A new family of life distributions for dependent data: Estimation

One of the biggest problems in reliability analysis is determining an appropriate distribution of life data. Therefore, this paper develops the estimation aspect of a family of life distributions obtained from spherical distributions. Additionally, a new family of life distributions is proposed for dependent life data, together with an optimization algorithm based on the simulated annealing method. This algorithm is very efficient for optimization purposes and does not require any manipulation of the log-likelihood functions for the distributions proposed in this study.     

Tensor rank one differential graph preserving analysis for facial expression recognition

This paper presents a new dimensionality reduction algorithm for multi-dimensional data based on the tensor rank-one decomposition and graph preserving criterion. Through finding proper rank-one tensors, the algorithm effectively enhances the pairwise inter-class margins and meanwhile preserves the intra-class local manifold structure. In the algorithm, a novel marginal neighboring graph is devised to describe the pairwise inter-class boundaries, and a differential formed objective function is adopted to ensure convergence. Furthermore, the algorithm has less computation in comparison with the vector representation based and the tensor-to-tensor projection based algorithms. The experiments for the basic facial expressions recognition show its effectiveness, especially when it is followed by a neural network classifier.     

Functional signatures:new definition and constructions

Functional signatures (FS) enable a master authority to delegate its signing privilege to an assistant.Concretely,the master authority uses its secret key skF t...     

A new definition of SLDNF-resolution

We propose a new, “top-down” definition of SLDNF-resolution that retains the spirit of the original definition, but avoids the difficulties noted in the literature. We compare it with the “bottom-up” definition of Kunen [7].     

Hierarchical tensor approximation of multi-dimensional visual data

Visual data comprise of multi-scale and inhomogeneous signals. In this paper, we exploit these characteristics and develop a compact data representation technique based on a hierarchical tensor-based transformation. In this technique, an original multi-dimensional dataset is transformed into a hierarchy of signals to expose its multi-scale structures. The signal at each level of the hierarchy is further divided into a number of smaller tensors to expose its spatially inhomogeneous structures. These smaller tensors are further transformed and pruned using a tensor approximation technique. Our hierarchical tensor approximation supports progressive transmission and partial decompression. Experimental results indicate that our technique can achieve higher compression ratios and quality than previous methods, including wavelet transforms, wavelet packet transforms, and single-level tensor approximation. We have successfully applied our technique to multiple tasks involving multi-dimensional visual data, including medical and scientific data visualization, data-driven rendering and texture synthesis.     

A new definition of shape similarity

Two reference points of a region are defined which do not depend on the position, size and orientation of the region. Reference points are used to get borders on the basis of which the shape distance and shape similarity are defined.     

Computing rank dependent utility in graphical models for sequential decision problems

This paper is devoted to automated sequential decision in AI. More precisely, we focus here on the Rank Dependent Utility (RDU) model. This model is able to encompass rational decision behaviors that the Expected Utility model cannot accommodate. However, the non-linearity of RDU makes it difficult to compute an RDU-optimal strategy in sequential decision problems. This has considerably slowed the use of RDU in operational contexts. In this paper, we are interested in providing new algorithmic solutions to compute an RDU-optimal strategy in graphical models. Specifically, we present algorithms for solving decision tree models and influence diagram models of sequential decision problems. For decision tree models, we propose a mixed integer programming formulation that is valid for a subclass of RDU models (corresponding to risk seeking behaviors). This formulation reduces to a linear program when mixed strategies are considered. In the general case (i.e., when there is no particular assumption on the parameters of RDU), we propose a branch and bound procedure to compute an RDU-optimal strategy among the pure ones. After highlighting the difficulties induced by the use of RDU in influence diagram models, we show how this latter procedure can be extended to optimize RDU in an influence diagram. Finally, we provide empirical evaluations of all the presented algorithms.     

Multiple rank multi-linear SVM for matrix data classification

Matrices, or more generally, multi-way arrays (tensors) are common forms of data that are encountered in a wide range of real applications. How to classify this kind of data is an important research topic for both pattern recognition and machine learning. In this paper, by analyzing the relationship between two famous traditional classification approaches, i.e., SVM and STM, a novel tensor-based method, i.e., multiple rank multi-linear SVM (MRMLSVM), is proposed. Different from traditional vector-based and tensor based methods, multiple-rank left and right projecting vectors are employed to construct decision boundary and establish margin function. We reveal that the rank of transformation can be regarded as a tradeoff parameter to balance the capacity of learning and generalization in essence. We also proposed an effective approach to solve the proposed non-convex optimization problem. The convergence behavior, initialization, computational complexity and parameter determination problems are analyzed. Compared with vector-based classification methods, MRMLSVM achieves higher accuracy and has lower computational complexity. Compared with traditional supervised tensor-based methods, MRMLSVM performs better for matrix data classification. Promising experimental results on various kinds of data sets are provided to show the effectiveness of our method.     

Equilibrium-independent passivity: A new definition and numerical certification

We extend the traditional notion of passivity to a forced system whose equilibrium is dependent on the control input by defining equilibrium-independent passivity, a system property characterized by a dissipation inequality centered at an arbitrary equilibrium point. We provide a necessary input/output condition which can be tested for systems of arbitrary dimension and sufficient conditions to certify this property for scalar systems. An example from network stability analysis is presented which demonstrates the utility of this new definition. We then proceed to numerical certification of equilibrium-independent passivity using sum-of-squares programming. Finally, through numerical examples we show that equilibrium-independent passivity is less restrictive than incremental passivity.     

Association patterns for data modeling and definition

Association patterns provide guidance for modeling the associations that occur among objects within both the real world and the solution domains of computer applications. The patterns help the designer better understand and more precisely define the semantics of these associations, which allows them to be more easily and properly implemented. This paper describes a number of association patterns using Object Relationship Notation (ORN) and by doing so provides evidence for the effectiveness of this notation. It also shows how the development of database systems can be improved by an approach that uses association patterns to build a database model and then implements the model by mapping it to an ORN-extended database definition that is supported by a DBMS. The feasibility of this approach and the applicability of our association patterns have been validated by DBMS research prototypes and by the modeling, implementing, and testing of numerous associations.     

Design and implementation of data abstraction definition facility

This paper describes the Implementation model of a data definition facility for abstract data types, implemented as an extension to PL/1. The facility is based on a modified version of the cluster mechanism for the implementation of types. The proposed version tries to address the issues of efficiency and portability in connection with the goal of systematic programming.     

A survey of multilinear subspace learning for tensor data

Increasingly large amount of multidimensional data are being generated on a daily basis in many applications. This leads to a strong demand for learning algorithms to extract useful information from these massive data. This paper surveys the field of multilinear subspace learning (MSL) for dimensionality reduction of multidimensional data directly from their tensorial representations. It discusses the central issues of MSL, including establishing the foundations of the field via multilinear projections, formulating a unifying MSL framework for systematic treatment of the problem, examining the algorithmic aspects of typical MSL solutions, and categorizing both unsupervised and supervised MSL algorithms into taxonomies. Lastly, the paper summarizes a wide range of MSL applications and concludes with perspectives on future research directions.     

Tensor decision trees for continual learning from drifting data streams

Machine Learning - Data stream classification is one of the most vital areas of contemporary machine learning, as many real-life problems generate data continuously and in large volumes. However,...     

Using bilingual Web data to mine and rank translations

The English Reading Wizard uses bilingual Web and local-dictionary data to help readers understand foreign languages by translating words and phrases. Methods include the expectation maximization algorithm and bilingual bootstrapping.     

Tensor distance based multilinear globality preserving embedding: A unified tensor based dimensionality reduction framework for image and video classification

Image and video classification tasks often suffer from the problem of high-dimensional feature space. How to discover the meaningful, low-dimensional representations of such high-order, high-dimensional observations remains a fundamental challenge. In this paper, we present a unified framework for tensor based dimensionality reduction including a new tensor distance (TD) metric and a novel multilinear globality preserving embedding (MGPE) strategy. Different with the traditional Euclidean distance, which is constrained by orthogonality assumption, TD measures the distance between data points by considering the relationships among different coordinates of high-order data. To preserve the natural tensor structure in low-dimensional space, MGPE directly works on the high-order form of input data and employs an iterative strategy to learn the transformation matrices. To provide faithful global representation for datasets, MGPE intends to preserve the distances between all pairs of data points. According to the proposed TD metric and MGPE strategy, we further derive two algorithms dubbed tensor distance based multilinear multidimensional scaling (TD-MMDS) and tensor distance based multilinear isometric embedding (TD-MIE). TD-MMDS finds the transformation matrices by keeping the TDs between all pairs of input data in the embedded space, while TD-MIE intends to preserve all pairwise distances calculated according to TDs along shortest paths in the neighborhood graph. By integrating tensor distance into tensor based embedding, TD-MMDS and TD-MIE perform tensor based dimensionality reduction through the whole learning procedure and achieve obvious performance improvement on various standard datasets.     

Pore scale definition and computation from tomography data

During recent years characterisation capabilities of porous media have been transformed by advances in computation and visualisation technologies. It is now possible to obtain detailed topological and hydrodynamic information of porous media by combining tomographic and computational fluid dynamic studies. Despite the existence of these new capabilities, the characterisation process itself is based on the same antiquated experimental macroscopic concepts.We are interested in an up-scaling process where we can keep key information for every pore size present in the media in order to feed multi-scale transport models. Hydrometallurgical, environmental, food, pharmaceutical and chemical engineering are industries with process outcomes based on homogeneous and heterogeneous reactions and therefore sensitive to the reaction and transport processes happening at different pore scales.The present work addresses a key step in the information up-scaling process, i.e. a pore identification algorithm similar to alternating sequential filters. In a preliminary study, topological and hydrodynamic variables are correlated with the pore size. Micrometre and millimetre resolution tomographies are used to characterise the pore size distribution of a packed column and different rocks. Finally, we compute the inter-pore-scale redistribution function which is a measure of the heterogeneity of the media and magnitude needed in multi-scale modelling.

Program summary

Program title: PoresizedistCatalogue identifier: AEJJ_v1_0Program summary URL:http://cpc.cs.qub.ac.uk/summaries/AEJJ_v1_0.htmlProgram obtainable from: CPC Program Library, Queen?s University, Belfast, N. IrelandLicensing provisions: Standard CPC licence, http://cpc.cs.qub.ac.uk/licence/licence.htmlNo. of lines in distributed program, including test data, etc.: 312No. of bytes in distributed program, including test data, etc.: 6534Distribution format: tar.gzProgramming language: MatLabComputer: Desktop or LaptopOperating system: Runs under MatLab (tested in Linux and Windows)RAM: Tested for problems up to 10¹⁰ bytesClassification: 7.9, 14External routines: MatLab Image ToolboxNature of problem: Identify individual pores from a foreground image representing void space.Solution method: Algorithm based on successive erosions with a shrinking erosion disk diameter.Restrictions: The tomographic data must fit in the available computer memory. The input tomographic data should have the open porosity space to characterise as foreground.Unusual features: Can be used together with the solution for the fluid flow for obtaining a combined topological-hydrodynamical characterisation.Additional comments: Our implementation was oriented for easy understanding, not computational speed. However, see section regarding memory implementation details, RAM-disk swapping strategy and parallelisation for details.Running time: Typical running time: 2 hours. Largest tested problem (10¹⁰ bytes): 1 day.     

Tensor and border rank of certain classes of matrices and the fast evaluation of determinant inverse matrix and eigenvalues

The tensor rankA of the linear spaceA generated by the set of linearly independent matricesA ₁, A₂, …, A_p, is the least integert for wich there existt diadsu ^(r) v ^(r)τ, τ=1,2,...,t, such that . IfA=n+k,k≪n then some computational problems concerning matricesA∈A can be solyed fast. For example the parallel inversion of almost any nonsingular matrixA∈A costs 3 logn+0(log² k) steps with max(n ²+p (n+k), k² n+nk) processors, the evaluation of the determinant ofA can be performed by a parallel algorithm in logp+logn+0 (log² k) parallel steps and by a sequential algorithm inn(1+k ²)+p (n+k)+0 (k ³) multiplications. Analogous results hold to accomplish one step of bisection method, Newton's iterations method and shifted inverse power method applied toA−λB in order to compute the (generalized) eigenvalues provided thatA, B∈A. The same results hold if tensor rank is replaced by border rank. Applications to the case of banded Toeplitz matrices are shown. Dedicated to Professor S. Faedo on his 70th birthday Part of the results of this paper has been presented at the Oberwolfach Conference on Komplexitatstheorie, November 1983