An error-entropy minimization algorithm for supervised training ofnonlinear adaptive systems

The paper investigates error-entropy-minimization in adaptive systems training. We prove the equivalence between minimization of error's Renyi (1970) entropy of order α and minimization of a Csiszar (1981) distance measure between the densities of desired and system outputs. A nonparametric estimator for Renyi's entropy is presented, and it is shown that the global minimum of this estimator is the same as the actual entropy. The performance of the error-entropy-minimization criterion is compared with mean-square-error-minimization in the short-term prediction of a chaotic time series and in nonlinear system identification     

Computationally Efficient Exact Calculation of Kernel Density Derivatives

Machine learning research related to the derivatives of the kernel density estimator has received limited attention compared to the density estimator itself. This is despite of the general consensus that most of the important features of a data distribution, such as modes, curvature or even cluster structure, are characterized by its derivatives. In this paper we present a computationally efficient algorithm to calculate kernel density estimates and their derivatives for linearly separable kernels, with significant savings especially for high dimensional data and higher order derivatives. It significantly reduces the number of operations (multiplications and derivative evaluations) to calculate the estimates, while keeping results exact (i.e. no approximations are involved). The main idea is that the calculation of multivariate separable kernels and their derivatives, such as the gradient vector and the Hessian matrix involves significant number of redundant operations that can be eliminated using the chain rule. A tree-based algorithm that calculates exact kernel density estimate and derivatives in the most efficient fashion is presented with the particular focus being on optimizing kernel evaluations for individual data pairs. In contrast, most approaches in the literature resort to approximations of functions or downsampling. Overall computational savings of the presented method could be further increased by incorporating such approximations, which aim to reduce the number of pairs of data considered. The theoretical computational complexity of the tree-based and direct methods that perform all multiplications are compared. In experimental results, calculating separable kernels and their derivatives is considered, as well as a measure that evaluates how close a point is to the principal curve of a density, which employs first and second derivatives. These results indicate considerable improvement in computational complexity, hence time over the direct approach.

     

Requirement for EGF receptor signalling in neural recruitment during formation of Drosophila chordotonal sense organ clusters

The different lesions of the glenoid labrum are described. They may involve the antero-inferior, the posterior or the superior (SLAP lesions) part of the labrum. CT-arthrography is the gold standard imaging modality in this field of shoulder abnormalities.     

Essentials of Software Process

Process trends can be placed inside a triangular map according to their emphasis on three aspects, represented by the vertices: people, technology, and rigor. Plan-oriented, engineering, and research-based approaches tend to view software as a rigid artifact, so they stress technology and rigor over people. Evolutionary approaches tend to view software development as an organic, skills-driven technical activity, so they stress people and technology over rigor. But this scheme of positioning process approaches is rather rough. A more complete scheme requires dissection in terms of seven essential, mutually reinforcing characteristics: human-centricity, technical orientation, discipline, pragmatism, empiricism, experimentation, and value orientation.     

So Many Languages, So Little Time

What's up and coming in the programming language arena? A rudimentary analysis of the 200+ sessions' titles and abstracts at OOPSLA 07 (22nd Int'l Conf. Object-Oriented Programming, Systems, Languages, and Applications) provides a rough idea of what's happening with object-oriented, functional, dynamic, and domain-specific languages.     

Asymmetric cell division

With the recent identification of intrinsic cell-fate determinants for asymmetric cell division in several systems, biologists have begun to gain insight into the cellular mechanisms by which these determinants are preferentially segregated into one of the two daughter cells during mitosis so that the daughter cells acquire different fates.     

Overexpression of cyclin D1 correlates with early recurrence in superficial bladder cancers

Cyclin D1 is a cell cycle regulator essential for G1 phase progression and is frequently overexpressed in several human tumour types as a consequence of gene amplification or chromosomal rearrangements. We analysed the expression of cyclin D1 in 75 patients with transitional cell carcinoma (TCC) to investigate the possible relationship between its expression and clinical outcome as well as histopathological findings using the immunohistochemical method. We observed strong staining (++, > 50% positive cells) for cyclin D1 in 19 cases (25.3%) and weak staining (+, 5-50% positive cells) in 19 cases (25.3%). Overexpression of cyclin D1 was not associated with tumour invasion. No significant association was found between overexpression of cyclin D1 and tumour grade (P > 0.05). We assessed the differences of disease-free interval in superficial tumours and actuarial survival probability in invasive tumours according to the status of cyclin D1 expression. Tumours with (++) staining for cyclin D1 recurred much more rapidly than (-) and/or (+) staining tumours (P < 0.01 for - vs ++; P < 0.05 for + vs ++). However, overexpression of cyclin D1 was not associated with a shortened overall survival of patients with invasive tumours (P < 0.1). These results suggest that genetic alteration of cyclin D1 appears to be an early event in the tumorigenesis of bladder TCC and is associated with early recurrence in superficial tumours.