systems in which an earlier classification modifies the outcome of a latter one.

Strahler, A. H.↗
DOE OSTI2020

Radiation Mapping for an Unmanned Aerial Vehicle: Development and Simulated Testing of Algorithms for Source Mapping and Navigation Path Generation

Image reconstruction algorithms were developed for radiation source mapping and used for generating the search path of a moving radiation detector, such as one onboard an unmanned aerial vehicle. Simulations consisted of first assuming radioactive sources of varying complexity and estimating the radiation fields that would then be produced by that source distribution. Next, the "measurements" that would result from a pair of adjacent spatial locations were computed. A crude estimate of the source distribution likely to have produced such "measurements" was reconstructed based upon the limited measurements. Location of the next "measurement" was then determined as halfway between the location of the estimated source and the current "measurement." With each additional sample, improved source distribution reconstructions were made and used to inform the immediate direction of detector motion. Source reconstruction or mapping was formulated as an inverse problem solved with either maximum a posteriori or least squares (LS) regression deconvolution methods. Different amounts of noise were added to the simulated "measurements," allowing evaluation of the methods' performances as functions of signal-to-noise ratio of the measured map. As expected, methods that promote sparsity were better suited in reconstructing point sources. Reliable prior information of the source distribution also improved the reconstruction results, especially with distributed sources. With a non-negative least square algorithm and the suggested paths it generated, location of sources was successfully estimated to an accuracy of 0.014 m within nine iterations in a single-source scenario and 12 iterations in a two-source scenario, given a 10% error on the integrated counts and a Poisson distribution of the noise associated with the measured counts.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗
NASA NTRS1999

Main Geomagnetic Field Models from Oersted and Magsat Data Via a Rigorous General Inverse Theory with Error Bounds

The purpose of the grant was to study how prior information about the geomagnetic field can be used to interpret surface and satellite magnetic measurements, to generate quantitative descriptions of prior information that might be so used, and to use this prior information to obtain from satellite data a model of the core field with statistically justifiable error estimates. The need for prior information in geophysical inversion has long been recognized. Data sets are finite, and faithful descriptions of aspects of the earth almost always require infinite-dimensional model spaces. By themselves, the data can confine the correct earth model only to an infinite-dimensional subset of the model space. Earth properties other than direct functions of the observed data cannot be estimated from those data without prior information about the earth. Prior information is based on what the observer already knows before the data become available. Such information can be "hard" or "soft". Hard information is a belief that the real earth must lie in some known region of model space. For example, the total ohmic dissipation in the core is probably less that the total observed geothermal heat flow out of the earth's surface. (In principle, ohmic heat in the core can be recaptured to help drive the dynamo, but this effect is probably small.) "Soft" information is a probability distribution on the model space, a distribution that the observer accepts as a quantitative description of her/his beliefs about the earth. The probability distribution can be a subjective prior in the sense of Bayes or the objective result of a statistical study of previous data or relevant theories.

Backus, George E.↗
DOE OSTI2022

Gradient-Based Novelty Detection Boosted by Self-Supervised Binary Classification

Novelty detection aims to automatically identify out-of-distribution (OOD) data, without any prior knowledge of them. It is a critical step in data monitoring, behavior analysis and other applications, helping enable continual learning in the field. Conventional methods of OOD detection perform multi-variate analysis on an ensemble of data or features, and usually resort to the supervision with OOD data to improve the accuracy. In reality, such supervision is impractical as one cannot anticipate the anomalous data. In this paper, we propose a novel, self-supervised approach that does not rely on any pre-defined OOD data: (1) The new method evaluates the Mahalanobis distance of the gradients between the in-distribution and OOD data. (2) It is assisted by a self-supervised binary classifier to guide the label selection to generate the gradients, and maximize the Mahalanobis distance. In the evaluation with multiple datasets, such as CIFAR-10, CIFAR-100, SVHN and TinyImageNet, the proposed approach consistently outperforms state-of-the-art supervised and unsupervised methods in the area under the receiver operating characteristic (AUROC) and area under the precision-recall curve (AUPR) metrics. We further demonstrate that this detector is able to accurately learn one OOD class in continual learning.

Sun, Jingbo↗