Search NASA⌕ Search

SEARCH · Search NASA

Results for “Randomized sketching”

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.

29 records · Page 2
All sourcesNASA NTRSDOE OSTI
NASA NTRSDate not supplied

Planning Bias: Planning as a Source of Sampling Bias

Many data-driven planning methods are trained on data generated by planners. It is well known that many statistical learning methods are sensitive to sampling bias, and yet there has been little or no attention to planning as a sampling method and its role in introducing sampling bias into planner-generated training data. Recently, it has been demonstrated that A**,* in the presence of problems with variable heuristic error, prefers some solutions over other equally cost-optimal solutions. But, as we discuss in this paper, mitigation may not be as simple as resolving arbitrary tie-breaking by sampling from ties uniformly at random. In this paper, we formalize an intuition of planning bias. We focus on problems which output a single solution. Diverse planning only complicates the problem by generalizing it to bias in the set of sets; we show how it is subject to bias in the single solution. We make some useful observations about deterministic algorithms in contrast to non-deterministic algorithms. We explain how information entropy may be a good way to measure planning bias, and discuss some issues in evaluating practical approaches to measurement. We address the intuition that uniform random tiebreaking should mitigate bias; and sketch a novel approach to constructing an appropriate random distribution for duplicate detection during forward search for unbiased A*. Finally, we suggest directions for future work.

Planning Scheduling Algorithms↗
DOE OSTI2024

Bringing randomized algorithms to mainstream numerical linear algebra

Numerical linear algebra (NLA) underpins huge swaths of computational science and engineering. For scientists and engineers to make the most of the DOE’s computing resources, it is essential that they have access to high-performance implementations of algorithms with best-in-class scalability and reliability. Despite this, prevailing NLA libraries have little to no support for breakthrough algorithms from the field of randomized numerical linear algebra (RandNLA) that have been developed over the past twenty years. The goal of this LDRD was to break a log-jam that had prevented broad adoption of RandNLA. Our work had two thrusts. The first was to develop RandBLAS: a trustworthy and high-performance C++ library for randomized dimension reduction (an operation widely known as sketching). The second was the development of a novel randomized algorithm for computing a challenging type of matrix decomposition known as Householder QR with column pivoting (Householder QRCP). In this one-year late-start LDRD we successfully delivered RandBLAS 1.0 and new CPU and GPU codes for Householder QRCP. RandBLAS has extensive documentation at https://randblas.readthedocs.io/en/stable/. Papers on RandBLAS and and our high-performance QRCP codes are forthcoming.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Randomized Algorithms for Low-Rank Matrix and Tensor Decompositions

This paper surveys randomized algorithms in numerical linear algebra for low-rank decompositions of matrices and tensors. The survey begins with a review of classical matrix algorithms that can be accelerated by randomized dimensionality reduction, such as the singular value decomposition (SVD) or interpolative (ID) and CUR decompositions. Recent advances in randomized dimensionality reduction are discussed, including new methods of fast matrix sketching and sampling techniques, which are incorporated into classical matrix algorithms for fast low-rank matrix approximations. The extension of randomized matrix algorithms to tensors is then explored for several low-rank tensor decompositions in the CP and Tucker formats, including the higher-order SVD, ID, and CUR decomposition.

Pearce, Katherine J. [The University of Texas at A↗
DOE OSTI2020

An investigation of Newton-Sketch and subsampled Newton methods

Sketching, a dimensionality reduction technique, has received much attention in the statistics community. In this paper, we study sketching in the context of Newton's method for solving finite-sum optimization problems in which the number of variables and data points are both large. In this work, we study two forms of sketching that perform dimensionality reduction in data space: Hessian subsampling and randomized Hadamard transformations. Each has its own advantages, and their relative tradeoffs have not been investigated in the optimization literature. Additionally, our study focuses on practical versions of the two methods in which the resulting linear systems of equations are solved approximately, at every iteration, using an iterative solver. The advantages of using the conjugate gradient method vs. a stochastic gradient iteration are revealed through a set of numerical experiments, and a complexity analysis of the Hessian subsampling method is presented.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Sketching Algorithms in Distributed Systems

In this position paper, we discuss exciting recent advancements in sketching algorithms applied to distributed systems. That is, we look at randomized algorithms that simultaneously reduce the data dimensionality, offer potential privacy benefits, while maintaining verifiably high levels of algorithm accuracy and performance in multi-node computational setups. We look at next steps and discuss the applicability to real systems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Randomized Algorithms for Scientific Computing (RASC)

Randomized algorithms have propelled advances in artificial intelligence (AI) and represent a foundational research area in advancing AI for Science. Future advancements in DOE Office of Science priority areas such as climate science, astrophysics, fusion, advanced materials, combustion, and quantum computing all require randomized algorithms for surmounting challenges of complexity, robustness, and scalability. Advances in data collection and numerical simulation have changed the dynamics of scientific research and motivate the need for randomized algorithms. For instance, advances in imaging technologies such as X-ray ptychography, electron microscopy, electron energy loss spectroscopy, or adaptive optics lattice light-sheet microscopy collect hyperspectral imaging and scattering data in terabytes, at breakneck speed enabled by state-of-the-art detectors. The data collection is exceptionally fast compared with its analysis. Likewise, advances in high-performance architectures have made exascale computing a reality and changed the economies of scientific computing in the process. Floating-point operations that create data are essentially free in comparison with data movement. Thus far, most approaches have focused on creating faster hardware. Ironically, this faster hardware has exacerbated the problem by making data still easier to create. Under such an onslaught, scientists often resort to heuristic deterministic sampling schemes (e.g., low-precision arithmetic, sampling every nth element) and sacrifice potentially valuable accuracy. Dramatically better results can be achieved via randomized algorithms, reducing the data size as much as or more than naive deterministic subsampling can achieve, while retaining the high accuracy of computing on the full data set. By randomized algorithms we mean those algorithms that employ some form of randomness in internal algorithmic decisions to accelerate time to solution, increase scalability, or improve reliability. Examples include matrix sketching for solving large-scale least-squares problems (see Figure 1) and stochastic gradient descent for training machine learning models. We are not recommending heuristic methods but rather randomized algorithms that have certificates of correctness and probabilistic guarantees of optimality and near-optimality. Such approaches can be useful beyond acceleration, for example, in understanding how to avoid measure zero worst-case scenarios that plague methods such as QR matrix factorization.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Two-Level Sketching Alternating Anderson Acceleration for Complex Physics Applications

We present a novel two-level sketching extension of the Alternating Anderson–Picard (AAP) method for accelerating fixed-point iterations in challenging single- and multiphysics simulations governed by discretized PDEs. Our approach combines a static, physics-based projection that reduces the least-squares (LS) problem to the most informative field (e.g., via Schur-complement insight) with a dynamic, algebraic sketching stage driven by a backward stability analysis under Lipschitz continuity. We introduce inexpensive estimators for stability thresholds and cache-aware randomized selection strategies to balance computational cost against memory access overhead. The resulting algorithm solves reduced LS systems in place, minimizes memory footprints, and seamlessly alternates between low-cost Picard updates and Anderson mixing. Implemented in Julia, our two-level sketching AAP achieves up to 50% time-to-solution reductions compared to standard Anderson acceleration—without degrading convergence rates—on benchmark problems including Stokes, 𝑝-Laplacian, bidomain, and Navier–Stokes formulations at varying problem sizes. These results demonstrate the method’s robustness, scalability, and potential for integration into high-performance scientific computing frameworks. Our implementation is available open source in the AAP.jl library.

Barnafi, Nicolas [University of Chile, Santiago]↗
DOE OSTI2025

Randomized Preconditioned Solvers for Strong Constraint 4D-Var Data Assimilation

The Strong Constraint 4D Variational (SC-4DVAR) data assimilation method is widely used in climate and weather applications. SC-4DVAR involves solving a minimization problem to compute the maximum a posteriori estimate, which we tackle using the Gauss-Newton method. The computation of the descent direction is expensive since it involves the solution of a large-scale and potentially ill-conditioned linear system, solved using the preconditioned conjugate gradient (PCG) method. Here, to address this cost, we efficiently construct scalable preconditioners using three different randomization techniques, which all rely on a certain low-rank structure involving the Gauss-Newton Hessian. The proposed techniques come with theoretical guarantees on the condition number, and at the same time, are amenable to parallelization. We also develop an adaptive approach to estimate the sketch size and choose between the reuse or recomputation of the preconditioner. We demonstrate the performance and effectiveness of our methodology on two representative model problems—the Burgers and barotropic vorticity equation—showing a drastic reduction in both the number of PCG iterations and the number of Gauss-Newton Hessian products after including the preconditioner construction cost.

Gauss-Newton↗
DOE OSTI2022

Dynamic mode decomposition with core sketch

With the increase in collected data volumes, either from experimental measurements or high fidelity simulations, there is an ever-growing need to develop computationally efficient tools to process, analyze, and interpret these datasets. Modal analysis techniques have gained great interest due to their ability to identify patterns in the data and extract valuable information about the system being considered. Dynamic mode decomposition (DMD) relies on elements of the Koopman approximation theory to compute a set of modes, each associated with a fixed oscillation frequency and a decay/growth rate. Extracting these details from large datasets can be computationally expensive due to the need to implement singular value decomposition of the input data matrix. Sketching algorithms have become popular in numerical linear algebra where statistical theoretic approaches are utilized to reduce the cost of major operations. A sketch of a matrix is another matrix, which is significantly smaller, but still sufficiently approximates the original system. We put forth an efficient DMD framework, SketchyDMD, based on a core sketching algorithm that captures information about the range and corange (their mutual relationship) of input data. The proposed sketching-based framework can accelerate various portions of the DMD routines, compared to classical methods that operate directly on the raw input data. We conduct numerical experiments using the spherical shallow water equations as a prototypical model in the context of geophysical flows. In conclusion, we show that the proposed SketchyDMD is superior to existing randomized DMD methods that are based on capturing only the range of the input data.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
NASA NTRS2007

Potential of VIIRS Time Series Data for Aiding the USDA Forest Service Early Warning System for Forest Health Threats: A Gypsy Moth Defoliation Case Study

This report details one of three experiments performed during FY 2007 for the NASA RPC (Rapid Prototyping Capability) at Stennis Space Center. This RPC experiment assesses the potential of VIIRS (Visible/Infrared Imager/Radiometer Suite) and MODIS (Moderate Resolution Imaging Spectroradiometer) data for detecting and monitoring forest defoliation from the non-native Eurasian gypsy moth (Lymantria dispar). The intent of the RPC experiment was to assess the degree to which VIIRS data can provide forest disturbance monitoring information as an input to a forest threat EWS (Early Warning System) as compared to the level of information that can be obtained from MODIS data. The USDA Forest Service (USFS) plans to use MODIS products for generating broad-scaled, regional monitoring products as input to an EWS for forest health threat assessment. NASA SSC is helping the USFS to evaluate and integrate currently available satellite remote sensing technologies and data products for the EWS, including the use of MODIS products for regional monitoring of forest disturbance. Gypsy moth defoliation of the mid-Appalachian highland region was selected as a case study. Gypsy moth is one of eight major forest insect threats listed in the Healthy Forest Restoration Act (HFRA) of 2003; the gypsy moth threatens eastern U.S. hardwood forests, which are also a concern highlighted in the HFRA of 2003. This region was selected for the project because extensive gypsy moth defoliation occurred there over multiple years during the MODIS operational period. This RPC experiment is relevant to several nationally important mapping applications, including agricultural efficiency, coastal management, ecological forecasting, disaster management, and carbon management. In this experiment, MODIS data and VIIRS data simulated from MODIS were assessed for their ability to contribute broad, regional geospatial information on gypsy moth defoliation. Landsat and ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) data were used to assess the quality of gypsy moth defoliation mapping products derived from MODIS data and from simulated VIIRS data. The project focused on use of data from MODIS Terra as opposed to MODIS Aqua mainly because only MODIS Terra data was collected during 2000 and 2001-years with comparatively high amounts of gypsy moth defoliation within the study area. The project assessed the quality of VIIRS data simulation products. Hyperion data was employed to assess the quality of MODIS-based VIIRS simulation datasets using image correlation analysis techniques. The ART (Application Research Toolbox) software was used for data simulation. Correlation analysis between MODIS-simulated VIIRS data and Hyperion-simulated VIIRS data for red, NIR (near-infrared), and NDVI (Normalized Difference Vegetation Index) image data products collectively indicate that useful, effective VIIRS simulations can be produced using Hyperion and MODIS data sources. The r(exp 2) for red, NIR, and NDVI products were 0.56, 0.63, and 0.62, respectively, indicating a moderately high correlation between the 2 data sources. Temporal decorrelation from different data acquisition times and image misregistration may have lowered correlation results. The RPC experiment also generated MODIS-based time series data products using the TSPT (Time Series Product Tool) software. Time series of simulated VIIRS NDVI products were produced at approximately 400-meter resolution GSD (Ground Sampling Distance) at nadir for comparison to MODIS NDVI products at either 250- or 500-meter GSD. The project also computed MODIS (MOD02) NDMI (Normalized Difference Moisture Index) products at 500-meter GSD for comparison to NDVI-based products. For each year during 2000-2006, MODIS and VIIRS (simulated from MOD02) time series were computed during the peak gypsy moth defoliation time frame in the study area (approximately June 10 through July 27). Gypsy moth defoliation mapping products from simated VIIRS and MOD02 time series were produced using multiple methods, including image classification and change detection via image differencing. The latter enabled an automated defoliation detection product computed using percent change in maximum NDVI for a peak defoliation period during 2001 compared to maximum NDVI across the entire 2000-2006 time frame. Final gypsy moth defoliation mapping products were assessed for accuracy using randomly sampled locations found on available geospatial reference data (Landsat and ASTER data in conjunction with defoliation map data from the USFS). Extensive gypsy moth defoliation patches were evident on screen displays of multitemporal color composites derived from MODIS data and from simulated VIIRS vegetation index data. Such defoliation was particularly evident for 2001, although widespread denuded forests were also seen for 2000 and 2003. These visualizations were validated using aforementioned reference data. Defoliation patches were visible on displays of MODIS-based NDVI and NDMI data. The viewing of apparent defoliation patches on all of these products necessitated adoption of a specialized temporal data processing method (e.g., maximum NDVI during the peak defoliation time frame). The frequency of cloud cover necessitated this approach. Multitemporal simulated VIIRS and MODIS Terra data both produced effective general classifications of defoliated forest versus other land cover. For 2001, the MOD02-simulated VIIRS 400-meter NDVI classification produced a similar yet slightly lower overall accuracy (87.28 percent with 0.72 Kappa) than the MOD02 250-meter NDVI classification (88.44 percent with 0.75 Kappa). The MOD13 250-meter NDVI classification had a lower overall accuracy (79.13 percent) and a much lower Kappa (0.46). The report discusses accuracy assessment results in much more detail, comparing overall classification and individual class accuracy statistics for simulated VIIRS 400-meter NDVI, MOD02 250-meter NDVI, MOD02-500 meter NDVI, MOD13 250-meter NDVI, and MOD02 500-meter NDMI classifications. Automated defoliation detection products from simulated VIIRS and MOD02 data for 2001 also yielded similar, relatively high overall classification accuracy (85.55 percent for the VIIRS 400-meter NDVI versus 87.28 percent for the MOD02 250-meter NDVI). In contrast, the USFS aerial sketch map of gypsy moth defoliation showed a lower overall classification accuracy at 73.64 percent. The overall classification Kappa values were also similar for the VIIRS (approximately 0.67 Kappa) versus the MOD02 (approximately 0.72 Kappa) automated defoliation detection product, which were much higher than the values exhibited by the USFS sketch map product (overall Kappa of approximately 0.47). The report provides additional details on the accuracy of automated gypsy moth defoliation detection products compared with USFS sketch maps. The results suggest that VIIRS data can be effectively simulated from MODIS data and that VIIRS data will produce gypsy moth defoliation mapping products that are similar to MODIS-based products. The results of the RPC experiment indicate that VIIRS and MODIS data products have good potential for integration into the forest threat EWS. The accuracy assessment was performed only for 2001 because of time constraints and a relative scarcity of cloud-free Landsat and ASTER data for the peak defoliation period of the other years in the 2000-2006 time series. Additional work should be performed to assess the accuracy of gypsy moth defoliation detection products for additional years.The study area (mid-Appalachian highlands) and application (gypsy moth forest defoliation) are not necessarily representative of all forested regions and of all forest threat disturbance agents. Additional work should be performed on other inland and coastal regions as well as for other major forest threats.

Spruce, Joseph P.↗