Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “linear embedding”

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 19 records

REDUCING THE DIMENSIONALITY OF DATA: LOCALLY LINEAR EMBEDDING OF SLOAN GALAXY SPECTRA

We introduce locally linear embedding (LLE) to the astronomical community as a new classification technique, using Sloan Digital Sky Survey spectra as an example data set. LLE is a nonlinear dimensionality reduction technique that has been studied in the context of computer perception. We compare the performance of LLE to well-known spectral classification techniques, e.g., principal component analysis and line-ratio diagnostics. We find that LLE combines the strengths of both methods in a single, coherent technique, and leads to improved classification of emission-line spectra at a relatively small computational cost. We also present a data subsampling technique that preserves local information content, and proves effective for creating small, efficient training samples from large, high-dimensional data sets. Software used in this LLE-based classification is made available.

79 ASTRONOMY AND ASTROPHYSICS↗

CLASSIFICATION OF STELLAR SPECTRA WITH LOCAL LINEAR EMBEDDING

We investigate the use of dimensionality reduction techniques for the classification of stellar spectra selected from the Sloan Digital Sky Survey. Using local linear embedding (LLE), a technique that preserves the local (and possibly nonlinear) structure within high-dimensional data sets, we show that the majority of stellar spectra can be represented as a one-dimensional sequence within a three-dimensional space. The position along this sequence is highly correlated with spectral temperature. Deviations from this 'stellar locus' are indicative of spectra with strong emission lines (including misclassified galaxies) or broad absorption lines (e.g., carbon stars). Based on this analysis, we propose a hierarchical classification scheme using LLE that progressively identifies and classifies stellar spectra in a manner that requires no feature extraction and that can reproduce the classic MK classifications to an accuracy of one type.

79 ASTRONOMY AND ASTROPHYSICS↗

KEPLER ECLIPSING BINARY STARS. III. CLASSIFICATION OF KEPLER ECLIPSING BINARY LIGHT CURVES WITH LOCALLY LINEAR EMBEDDING

We present an automated classification of 2165 Kepler eclipsing binary (EB) light curves that accompanied the second Kepler data release. The light curves are classified using locally linear embedding, a general nonlinear dimensionality reduction tool, into morphology types (detached, semi-detached, overcontact, ellipsoidal). The method, related to a more widely used principal component analysis, produces a lower-dimensional representation of the input data while preserving local geometry and, consequently, the similarity between neighboring data points. We use this property to reduce the dimensionality in a series of steps to a one-dimensional manifold and classify light curves with a single parameter that is a measure of 'detachedness' of the system. This fully automated classification correlates well with the manual determination of morphology from the data release, and also efficiently highlights any misclassified objects. Once a lower-dimensional projection space is defined, the classification of additional light curves runs in a negligible time and the method can therefore be used as a fully automated classifier in pipeline structures. The classifier forms a tier of the Kepler EB pipeline that pre-processes light curves for the artificial intelligence based parameter estimator.

79 ASTRONOMY AND ASTROPHYSICS↗

Linear embedding of nonlinear dynamical systems and prospects for efficient quantum algorithms

The simulation of large nonlinear dynamical systems, including systems generated by discretization of hyperbolic partial differential equations, can be computationally demanding. Such systems are important in both fluid and kinetic computational plasma physics. This motivates exploring whether a future error-corrected quantum computer could perform these simulations more efficiently than any classical computer. In this work, we describe a method for mapping any finite nonlinear dynamical system to an infinite linear dynamical system (embedding) and detail three specific cases of this method that correspond to previously studied mappings. Then we explore an approach for approximating the resulting infinite linear system with finite linear systems (truncation). Using a number of qubits only logarithmic in the number of variables of the nonlinear system, a quantum computer could simulate truncated systems to approximate output quantities if the nonlinearity is sufficiently weak. Other aspects of the computational efficiency of the three detailed embedding strategies are also discussed.

97 MATHEMATICS AND COMPUTING↗

Learning Canonical Embeddings for Unsupervised Shape Correspondence With Locally Linear Transformations

We present a new approach to unsupervised shape correspondence learning between pairs of point clouds. We make the first attempt to adapt the classical locally linear embedding algorithm (LLE)-originally designed for nonlinear dimensionality reduction-for shape correspondence. The key idea is to find dense correspondences between shapes by first obtaining high-dimensional neighborhood-preserving embeddings of low-dimensional point clouds and subsequently aligning the source and target embeddings using locally linear transformations. We demonstrate that learning the embedding using a new LLE-inspired point cloud reconstruction objective results in accurate shape correspondences. More specifically, the approach comprises an end-to-end learnable framework of extracting high-dimensional neighborhood-preserving embeddings, estimating locally linear transformations in the embedding space, and reconstructing shapes via divergence measure-based alignment of probability density functions built over reconstructed and target shapes. Our approach enforces embeddings of shapes in correspondence to lie in the same universal/canonical embedding space, which eventually helps regularize the learning process and leads to a simple nearest neighbors approach between shape embeddings for finding reliable correspondences. Comprehensive experiments show that the new method makes noticeable improvements over state-of-the-art approaches on standard shape correspondence benchmark datasets covering both human and nonhuman shapes.

deformation↗

Uncovering interpretable relationships in high-dimensional scientific data through function preserving projections

Abstract In many fields of science and engineering, we frequently encounter experiments or simulations datasets that describe the behavior of complex systems and uncovering human interpretable patterns between their inputs and outputs via exploratory data analysis is essential for building intuition and facilitating discovery. Often, we resort to 2D embeddings for examining these high-dimensional relationships (e.g. dimensionality reduction). However, most existing embedding methods treat the dimensions as coordinates for samples in a high-dimensional space, which fail to capture the potential functional relationships, and the few methods that do take function into consideration either only focus on linear patterns or produce non-linear embeddings that are hard to interpret. To address these challenges, we proposed function preserving projections (FPP), which construct 2D linear embeddings optimized to reveal interpretable yet potentially non-linear patterns between the domain and the range of a high-dimensional function. The intuition here is that humans are good at understanding potentially non-linear patterns in 2D but unable to interpret non-linear mapping from high-dimensional space to 2D. Therefore, we should restrict the projection to linear but not the pattern we are seeking. Using FPP on real-world datasets, one can obtain fundamentally new insights about high-dimensional relationships in extremely large datasets that could not be processed with existing dimension reduction methods.

97 MATHEMATICS AND COMPUTING↗

Function Preserving Projection

With function preserving projection (FPP), we aim to address high-dimensional data exploration challenges common in many scientific applications. Instead of preserving the global or local geometric structure of a dataset like traditional dimensionality reduction methods (e.g., principal component analysis, t-stochastic neighborhood embedding), the proposed approach finds a linear projection that reveal interpretable (yet potentially nonlinear) pattern of the properties of interests defined on a highdimensional domain (e.g., simulation yield defined on the input parameter space). The key insight driving FPP is the fact that while non-linear embeddings are difficult to interpret, humans are highly skilled in understanding complex visual patterns. FPP lets users select the complexity of a pattern they consider interpretable, allows to find a joint projection that explains multiple functions, and provides a confidence statistic (akin to a p-value) indicating how salient the resulting pattern is. We demonstrate the efficacy of the proposed method on several synthetic and scientific datasets for both continuous (regression) and discrete (classification) problems of interests and illustrate the scalability of FPP by finding an optimal projection for close to 40 gigabytes in around 10 minutes.

Liu, SHUSEN↗

Comparative analysis of nonlinear dimensionality reduction techniques for breast MRI segmentation

Purpose: Visualization of anatomical structures using radiological imaging methods is an important tool in medicine to differentiate normal from pathological tissue and can generate large amounts of data for a radiologist to read. Integrating these large data sets is difficult and time-consuming. A new approach uses both supervised and unsupervised advanced machine learning techniques to visualize and segment radiological data. This study describes the application of a novel hybrid scheme, based on combining wavelet transform and nonlinear dimensionality reduction (NLDR) methods, to breast magnetic resonance imaging (MRI) data using three well-established NLDR techniques, namely, ISOMAP, local linear embedding (LLE), and diffusion maps (DfM), to perform a comparative performance analysis. Methods: Twenty-five breast lesion subjects were scanned using a 3T scanner. MRI sequences used were T1-weighted, T2-weighted, diffusion-weighted imaging (DWI), and dynamic contrast-enhanced (DCE) imaging. The hybrid scheme consisted of two steps: preprocessing and postprocessing of the data. The preprocessing step was applied for B{sub 1} inhomogeneity correction, image registration, and wavelet-based image compression to match and denoise the data. In the postprocessing step, MRI parameters were considered data dimensions and the NLDR-based hybrid approach was applied to integrate the MRI parameters into a single image, termed the embedded image. This was achieved by mapping all pixel intensities from the higher dimension to a lower dimensional (embedded) space. For validation, the authors compared the hybrid NLDR with linear methods of principal component analysis (PCA) and multidimensional scaling (MDS) using synthetic data. For the clinical application, the authors used breast MRI data, comparison was performed using the postcontrast DCE MRI image and evaluating the congruence of the segmented lesions. Results: The NLDR-based hybrid approach was able to define and segment both synthetic and clinical data. In the synthetic data, the authors demonstrated the performance of the NLDR method compared with conventional linear DR methods. The NLDR approach enabled successful segmentation of the structures, whereas, in most cases, PCA and MDS failed. The NLDR approach was able to segment different breast tissue types with a high accuracy and the embedded image of the breast MRI data demonstrated fuzzy boundaries between the different types of breast tissue, i.e., fatty, glandular, and tissue with lesions (>86%). Conclusions: The proposed hybrid NLDR methods were able to segment clinical breast data with a high accuracy and construct an embedded image that visualized the contribution of different radiological parameters.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Hyperspectral target detection using manifold learning and multiple target spectra

Imagery collected from satellites and airborne platforms provides an important tool for remotely analyzing the content of a scene. In particular, the ability to remotely detect a specific material within a scene is of critical importance in nonproliferation and other applications. The sensor systems that process hyperspectral images collect the high-dimensional spectral information necessary to perform these detection analyses. For a d-dimensional hyperspectral image, however, where d is the number of spectral bands, it is common for the data to inherently occupy an m-dimensional space with m << d. In the remote sensing community, this has led to recent interest in the use of manifold learning, which seeks to characterize the embedded lower-dimensional, nonlinear manifold that the data discretely approximate. The research presented in this paper focuses on a graph theory and manifold learning approach to target detection, using an adaptive version of locally linear embedding that is biased to separate target pixels from background pixels. Finally, this approach incorporates multiple target signatures for a particular material, accounting for the spectral variability that is often present within a solid material of interest.

97 MATHEMATICS AND COMPUTING↗

Time-series forecasting using manifold learning, radial basis function interpolation, and geometric harmonics

We address a three-tier numerical framework based on nonlinear manifold learning for the forecasting of high-dimensional time series, relaxing the “curse of dimensionality” related to the training phase of surrogate/machine learning models. At the first step, we embed the high-dimensional time series into a reduced low-dimensional space using nonlinear manifold learning (local linear embedding and parsimonious diffusion maps). Then, we construct reduced-order surrogate models on the manifold (here, for our illustrations, we used multivariate autoregressive and Gaussian process regression models) to forecast the embedded dynamics. Finally, we solve the pre-image problem, thus lifting the embedded time series back to the original high-dimensional space using radial basis function interpolation and geometric harmonics. The proposed numerical data-driven scheme can also be applied as a reduced-order model procedure for the numerical solution/propagation of the (transient) dynamics of partial differential equations (PDEs). In conclusion, we assess the performance of the proposed scheme via three different families of problems: (a) the forecasting of synthetic time series generated by three simplistic linear and weakly nonlinear stochastic models resembling electroencephalography signals, (b) the prediction/propagation of the solution profiles of a linear parabolic PDE and the Brusselator model (a set of two nonlinear parabolic PDEs), and (c) the forecasting of a real-world data set containing daily time series of ten key foreign exchange rates spanning the time period 3 September 2001–29 October 2020.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Inverse design of two-dimensional graphene/h-BN hybrids by a regressional and conditional GAN

Design of materials with desired properties is currently laborious and heavily relies on intuition of researchers through a trial-and-error process. To tackle this challenge, in this work we propose a novel regressional and conditional generative adversarial network (RCGAN) for inverse design of representative two-dimensional materials, the graphene and boron-nitride (BN) hybrids. RCGAN incorporates a supervised regressor network, thus overcoming the common technical barrier in the traditional unsupervised GANs, which cannot generate data when fed with continuous and quantitative labels. RCGAN can autonomously generate graphene/BN hybrids given any target bandgap values. These structures are distinguished from the ones used for training and exhibit high diversity for a given bandgap. Moreover, they exhibit high fidelity, yielding bandgaps within ~10% MAE F of the desired bandgaps as validated by density functional theory (DFT) calculations. Analysis by the principle component analysis (PCA) and modified locally linear embedding (MLLE) reveals that the generator has successfully generated structures following the statistical distribution of the real structures. It implies the possibility of the RCGAN in recognizing physical rules hidden in the high-dimensional data. The novel strategy for designing regressional GAN architecture together with the successful application to inverse design of materials would inspire further exploration in research fields beyond materials.

36 MATERIALS SCIENCE↗

An efficient explicit implementation of a near-optimal quantum algorithm for simulating linear dissipative differential equations

We propose an efficient block-encoding technique for the implementation of the Linear Combination of Hamiltonian Simulations (LCHS) for simulating dissipative initial-value problems. This algorithm approximates a target nonunitary operator as a weighted sum of Hamiltonian evolutions, thereby emulating a dissipative problem by mixing various time scales. We introduce an efficient encoding of the LCHS into a quantum circuit based on a simple coordinate transformation that turns the dependence on the summation index into a trigonometric function. Classically, this method is equivalent to the use of a highly accurate Fejér-Clenshaw-Curtis quadrature formula. Quantumly, this significantly simplifies block-encoding of a dissipative problem and allows one to perform an exponential number of Hamiltonian simulations by a single Quantum Signal Processing (QSP) circuit. The resulting LCHS circuit has high success probability and the selector scales logarithmically with the number of terms in the LCHS sum and linearly with time. Careful analysis of error convergence proves that this method is more efficient than other LCHS circuits that have recently appeared in the literature. We verify the quantum circuit and its scaling by simulating it on a digital emulator of fault-tolerant quantum computers and, as a test problem, solve the advection-diffusion equation. The proposed algorithm can be used for simulating a wide class of nonunitary initial-value problems including the Liouville equation with added dissipation and linear embeddings of nonlinear systems, such as the Koopman-von Neumann and Carleman embeddings.

Novikau, I [Lawrence Livermore National Laboratory↗

CHROMOSPHERICALLY ACTIVE STARS IN THE RADIAL VELOCITY EXPERIMENT (RAVE) SURVEY. I. THE CATALOG

RAVE, the unbiased magnitude limited survey of southern sky stars, contained 456,676 medium-resolution spectra at the time of our analysis. Spectra cover the Ca II infrared triplet (IRT) range, which is a known indicator of chromospheric activity. Our previous work classified all spectra using locally linear embedding. It identified 53,347 cases with a suggested emission component in calcium lines. Here, we use a spectral subtraction technique to measure the properties of this emission. Synthetic templates are replaced by the observed spectra of non-active stars to bypass the difficult computations of non-local thermal equilibrium profiles of the line cores and stellar parameter dependence. We derive both the equivalent width of the excess emission for each calcium line on a 5 Å wide interval and their sum EW{sub IRT} for ∼44,000 candidate active dwarf stars with signal-to-noise ratio >20, with no cuts on the basis of the source of their emission flux. From these, ∼14,000 show a detectable chromospheric flux with at least a 2σ confidence level. Our set of active stars vastly enlarges previously known samples. Atmospheric parameters and, in some cases, radial velocities of active stars derived from automatic pipelines suffer from systematic shifts due to their shallower calcium lines. We re-estimate the effective temperature, metallicity, and radial velocities for candidate active stars. The overall distribution of activity levels shows a bimodal shape, with the first peak coinciding with non-active stars and the second with the pre-main-sequence cases. The catalog will be made publicly available with the next RAVE public data releases.

79 ASTRONOMY AND ASTROPHYSICS↗

EXPLORING THE MORPHOLOGY OF RAVE STELLAR SPECTRA

The RAdial Velocity Experiment (RAVE) is a medium-resolution (R {approx} 7500) spectroscopic survey of the Milky Way that has already obtained over half a million stellar spectra. They present a randomly selected magnitude-limited sample, so it is important to use a reliable and automated classification scheme that identifies normal single stars and discovers different types of peculiar stars. To this end, we present a morphological classification of {approx}350, 000 RAVE survey stellar spectra using locally linear embedding, a dimensionality reduction method that enables representing the complex spectral morphology in a low-dimensional projected space while still preserving the properties of the local neighborhoods of spectra. We find that the majority of all spectra in the database ({approx} 90%-95%) belong to normal single stars, but there is also a significant population of several types of peculiars. Among them, the most populated groups are those of various types of spectroscopic binary and chromospherically active stars. Both of them include several thousands of spectra. Particularly the latter group offers significant further investigation opportunities since activity of stars is a known proxy of stellar ages. Applying the same classification procedure to the sample of normal single stars alone shows that the shape of the projected manifold in two-dimensional space correlates with stellar temperature, surface gravity, and metallicity.

79 ASTRONOMY AND ASTROPHYSICS↗

Reinforcement Learning via Gaussian Processes with Neural Network Dual Kernels

While deep neural networks (DNNs) and Gaussian Processes (GPs) are both popularly utilized to solve problems in reinforcement learning, both approaches feature undesirable drawbacks for challenging problems. DNNs learn complex non-linear embeddings, but do not naturally quantify uncertainty and are often data-inefficient to train. GPs infer posterior distributions over functions, but popular kernels exhibit limited expressivity on complex and high-dimensional data. Fortunately, recently discovered conjugate and neural tangent kernel functions encode the behavior of overparameterized neural networks in the kernel domain. We demonstrate that these kernels can be efficiently applied to regression and reinforcement learning problems by analyzing a baseline case study.We apply GPs with neural network dual kernels to solve reinforcement learning tasks for the first time. We demonstrate, using the well understood mountain-car problem, that GPs empowered with dual kernels perform at least as well as those using the conventional radial basis function kernel. Finally, we conjecture that by inheriting the probabilistic rigor of GPs and the powerful embedding properties of DNNs, GPs using NN dual kernels will empower future reinforcement learning models on difficult domains.

97 MATHEMATICS AND COMPUTING↗

Dimensionality Reduction Library v 0.2

Dimensionality Reduction Library is a C++ library for dimensionality reduction. In the context of this library, dimensionality reduction is considered to consist of 1)estimation of the intrinsic dimensionality using sampled data, 2) Finding maps that reduce the dimensionality of data (forward map) or increase the dimensionality of data (reverse map) and 3) mapping arbitray coordinates to high and low dimensionalities. The library is intended to provide a consistent interface to multiple dimensionality reduction algorithms with an efficient C++ interface that runs efficiently on multicore architectures. A few routines have been optimized with an option for GPU acceleration or distributed computation. Currently the library offers intrinsic dimensionality estimation using point-PCA reconstruction error and/ residual variance. The following dimensionality reduction methods have been implemented: Principal Component Analysis Multidimensional Scaling Locally Linear Embedding IsoMap Autoencoder Neutral Networks An executable is also supplied that can be built to allow for command-line access to the library routines. A description for an application of the library for molecular structure analysis has been published.

BROWN, WILLIAM↗

Dimensionality Reduction Library v 0.2

Dimensionality Reduction Library is a C++ library for dimensionality reduction. In the context of this library, dimensionality reduction is considered to consist of 1)estimation of the intrinsic dimensionality using sampled data, 2) Finding maps that reduce the diemsionality of data (forward map) or increase the dimensionality of data (reverse map) and 3) mapping arbitray coordiantes to high and low dimensionalities. The library is intended toprovide a consistent interface to multiple dimensionality reduction algorithms with an efficient C++ interface that runs efficiently on multicore architectures. A few routines have been optimized with an option for GPU acceleration or distributed computation. Currently the library offers intrinsic dimensionality estimation using point-PCA reconstruction error and/ residual variance. The following dimensionality reduction methods have been implemented: Principal Component Analysis Multidimensional Scaling Locally Linear Embedding IsoMap Autoencoder Neutral Networks An executable is also supplied that can be built to allow for command-line access to the library routines. A description for an applciation of the library for molecular structure analysis has been published.

BROWN, WILLIAM↗