Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “algorithm selection”

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 991 records · Page 55

Structural Health Management of Damaged Aircraft Structures Using the Digital Twin Concept

The development of multidisciplinary integrated Structural Health Management (SHM) tools will enable accurate detection, and prognosis of damaged aircraft under normal and adverse conditions during flight. As part of the digital twin concept, methodologies are developed by using integrated multiphysics models, sensor information and input data from an in-service vehicle to mirror and predict the life of its corresponding physical twin. SHM tools are necessary for both damage diagnostics and prognostics for continued safe operation of damaged aircraft structures. The adverse conditions include loss of control caused by environmental factors, actuator and sensor faults or failures, and structural damage conditions. A major concern in these structures is the growth of undetected damage/cracks due to fatigue and low velocity foreign object impact that can reach a critical size during flight, resulting in loss of control of the aircraft. To avoid unstable, catastrophic propagation of damage during a flight, load levels must be maintained that are below a reduced load-carrying capacity for continued safe operation of an aircraft. Hence, a capability is needed for accurate real-time predictions of damage size and safe load carrying capacity for structures with complex damage configurations. In the present work, a procedure is developed that uses guided wave responses to interrogate damage. As the guided wave interacts with damage, the signal attenuates in some directions and reflects in others. This results in a difference in signal magnitude as well as phase shifts between signal responses for damaged and undamaged structures. Accurate estimation of damage size, location, and orientation is made by evaluating the cumulative signal responses at various pre-selected sensor locations using a genetic algorithm (GA) based optimization procedure. The damage size, location, and orientation is obtained by minimizing the difference between the reference responses and the responses obtained by wave propagation finite element analysis of different representative cracks, geometries, and sizes.

Seshadri, Banavara R.↗

An Optimized Photoelectron Track Reconstruction Method for Photoelectric X-Ray Polarimeters

We present a data processing algorithm for angular reconstruction and event selection applied to 2-D photoelectron track images from X-ray polarimeters. The method reconstructs the initial emission angle of a photoelectron from the initial portion of the track, which is obtained by continuously cutting a track until the image moments or number of pixels fall below tunable thresholds. In addition, event selection which rejects round tracks quantified with eccentricity and circularity is performed so that polarimetry sensitivity considering a trade-off between the modulation factor and signal acceptance is maximized. The modulation factors with applying track selection are 26.6 ± 0.4, 46.1 ± 0.4, 62.3 ± 0.4, and 61.8 ± 0.3% at 2.7, 4.5, 6.4, and 8.0 keV, respectively, using the same data previously analyzed by Iwakiri et al. (2016), where the corresponding numbers are 26.9±0.4, 43.4±0.4, 54.4±0.3, and 59.1 ± 0.3%. The method improves polarimeter sensitivity by 5%–10% at the high energy end of the band previously presented (Iwakiri et al. 2016).

Kitaguchi, Takao↗

The Cumulus and Stratocumulus CloudSat-CALIPSO Dataset (CASCCAD)

Low clouds continue to contribute greatly to the uncertainty in cloud feedback estimates. Depending on whether a region is dominated by cumulus (Cu) or stratocumulus (Sc) clouds, the interannual low-cloud feedback is somewhat different in both spaceborne and large-eddy simulation studies. Therefore, simulating the correct amount and variation of the Cu and Sc cloud distributions could be crucial to predict future cloud feedbacks. Here we document spatial distributions and profiles of Sc and Cu clouds derived from Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) and CloudSat measurements. For this purpose, we create a new dataset called the Cumulus And Stratocumulus CloudSat-CALIPSO Dataset (CASCCAD), which identifies Sc, broken Sc, Cu under Sc, Cu with stratiform outflow and Cu. To separate the Cu from Sc, we design an original method based on the cloud height, horizontal extent, vertical variability and horizontal continuity, which is separately applied to both CALIPSO and combined CloudSat–CALIPSO observations. First, the choice of parameters used in the discrimination algorithm is investigated and validated in selected Cu, Sc and Sc–Cu transition case studies. Then, the global statistics are compared against those from existing passive- and active-sensor satellite observations. Our results indicate that the cloud optical thickness – as used in passive-sensor observations – is not a sufficient parameter to discriminate Cu from Sc clouds, in agreement with previous literature. Using clustering-derived datasets shows better results although one cannot completely separate cloud types with such an approach. On the contrary, classifying Cu and Sc clouds and the transition between them based on their geometrical shape and spatial heterogeneity leads to spatial distributions consistent with prior knowledge of these clouds, from ground-based, ship-based and field campaigns. Furthermore, we show that our method improves existing Sc–Cu classifications by using additional information on cloud height and vertical cloud fraction variation. Finally, the CASCCAD datasets provide a basis to evaluate shallow convection and stratocumulus clouds on a global scale in climate models and potentially improve our understanding of low-level cloud feedbacks. The CASCCAD dataset (Cesana, 2019, https://doi.org/10.5281/zenodo.2667637) is available on the Goddard Institute for Space Studies (GISS) website at https://data.giss.nasa.gov/clouds/casccad/ (last access: 5 November 2019) and on the zenodo website at https://zenodo.org/record/2667637 (last access: 5 November 2019).

Cesana, Gregory V.↗

Creating Near Real-time Hourly VIIRS Radiance Mosaic over Polar Regions from SatCORPS Global Cloud Composite (GCC)

While polar orbiting satellites are essential in providing the needed observations over polar regions for data assimilation, weather diagnoses/forecasting, and flight guidance and situation awareness in airborne field campaigns, there is limited temporal coverage. Often, polar mosaics are created by overlaying multiple satellite imagers on top of each other which can lead to transition lines and discontinuities in the data products. To help ease some of these shortcomings, we utilize the NASA Langley Satellite ClOud and Radiation Property retrieval System (SatCORPS) Mosaic algorithm for optimally fusing and merging selected imager radiances from the Visible Infrared Imaging Radiometer Suite (VIIRS) instrument onboard the Suomi National polar-orbiting Partnership (SNPP), Joint Polar Satellite System (JPSS-1), and JPSS-2 satellites, and the MODerate resolution Imaging Spectroradiometer (MODIS) instruments onboard AQUA and TERRA satellites. Since multiple satellites overpasses and instruments are used in creating hourly 3-km composites over the polar regions, selection of the best satellite data for each 3-km composite pixel is based on aggregated ranking of satellite resolution, time, viewing zenith angle, and distance from terminator. This ranking system provides a smoother transition and improves data discontinuity in the merged output resulting in a more seamless polar region dataset. We will discuss the use of this near real-time dataset to support polar airborne field campaigns and SatCORPS satellite-derived Global Cloud Composite (GCC) generation.

Satellite composites↗

R and DE Robotic Sensor Intern Team: Remote 3D Mapping Overview

The Department of Energy (DOE) nuclear complex includes many hazardous and challenging environments that must be inspected remotely and routinely for structural integrity evaluation and long-term planning purposes. The H-Canyon Air Exhaust (HCAEX) Tunnel biannual inspection is an example of such an environment. Presently inspections are performed by teleoperated robotic crawlers provide low resolution video feedback to the structural engineers. Advanced sensors such as Lidar and panospheric hi-res cameras have the potential to provide additional valuable data in future inspections. UT-Austin Lidar work in support of an H-Canyon deployment began with the DOE EM sponsored H-Canyon Integrated Research Project. The R and DE Robotic Sensor intern team includes two of the UT-Austin Masters students working on Lidar development to facilitate the pulling of university developed solutions to the field. The goal of this project is to develop a suite of sensors that could be used to gather advanced data in future HCAEX tunnel inspections. Of particular interest is the ability to map the tunnel walls at an accuracy that will allow for the comparison of two scans performed at different time periods to estimate the concrete erosion rate. Additionally, the development of the 3D point cloud in the GPS denied environment is the first step in being able to collect and overlay other 3D data such as radiation mapping. Project specific goals include: - Design and develop 3D mapping sensor suite. Merge Lidar point cloud with high res camera images Incorporate Inertial Measurement Unit (IMU) to assist with positioning. - Develop sensor software and user GUI Sensor data collection and integration software User GUI for reach back system. - Test solution on crawler in tunnel test bed Use crawler base similar to H-Canyon planned crawler Design and build test bed simulating tunnel conditions. To facilitate rapid sensor integration, software coding and user GUI development, the Robot Operating System (ROS) was employed as the development platform. ROS is a set of open source software libraries and tools that assist with the building of robot applications. ROS includes drivers for many sensors, state-of-the-art algorithms, and powerful developer tools. Sensors selected for the summer intern solution included assessment as to whether they possessed ROS drivers and associated sample code. Additionally, the ROS 3D Robot Visualizer (rviz) tool has been employed.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Automating Risk Assessments of Hazardous Material Shipments for Transportation Routes and Mode Selection

The METEOR project at Idaho National Laboratory (INL) successfully addresses the difficult problem in risk assessment analyses of combining the results from bounding deterministic simulation results with probabilistic (Monte Carlo) risk assessment techniques. This paper describes a software suite designed to perform sensitivity and cost/benefit analyses on selected transportation routes and vehicles to minimize risk associated with the shipment of hazardous materials. METEOR uses Monte Carlo techniques to estimate the probability of an accidental release of a hazardous substance along a proposed transportation route. A METEOR user selects the mode of transportation, origin and destination points, and charts the route using interactive graphics. Inputs to METEOR (many selections built in) include crash rates for the specific aircraft, soil/rock type and population densities over the proposed route, and bounding limits for potential accident types (velocity, temperature, etc.). New vehicle, materials, and location data are added when available. If the risk estimates are unacceptable, the risks associated with alternate transportation modes or routes can be quickly evaluated and compared. Systematic optimizing methods will provide the user with the route and vehicle selection identified with the lowest risk of hazardous material release. The effects of a selected range of potential accidents such as vehicle impact, fire, fuel explosions, excessive containment pressure, flooding, etc. are evaluated primarily using hydrocodes capable of accurately simulating the material response of critical containment components. Bounding conditions that represent credible accidents (i.e; for an impact event, velocity, orientations, and soil conditions) are used as input parameters to the hydrocode models yielding correlation functions relating accident parameters to component damage. The Monte Carlo algorithms use random number generators to make selections at the various decision points such as; crash, location, etc. For each pass through the routines, when a crash is randomly selected, crash parameters are then used to determine if failure has occurred using either external look up tables, correlations functions from deterministic calculations, or built in data libraries. The effectiveness of the software was recently demonstrated in safety analyses of the transportation of radioisotope systems for the US Dept. of Energy. These methods are readily adaptable to estimating risks associated with a variety of hazardous shipments such as spent nuclear fuel, explosives, and chemicals.

42 ENGINEERING↗

System for Processing Coded OFDM Under Doppler and Fading

An advanced communication system has been proposed for transmitting and receiving coded digital data conveyed as a form of quadrature amplitude modulation (QAM) on orthogonal frequency-division multiplexing (OFDM) signals in the presence of such adverse propagation-channel effects as large dynamic Doppler shifts and frequency-selective multipath fading. Such adverse channel effects are typical of data communications between mobile units or between mobile and stationary units (e.g., telemetric transmissions from aircraft to ground stations). The proposed system incorporates novel signal processing techniques intended to reduce the losses associated with adverse channel effects while maintaining compatibility with the high-speed physical layer specifications defined for wireless local area networks (LANs) as the standard 802.11a of the Institute of Electrical and Electronics Engineers (IEEE 802.11a). OFDM is a multi-carrier modulation technique that is widely used for wireless transmission of data in LANs and in metropolitan area networks (MANs). OFDM has been adopted in IEEE 802.11a and some other industry standards because it affords robust performance under frequency-selective fading. However, its intrinsic frequency-diversity feature is highly sensitive to synchronization errors; this sensitivity poses a challenge to preserve coherence between the component subcarriers of an OFDM system in order to avoid intercarrier interference in the presence of large dynamic Doppler shifts as well as frequency-selective fading. As a result, heretofore, the use of OFDM has been limited primarily to applications involving small or zero Doppler shifts. The proposed system includes a digital coherent OFDM communication system that would utilize enhanced 802.1la-compatible signal-processing algorithms to overcome effects of frequency-selective fading and large dynamic Doppler shifts. The overall transceiver design would implement a two-frequency-channel architecture (see figure) that would afford frequency diversity for reducing the adverse effects of multipath fading. By using parallel concatenated convolutional codes (also known as Turbo codes) across the dual-channel and advanced OFDM signal processing within each channel, the proposed system is intended to achieve at least an order of magnitude improvement in received signal-to-noise ratio under adverse channel effects while preserving spectral efficiency.

Tsou, Haiping↗

WE-A-17A-06: Evaluation of An Automatic Interstitial Catheter Digitization Algorithm That Reduces Treatment Planning Time and Provide Means for Adaptive Re-Planning in HDR Brachytherapy of Gynecologic Cancers

Purpose: To evaluate an automatic interstitial catheter digitization algorithm that reduces treatment planning time and provide means for adaptive re-planning in HDR Brachytherapy of Gynecologic Cancers. Methods: The semi-automatic catheter digitization tool utilizes a region growing algorithm in conjunction with a spline model of the catheters. The CT images were first pre-processed to enhance the contrast between the catheters and soft tissue. Several seed locations were selected in each catheter for the region growing algorithm. The spline model of the catheters assisted in the region growing by preventing inter-catheter cross-over caused by air or metal artifacts. Source dwell positions from day one CT scans were applied to subsequent CTs and forward calculated using the automatically digitized catheter positions. This method was applied to 10 patients who had received HDR interstitial brachytherapy on an IRB approved image-guided radiation therapy protocol. The prescribed dose was 18.75 or 20 Gy delivered in 5 fractions, twice daily, over 3 consecutive days. Dosimetric comparisons were made between automatic and manual digitization on day two CTs. Results: The region growing algorithm, assisted by the spline model of the catheters, was able to digitize all catheters. The difference between automatic and manually digitized positions was 0.8±0.3 mm. The digitization time ranged from 34 minutes to 43 minutes with a mean digitization time of 37 minutes. The bulk of the time was spent on manual selection of initial seed positions and spline parameter adjustments. There was no significance difference in dosimetric parameters between the automatic and manually digitized plans. D90% to the CTV was 91.5±4.4% for the manual digitization versus 91.4±4.4% for the automatic digitization (p=0.56). Conclusion: A region growing algorithm was developed to semi-automatically digitize interstitial catheters in HDR brachytherapy using the Syed-Neblett template. This automatic digitization tool was shown to be accurate compared to manual digitization.

60 APPLIED LIFE SCIENCES↗

Spontaneous Intramuscular Hematomas of the Abdomen and Pelvis: A New Multilevel Algorithm to Direct Transarterial Embolization and Patient Management

PurposeTo report our experience using a multilevel patient management algorithm to direct transarterial embolization (TAE) in managing spontaneous intramuscular hematoma (SIMH).Materials and MethodsFrom May 2006 to January 2014, twenty-seven patients with SIMH had been referred for TAE to our Radiology department. Clinical status and coagulation characteristics of the patients are analyzed. An algorithm integrating CT findings is suggested to manage SIMH. Patients were classified into three groups: Type I, SIMH with no active bleeding (AB); Type II, SIMH with AB and no muscular fascia rupture (MFR); and Type III, SIMH with MFR and AB. Type II is furthermore subcategorized as IIa, IIb and IIc. Types IIb, IIc and III were considered for TAE. The method of embolization as well as the material been used are described. Continuous variables are presented as mean ± SD. Categorical variables are reported as percentages. Technical success, clinical success, complications and 30-day mortality (d30 M) were analyzed.ResultsTwo patients (7.5%) had Type IIb, four (15%) Type IIc and 21 (77.5%) presented Type III. The detailed CT and CTA findings, embolization procedure and materials used are described. Technical success was 96% with a complication rate of 4%. Clinical success was 88%. The bleeding-related thirty-day mortality was 15% (all with Type III).ConclusionTAE is a safe and efficient technique to control bleeding that should be considered in selected SIMH as soon as possible. The proposed algorithm integrating CT features provides a comprehensive chart to select patients for TAE.Level of Evidence4.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

TU-F-CAMPUS-T-05: A Cloud-Based Monte Carlo Dose Calculation for Electron Cutout Factors

Purpose: For electron cutouts of smaller sizes, it is necessary to verify electron cutout factors due to perturbations in electron scattering. Often, this requires a physical measurement using a small ion chamber, diode, or film. The purpose of this study is to develop a fast Monte Carlo based dose calculation framework that requires only a smart phone photograph of the cutout and specification of the SSD and energy to determine the electron cutout factor, with the ultimate goal of making this cloud-based calculation widely available to the medical physics community. Methods: The algorithm uses a pattern recognition technique to identify the corners of the cutout in the photograph as shown in Figure 1. It then corrects for variations in perspective, scaling, and translation of the photograph introduced by the user’s positioning of the camera. Blob detection is used to identify the portions of the cutout which comprise the aperture and the portions which are cutout material. This information is then used define physical densities of the voxels used in the Monte Carlo dose calculation algorithm as shown in Figure 2, and select a particle source from a pre-computed library of phase-spaces scored above the cutout. The electron cutout factor is obtained by taking a ratio of the maximum dose delivered with the cutout in place to the dose delivered under calibration/reference conditions. Results: The algorithm has been shown to successfully identify all necessary features of the electron cutout to perform the calculation. Subsequent testing will be performed to compare the Monte Carlo results with a physical measurement. Conclusion: A simple, cloud-based method of calculating electron cutout factors could eliminate the need for physical measurements and substantially reduce the time required to properly assure accurate dose delivery.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Flexible language constructs for large parallel programs

The goal of the research described is to develop flexible language constructs for writing large data parallel numerical programs for distributed memory (MIMD) multiprocessors. Previously, several models have been developed to support synchronization and communication. Models for global synchronization include SIMD (Single Instruction Multiple Data), SPMD (Single Program Multiple Data), and sequential programs annotated with data distribution statements. The two primary models for communication include implicit communication based on shared memory and explicit communication based on messages. None of these models by themselves seem sufficient to permit the natural and efficient expression of the variety of algorithms that occur in large scientific computations. An overview of a new language that combines many of these programming models in a clean manner is given. This is done in a modular fashion such that different models can be combined to support large programs. Within a module, the selection of a model depends on the algorithm and its efficiency requirements. An overview of the language and discussion of some of the critical implementation details is given.

Rosing, Matthew↗

Encoding nonlinear and unsteady aerodynamics of limit cycle oscillations using nonlinear sparse Bayesian learning

This article investigates the applicability of a recently proposed, nonlinear sparse Bayesian learning (NSBL) algorithm to identify and estimate the complex aerodynamics of limit cycle oscillations. NSBL provides a semi-analytical framework for determining the data-optimal sparse model nested within a (potentially) over-parameterized model. This is particularly relevant to nonlinear dynamical systems where modelling approaches involve the use of physics-based and data-driven components. In such cases, the data-driven components, where analytical descriptions of the physical processes are not readily available, are often prone to overfitting, meaning that the empirical aspects of these models will often involve the calibration of an unnecessarily large number of parameters. While an overparameterized model may fit the observed data well, such models may be inadequate for making predictions in regimes that are different from those wherein the data were recorded. In view of this, it is desirable to not only calibrate the model parameters, but also identify the optimal compromise between data fit and model complexity. In this article, we exhibit the optimal model discovery for an aeroelastic system wherein the structural dynamics are well-known and described by a differential equation model, coupled with a semi-empirical aerodynamic model for laminar separation flutter, resulting in low-amplitude limit cycle oscillations (LCO). To illustrate the performance of the algorithm, in this article, we use synthetic data and demonstrate the ability of the algorithm to correctly rediscover the optimal model and model parameters, given a known data-generating model. The synthetic data are generated from a forward simulation of a known differential equation model with parameters selected so as to mimic the dynamics observed in wind-tunnel experiments. Subsequently, we demonstrate the performance of the algorithm for model selection using noisy LCO data from wind tunnel experiments. As there is no ground truth available for the experimental data case, we provide a comparison between NSBL and Bayesian model selection to validate the results, and demonstrate the use of NSBL as an efficient alternative to traditional methods.

97 MATHEMATICS AND COMPUTING↗

Avoiding Selection Bias in Generating Examples of Plans in the Presence of Heuristic Error

It is generally understood that heuristic error hurts the performance of search algorithms, measured in terms of search effort. Hence there is an interest in understanding how to reduce heuristic error. One way to do this is to learn a heuristic from a set of examples of plans generated offline, e.g. bootstrapping methods. In this paper, we consider how some methods for generating examples of plans may skew the training set in the presence of heuristic errors. Initial theoretical results show that duplicate detection is one source of selection bias in the canonical A* algorithm. We introduce a duplicate selection scheme for A* that avoids selection bias in generating cost-optimal examples, without compromising memory efficiency, and develop ideas in the satisficing setting. We evaluate our approach on n x m grids with multiple cost-optimal solutions and synthetic heuristic error. Finally, we attempt to extend these ideas to the problem of generating extreme examples of plans.

Alison S Paredes↗

Active learning in the presence of unlabelable examples

We propose a new active learning framework where the expert labeler is allowed to decline to label any example. This may be necessary because the true label is unknown or because the example belongs to a class that is not part of the real training problem. We show that within this framework, popular active learning algorithms (such as Simple) may perform worse than random selection because they make so many queries to the unlabelable class. We present a method by which any active learning algorithm can be modified to avoid unlabelable examples by training a second classifier to distinguish between the labelable and unlabelable classes. We also demonstrate the effectiveness of the method on two benchmark data sets and a real-world problem.

active learning↗

Analysis of open-loop conical scan pointing error and variance estimators

General pointing error and variance estimators for an open-loop conical scan (conscan) system are derived and analyzed. The conscan algorithm is modeled as a weighted least-squares estimator whose inputs are samples of receiver carrier power and its associated measurement uncertainty. When the assumptions of constant measurement noise and zero pointing error estimation are applied, the variance equation is then strictly a function of the carrier power to uncertainty ratio and the operator selectable radius and period input to the algorithm. The performance equation is applied to a 34-m mirror-based beam-waveguide conscan system interfaced with the Block V Receiver Subsystem tracking a Ka-band (32-GHz) downlink. It is shown that for a carrier-to-noise power ratio greater than or equal to 30 dB-Hz, the conscan period for Ka-band operation may be chosen well below the current DSN minimum of 32 sec. The analysis presented forms the basis of future conscan work in both research and development as well as for the upcoming DSN antenna controller upgrade for the new DSS-24 34-m beam-waveguide antenna.

Alvarez, L. S.↗

A Cosmic Void Catalog of SDSS DR12 BOSS Galaxies

We present a cosmic void catalog using the large-scale structure galaxy catalog from the Baryon Oscillation Spectroscopic Survey (BOSS). This galaxy catalog is part of the Sloan Digital Sky Survey (SDSS) Data Release 12 and is the final catalog of SDSS-III. We take into account the survey boundaries, masks, and angular and radial selection functions, and apply the ZOBOV void finding algorithm to the Galaxy catalog. We identify a total of 10,643 voids. After making quality cuts to ensure that the voids represent real underdense regions, we obtain 1,228 voids with effective radii spanning the range 20–100 h{sup −1} Mpc and with central densities that are, on average, 30% of the mean sample density. We release versions of the catalogs both with and without quality cuts. We discuss the basic statistics of voids, such as their size and redshift distributions, and measure the radial density profile of the voids via a stacking technique. In addition, we construct mock void catalogs from 1000 mock galaxy catalogs, and find that the properties of BOSS voids are in good agreement with those in the mock catalogs. We compare the stellar mass distribution of galaxies living inside and outside of the voids, and find no large difference. These BOSS and mock void catalogs are useful for a number of cosmological and galaxy environment studies.

79 ASTRONOMY AND ASTROPHYSICS↗

Actuator selection for large space structures

The paper discusses the process of selecting the actuator locations and the determination of the required number of actuators for large space structures. The selection is based on the definitions of the degree of controllability, the independence of actuators, and the effectiveness of the individual actuators. An algorithm is developed that can be used for the selection of the essential number of actuators and for finding some defects of the system, such as the insuffiency of the available actuator locations for effective control of the whole system or a too crowded frequency distribution. The efficiency of the algorithm was demonstrated by an application to the Space Station.

Reddy, A. S. S. R.↗

Impala: A Flexible Adaptive Sampling Algorithm for Active Learning

The Implicit Potential Active Learning Algorithm, or impala, is a novel approach to adaptive sampling that allows for the blending of multiple independent query strategies into a common batch active learning framework. This methodology is agnostic to both the underlying learning model and the employed query strategies. In essence, impala treats query strategies as potential functions, which are combined into a probability of candidate selection. The end result is a flexible algorithm that can dynamically balance exploration and exploitation whose inherent variability guards against over-zealous greedy searching.

97 MATHEMATICS AND COMPUTING↗