Common Kernel Research Articles

Cryo-EM of full-length \u03b1-synuclein reveals fibril polymorphs with a common structural kernel

α-Synuclein (aSyn) fibrillar polymorphs have distinct in vitro and in vivo seeding activities, contributing differently to synucleinopathies. Despite numerous prior attempts, how polymorphic aSyn fibrils differ in atomic structure remains elusive. Here, we present fibril polymorphs from the full-length recombinant human aSyn and their seeding capacity and cytotoxicity in vitro. By cryo-electron microscopy helical reconstruction, we determine the structures of the two predominant species, a rod and a twister, both at 3.7 Å resolution. Our atomic models reveal that both polymorphs share a kernel structure of a bent β-arch, but differ in their inter-protofilament interfaces. Thus, different packing of the same kernel structure gives rise to distinct fibril polymorphs. Analyses of disease-related familial mutations suggest their potential contribution to the pathogenesis of synucleinopathies by altering population distribution of the fibril polymorphs. Drug design targeting amyloid fibrils in neurodegenerative diseases should consider the formation and distribution of concurrent fibril polymorphs.

Marking imprecision, conveying surprise:Likebetween hedging and mirativity

Mirativeexpressions, which mark surprising information (DeLancey 1997), are often expressed through linguistic forms that are also used to encode other, seemingly unrelated, meanings – e.g. evidential markers that mark lack of direct evidence (Turkish: Slobin & Aksu 1982, Peterson 2010; Cheyenne: Rett & Murray 2013; Cuzco Quechua: Faller 2002; Ostyak: Nikolaeva 1999; among others). In this paper, we show that the English particlelikefeatures a parallel polysemy between a mirative use and its better-knownhedginguse, which expresses weakened commitment to the strict denotation of a linguistic expression. After presenting several diagnostics that point to a genuine empirical difference between the hedging and mirative functions oflike, we propose that both uses widen the size of a contextually restricted set, admitting elements that were previously excluded. More specifically, hedginglikeexpands the set of ‘similar enough’ interpretations that we can apply to a linguistic expression in a context, including interpretations that we would normally consider to be too different from the context at hand. Mirativelike, on the other hand, expands the set of worlds that we are willing to consider as candidates for the actual world in the conversation, resulting in the inclusion of worlds that interlocutors have previously ruled out due to perceived outlandishness. We therefore suggest that the two uses are best treated as sharing a common semantic kernel, deriving hedging and mirativity as effects of the particular type of object to whichlikeapplies.

Numerically Calculated 3D Space-Weighting Functions to Image Crustal Volcanic Structures Using Diffuse Coda Waves

Seismic coda measurements retrieve parameters linked to the physical characteristics of rock volumes illuminated by high frequency scattered waves. Space weighting functions (SWF) and kernels are different tools that model the spatial sensitivity of coda envelopes to scattering and absorption anomalies in these rock matrices, allowing coda-wave attenuation ( Q c o d a ) imaging. This note clarifies the difference between SWF and sensitivity kernels developed for coda wave imaging. It extends the SWF previously developed in 2D to the third dimension by using radiative transfer and the diffusion equation, based on the assumption that variations of Q c o d a depend solely on variations of the extinction length. When applied to active data (Deception Island, Antarctica), 3D SWF images strongly resemble 2D images, making this 3D extension redundant. On the other hand, diffusion does not efficiently model coda waveforms when using earthquake datasets spanning depths between 0 and 20 km, such as at Mount St. Helens volcano. In this setting, scattering attenuation and absorption suffer tradeoffs and cannot be separated by fitting a single seismogram energy envelope for SWF imaging. We propose that an approximate analytical 3D SWF, similar in shape to the common coda kernels used in literature, can still be used in a space weighted back-projection approach. While Q c o d a is not a physical parameter of the propagation medium, its spatially-dependent modeling allows improved reconstruction of crustal-scale tectonic and geological features. It is even more efficient as a velocity independent imaging tool for magma and fluid storage when applied to deep volcanism.

QR codes blind deconvolution algorithm based on binary characteristic and L0 norm minimization

In recent years, QR code has been widely used in various society fields, and has provided great convenience in information exchange, commodity payment and website registration. However, the degradation of QR code images can cause difficulty in reading the code information. In this paper, we propose an alternating minimization restoration model for QR code images. Based on the binary prior, we use the binary characteristic and L0 norm as the regularization terms, and introduce different auxiliary variables to make the complex model become solvable. The bi-level constraint of bar code image maintains the sparseness of image pixels and gradient. The experimental result shows that our algorithm has excellent effect on common blur kernels and multiple noise, and its performance is better than the state-of-the-art algorithms.

Fourier Feature Approximations for Periodic Kernels in Time-Series Modelling

Gaussian Processes (GPs) provide an extremely powerful mechanism to model a variety of problems but incur an O(N3) complexity in the number of data samples. Common approximation methods rely on what are often termed inducing points but still typically incur an O(NM2) complexity in the data and corresponding inducing points. Using Random Fourier Feature (RFF) maps, we overcome this by transforming the problem into a Bayesian Linear Regression formulation upon which we apply a Bayesian Variational treatment that also allows learning the corresponding kernel hyperparameters, likelihood and noise parameters. In this paper we introduce an alternative method using Fourier series to obtain spectral representations of common kernels, in particular for periodic warpings, which surprisingly have a convergent, non-random form using special functions, requiring fewer spectral features to approximate their corresponding kernel to high accuracy. Using this, we can fuse the Random Fourier Feature spectral representations of common kernels with their periodic counterparts to show how they can more effectively and expressively learn patterns in time-series for both interpolation and extrapolation. This method combines robustness, scalability and equally importantly, interpretability through a symbolic declarative grammar that is both functionally and humanly intuitive — a property that is crucial for explainable decision making. Using probabilistic programming and Variational Inference we are able to efficiently optimise over these rich functional representations. We show significantly improved Gram matrix approximation errors, and also demonstrate the method in several time-series problems comparing other commonly used approaches such as recurrent neural networks.

Concentric layered Hermite scatterers

The long wavelength limit of scattering from spheres has a rich history in optics, electromagnetics, and acoustics. Recently it was shown that a common integral kernel pertains to formulations of weak spherical scatterers in both acoustics and electromagnetic regimes. Furthermore, the relationship between backscattered amplitude and wavenumber k was shown to follow power laws higher than the Rayleigh scattering k2 power law, when the inhomogeneity had a material composition that conformed to a Gaussian weighted Hermite polynomial. Although this class of scatterers, called Hermite scatterers, are plausible, it may be simpler to manufacture scatterers with a core surrounded by one or more layers. In this case the inhomogeneous material property conforms to a piecewise continuous constant function. We demonstrate that the necessary and sufficient conditions for supra-Rayleigh scattering power laws in this case can be stated simply by considering moments of the inhomogeneous function and its spatial transform. This development opens an additional path for construction of, and use of scatterers with unique power law behavior.

A novel Monogenic Directional Pattern (MDP) and pseudo-Voigt kernel for facilitating the identification of facial emotions

Facial expressions are the best way of communicating human emotions. This paper proposes a novel Monogenic Directional Pattern (MDP) for extracting features from the face. To reduce the time spent on choosing the best kernel, a novel pseudo-Voigt kernel is chosen as the common kernel for dimension reduction proposed as pseudo-Voigt kernel-based Generalized Discriminant Analysis (PVK-GDA). The pseudo-Voigt kernel-based Extreme Learning Machine (PVK-ELM) is used for better recognition of facial emotions. The efficiency of the approach is proved by experimenting with the Japanese Female Facial Expression (JAFFE), Cohn Kanade (CK+), Multimedia Understanding Group (MUG), Static Facial Expressions in the Wild (SFEW) and Oulu-Chinese Academy of Science, Institute of Automation (Oulu-CASIA) datasets. This approach achieves better classification accuracy of 96.7% for JAFFE, 99.4% for CK+, 98.6% for MUG, 35.6% for SFEW and 88% for Oulu-CASIA, which is certainly higher when compared to other techniques in the literature.

Motor imagery EEG classification with optimal subset of wavelet based common spatial pattern and kernel extreme learning machine.

Performance of motor imagery based brain-computer interfaces (MI BCIs) greatly depends on how to extract the features. Various versions of filter-bank based common spatial pattern have been proposed and used in MI BCIs. Filter-bank based common spatial pattern has more number of features compared with original common spatial pattern. As the number of features increases, the MI BCIs using filter-bank based common spatial pattern can face overfitting problems. In this study, we used eigenvector centrality feature selection method, wavelet packet decomposition common spatial pattern, and kernel extreme learning machine to improve the performance of MI BCIs and avoid overfitting problems. Furthermore, the computational speed was improved by using kernel extreme learning machine.

Hiding the Long Latency of Persist Barriers Using Speculative Execution

Byte-addressable non-volatile memory technology is emerging as an alternative for DRAM for main memory. This new Non-Volatile Main Memory (NVMM) allows programmers to store important data in data structures in memory instead of serializing it to the file system, thereby providing a substantial performance boost. However, modern systems reorder memory operations and utilize volatile caches for better performance, making it difficult to ensure a consistent state in NVMM. Intel recently announced a new set of persistence instructions, clflushopt, clwb, and pcommit. These new instructions make it possible to implement fail-safe code on NVMM, but few workloads have been written or characterized using these new instructions. In this work, we describe how these instructions work and how they can be used to implement write-ahead logging based transactions. We implement several common data structures and kernels and evaluate the performance overhead incurred over traditional non-persistent implementations. In particular, we find that persistence instructions occur in clusters along with expensive fence operations, they have long latency, and they add a significant execution time overhead, on average by 20.3% over code with logging but without fence instructions to order persists. To deal with this overhead and alleviate the performance bottleneck, we propose to speculate past long latency persistency operations using checkpoint-based processing. Our speculative persistence architecture reduces the execution time overheads to only 3.6%.

A Novel Extreme Learning Machine Classification Model for e-Nose Application Based on the Multiple Kernel Approach.

A novel classification model, named the quantum-behaved particle swarm optimization (QPSO)-based weighted multiple kernel extreme learning machine (QWMK-ELM), is proposed in this paper. Experimental validation is carried out with two different electronic nose (e-nose) datasets. Being different from the existing multiple kernel extreme learning machine (MK-ELM) algorithms, the combination coefficients of base kernels are regarded as external parameters of single-hidden layer feedforward neural networks (SLFNs). The combination coefficients of base kernels, the model parameters of each base kernel, and the regularization parameter are optimized by QPSO simultaneously before implementing the kernel extreme learning machine (KELM) with the composite kernel function. Four types of common single kernel functions (Gaussian kernel, polynomial kernel, sigmoid kernel, and wavelet kernel) are utilized to constitute different composite kernel functions. Moreover, the method is also compared with other existing classification methods: extreme learning machine (ELM), kernel extreme learning machine (KELM), k-nearest neighbors (KNN), support vector machine (SVM), multi-layer perceptron (MLP), radical basis function neural network (RBFNN), and probabilistic neural network (PNN). The results have demonstrated that the proposed QWMK-ELM outperforms the aforementioned methods, not only in precision, but also in efficiency for gas classification.

Predicting the spread of all invasive forest pests in the United States.

We tested whether a general spread model could capture macroecological patterns across all damaging invasive forest pests in the United States. We showed that a common constant dispersal kernel model, simulated from the discovery date, explained 67.94% of the variation in range size across all pests, and had 68.00% locational accuracy between predicted and observed locational distributions. Further, by making dispersal a function of forest area and human population density, variation explained increased to 75.60%, with 74.30% accuracy. These results indicated that a single general dispersal kernel model was sufficient to predict the majority of variation in extent and locational distribution across pest species and that proxies of propagule pressure and habitat invasibility - well-studied predictors of establishment - should also be applied to the dispersal stage. This model provides a key element to forecast novel invaders and to extend pathway-level risk analyses to include spread.

Energy-Efficient Adaptive Computing With Multifunctional Memory

Digital memory arrays, which serve as an integral part of modern computing systems, are traditionally used for information storage. However, recent reports show that memory can be used on demand as a reconfigurable computing resource, drastically improving energy efficiency for many applications. In this case, memory usage is limited to storing function responses as multi-input multi-output lookup tables. In this brief, we propose a novel multifunctional memory (MFM) framework, which can function as typical memory for storage as well as in a neuroinspired computing mode. The system is based on a modified memory array, which can be dynamically switched between these two modes. Using a promising emerging memory device, namely, resistive random access memory, we present device-level engineering, circuit-level modifications, and appropriate architecture to realize the MFM framework. Simulation results demonstrate significant improvements in both energy efficiency and performance compared to a general-purpose processor, a field-programmable gate array, and a recent memory-based, reconfigurable computing framework (MAHA) for a set of common application kernels.

An Extended Shared Logarithmic Unit for Nonlinear Function Kernel Acceleration in a 65-nm CMOS Multicore Cluster

Energy-efficient computing and ultralow-power computing are strong requirements for various application areas, such as internet of things and wearables. While for some applications integer and fixed-point arithmetic suffice, others require a larger dynamic range, typically obtained using floating-point (FP) numbers. Logarithmic number systems (LNSs) have been proposed as energy-efficient alternative, since several complex FP operations translate into simple integer operations. However, additions and subtractions become nonlinear operations, which have to be approximated via interpolation. Even efficient LNS units (LNUs) are still larger than standard FP units (FPUs), rendering them impractical for most general-purpose processors. We show that, when shared among several cores, LNUs become a very attractive solution. A series of compact LNUs is developed, which provide significantly more functionality (such as transcendental functions) than other state-of-the-art designs. This allows, for example, to evaluate the atan2 function with three instructions for only 183.2 pJ/op at 0.8 V. We present the first shared-LNU architecture where these LNUs have been integrated into a multicore system with four 32-b-OpenRISC cores and show measurement results demonstrating that the shared-LNU design can be up to 4.1× more energy-efficient in common nonlinear processing kernels, compared with a similar area design with four private FPUs.

Energy-Efficient Adaptive Hardware Accelerator for Text Mining Application Kernels

Text mining is a growing field of applications, which enables the analysis of large text data sets using statistical methods. In recent years, exponential increase in the size of these data sets has strained existing systems, requiring more computing power, server hardware, networking interconnects, and power consumption. For practical reasons, this trend cannot continue in the future. Instead, we propose a reconfigurable hardware accelerator designed for text analytics systems, which can simultaneously improve performance and reduce power consumption. Situated near the last level of memory, it mitigates the need for high-bandwidth processor-to-memory connections, instead capitalizing on close data proximity, massively parallel operation, and analytic-inspired functional units to maximize energy efficiency, while remaining flexible to easily map common text analytic kernels. A field-programmable gate array-based emulation framework demonstrates the functional correctness of the system, and a full eight-core accelerator is synthesized for power, area, and delay estimates. The accelerator can achieve two to three orders of magnitude improvement in energy efficiency versus CPU and general-purpose graphics processing unit (GPU) for various text mining kernels. As a case study, we demonstrate how indexing performance of Lucene, a popular text search and analytics platform, can be improved by an average of 70% over CPU and GPU while significantly reducing data transfer energy and latency.

GPU accelerated spectral finite elements on all-hex meshes

This paper presents a spectral element finite element scheme that efficiently solves elliptic problems on unstructured hexahedral meshes. The discrete equations are solved using a matrix-free preconditioned conjugate gradient algorithm. An additive Schwartz two-scale preconditioner is employed that allows h-independence convergence. An extensible multi-threading programming API is used as a common kernel language that allows runtime selection of different computing devices (GPU and CPU) and different threading interfaces (CUDA, OpenCL and OpenMP). Performance tests demonstrate that problems with over 50 million degrees of freedom can be solved in a few seconds on an off-the-shelf GPU.

Tuning kernel parameters for SVM based on expected square distance ratio

The performance of a support vector machine (SVM) depends highly on the selection of the kernel function type and relevant parameters. To choose the kernel parameters properly, methods analyzing the class separability have been widely adopted for their efficiency compared with other methods, such as the popular grid search algorithm. This paper proposes a novel index called the Expected Square Distance Ratio (ESDR), which can serve as a better class separability criterion than the existing ones. Experiments on real-world datasets show that, compared with common kernel parameter selection methods that utilize the between-class separation, the variations in ESDR with respect to the kernel parameter are much more in line with those of the classification accuracy, leading to better kernel parameters. Moreover, ESDR takes the exact data distribution into account and can thus be used to study the model selection problem of an SVM for certain forms of data distribution. As an example, we employ the ESDR to analyze the selection of RBF (Radial Basis Function) kernel parameters for Gaussian data classification.

New Hermite orthogonal polynomial kernel and combined kernels in Support Vector Machine classifier

Support Vector Machine is a desired method for classification of different types of data, but the main obstacle to using this method is the considerable reduction of classification speed upon increase in the size of problem. In this paper, a new kernel function is proposed for SVM which is derived from Hermite orthogonal polynomials. This function improves classification accuracy as well as reduction in the number of support vectors and increases the classification speed. The overall kernel performance has been evaluated on real world data sets from UCI repository by the ten-fold cross-validation method. The combinations of Hermite function with common kernels are proposed in this paper. Experimental results reveal that the Hermite–Chebyshev kernel which is obtained from combination of Hermite and Chebyshev kernel, has the best performance in the number of support vectors and Hermite–Gussian kernel that is produced from combination Hermite and Gussian kernel, has the first rank in the error rate among all experimental kernels. On the other hand the Hermite proposed method has the least number of support vectors and best performance in the error rate in comparison with common kernels, and lowest required time among all kernels.

A Gaussian-like immersed-boundary kernel with three continuous derivatives and improved translational invariance

The immersed boundary (IB) method is a general mathematical framework for studying problems involving fluid–structure interactions in which an elastic structure is immersed in a viscous incompressible fluid. In the IB formulation, the fluid described by Eulerian variables is coupled with the immersed structure described by Lagrangian variables via the use of the Dirac delta function. From a numerical standpoint, the Lagrangian force spreading and the Eulerian velocity interpolation are carried out by a regularized, compactly supported discrete delta function, which is assumed to be a tensor product of a single-variable immersed-boundary kernel. IB kernels are derived from a set of postulates designed to achieve approximate grid translational invariance, interpolation accuracy and computational efficiency. In this note, we present a new 6-point immersed-boundary kernel that is C3 and yields a substantially improved translational invariance compared to other common IB kernels.

Properties of Transmetric Density Estimation

Transmetric density estimation is a generalization of kernel density estimation that is proposed in Hovda(2014) and Hovda (2016), This framework involves the possibility of making assumptions on the kernel of the distribution to improve convergence orders and to reduce the number of dimensions in the graphical display. In this paper we show that several state-of-the-art nonparametric, semiparametric and even parametric methods are special cases of this formulation, meaning that there is a unified approach. Moreover, it is shown that parameters can be trained using unbiased cross-validation. When parameter estimation is included, the mean integrated squared error of the transmetric density estimator is lower than for the common kernel density estimator, when the number of dimensions is larger than two.

Shadow Kernels

Existing operating systems share a common kernel text section amongst all processes. It is not possible to perform kernel specialization or tuning such that different applications execute text optimized for their kernel use despite the benefits of kernel specialization for performance guided optimization, exokernels, kernel fastpaths, and cheaper hardware access. Current specialization primitives involve system wide changes to kernel text, which can have adverse effects on other processes sharing the kernel due to the global side-effects. We present shadow kernels: a primitive that allows multiple kernel text sections to coexist in a contemporary operating system. By remapping kernel virtual memory on a context-switch, or for individual system calls, we specialize the kernel on a fine-grained basis. Our implementation of shadow kernels uses the Xen hypervisor so can be applied to any operating system that runs on Xen.