A discriminative learning framework with pairwise constraints for video object classification

To deal with the problem of insufficient labeled data in video object classification, one solution is to utilize additional pairwise constraints that indicate the relationship between two examples, i.e., whether these examples belong to the same class or not. In this paper, we propose a discriminative learning approach which can incorporate pairwise constraints into a conventional margin-based learning framework. Different from previous work that usually attempts to learn better distance metrics or estimate the underlying data distribution, the proposed approach can directly model the decision boundary and, thus, require fewer model assumptions. Moreover, the proposed approach can handle both labeled data and pairwise constraints in a unified framework. In this work, we investigate two families of pairwise loss functions, namely, convex and nonconvex pairwise loss functions, and then derive three pairwise learning algorithms by plugging in the hinge loss and the logistic loss functions. The proposed learning algorithms were evaluated using a people identification task on two surveillance video data sets. The experiments demonstrated that the proposed pairwise learning algorithms considerably outperform the baseline classifiers using only labeled data and two other pairwise learning algorithms with the same amount of pairwise constraints.     

A combined dislocation—cohesive zone model for fracture in a confined ductile layer

The collective dislocation behavior near a crack tip in a ductile layer sandwiched between two brittle solids is analyzed via two-dimensional dislocation dynamics (DD) simulations that incorporate a cohesive zone (CZ) model. The cohesive crack tip is treated as part of a much larger finite crack confined in the ductile layer. The underlying boundary value problem is formulated with a set of boundary integral equations and numerically evaluated with a collocation method. The fracture energy of the layered composite material is shown to be strongly correlated with the layer thickness and is directly influenced by the cohesive strength of the ductile layer (Hsia KJ et al. (1994) J Mech Phys Solids 6 877–896).     

Effects of Ti on N distribution and diffusion in FeTiN thin films

We studied the distribution and diffusion of N atoms in FeTiN single-layer and bilayer thin films by X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS). We found that in as-deposited films N atoms are first absorbed by Ti atoms, the rest being dissolved into FeTi lattices. Ti not only directly absorbs N by chemical bonding, but also decreases the energy of the N atoms that are dissolved into FeTi lattices. The diffusion study of N atoms in single layer and bilayer films showed that although the presence of Ti stabilized the /spl alpha/ phase of FeTi lattices during 200/spl deg/C annealing, the diffusion length of dissolved N in FeTi (Ti/spl les/8 at.%) lattices was still comparable to the diffusion length of N in pure Fe. Thus, the addition of a small amount of Ti in /spl alpha/-Fe lattice can not completely stabilize N atoms in the film, and the induced magnetic anisotropy of the films can still be unstable, although the /spl alpha/ phase of FeTiN is stabilized by the addition of Ti. The distribution of N atoms in bilayer films can be fitted by a simple enthalpy model.     

Euler Critical Force Calculation for Laced Columns

The paper presents a method of solving the buckling problem of laced column as a statically indeterminate structure without analyzing determinants of high order. The flexural and torsional buckling problems of laced column are reduced to the two-point boundary value problem for a difference equation system. The value of Euler critical load is determined as a result of analyzing the fourth order determinant for column with any degree of static indeterminacy. The solution is based on the method of initial values. Stability of columns with any types of lattice (crosswise, serpentine, with batten struts); with any number of lattice panels and with variable lattice spacing can be examined by this manner. The analogy between the flexural and torsional buckling of the laced column is established. It enables one to use the same relations for consideration of both kinds of buckling. The obtained numerical results show that the Euler critical loads calculated by this method can be substantially differed from those based on the approximated Engesser’s approach. A PC program for checking stability of laced column by designer can be developed on the basis of the present method.     

Software release management for component‐based software

Software release management is the process through which software is made available to and obtained by its users. Until now, this process has been relatively straightforward. However, the emergence of component‐based software is complicating software release management. Increasingly, software is constructed via the assembly of pre‐existing, independently produced, and independently released components. Both developers and users of such software are affected by these complications. Developers need to accurately document the complex and changing dependencies among the components constituting the software. Users must be involved in locating, retrieving, and assembling components in order to appropriately bring the software into their particular environment. In this paper, we introduce the problem of release management for component‐based software and discuss SRM, a prototype software release management tool we have developed that supports both developers and users in the software release management process. Copyright © 2002 John Wiley & Sons, Ltd.     

Retrieval Of Morphological Variants In Searches Of Latin Text Databases

This paper reports a detailed evaluation of the effectiveness of a system that has been developed for the identification and retrieval of morphological variants in searches of Latin text databases. A user of the retrieval system enters the principal parts of the search term (two parts for a noun or adjective, three parts for a deponent verb, and four parts for other verbs), this enabling the identification of the type of word that is to be processed and of the rules that are to be followed in determining the morphological variants that should be retrieved. Two different search algorithms are described. The algorithms are applied to the Latin portion of the Hartlib Papers Collection and to a range of classical, vulgar and medieval Latin texts drawn from the Patrologia Latina and from the PHI Disk 5.3 datasets. The effectiveness of these searches demonstrates the effectiveness of our procedures in providing access to the full range of classical and post-classical Latin text databases. This revised version was published online in July 2006 with corrections to the Cover Date.     

Detecting infeasibility and generating cuts for mixed integer programming using constraint programming

We study a hybrid MIP/CP solution approach in which CP is used for detecting infeasibilities and generating cuts within a branch-and-cut algorithm for MIP. Our framework applies to MIP problems augmented by monotone constraints that can be handled by CP. We illustrate our approach on a generic multiple machine scheduling problem, and present a number of computational experiments.     

Decentralized adaptive recurrent neural control structure

This paper presents a novel decentralized variable structure neural control approach for large-scale uncertain systems, which is developed using recurrent high-order neural networks (RHONN). It is assumed that each subsystem belongs to a class of block-controllable nonlinear systems whose vector fields includes interconnection terms, which are bounded by nonlinear functions. A decentralized RHONN structure and the respective learning law are proposed in order to approximate online the dynamical behavior of each nonlinear subsystem. The control law, which is able to regulate and to track the desired reference signals, is designed using the well-known variable structure theory. The stability of the whole system is analyzed via the Lyapunov methodology. The applicability of the proposed decentralized identification and control algorithm is illustrated via simulations as applied to an interconnected double inverted pendulum.