Interactive music composition driven by feature evolution.

Evolutionary music composition is a prominent technique for automatic music generation. The immense adaptation potential of evolutionary algorithms has allowed the realisation of systems that automatically produce music through feature and interactive-based composition approaches. Feature-based composition employs qualitatively descriptive music features as fitness landmarks. Interactive composition systems on the other hand, derive fitness directly from human ratings and/or selection. The paper at hand introduces a methodological framework that combines the merits of both evolutionary composition methodologies. To this end, a system is presented that is organised in two levels: the higher level of interaction and the lower level of composition. The higher level incorporates the particle swarm optimisation algorithm, along with a proposed variant and evolves musical features according to user ratings. The lower level realizes feature-based music composition with a genetic algorithm, according to the top level features. The aim of this work is not to validate the efficiency of the currently utilised setup in each level, but to examine the convergence behaviour of such a two-level technique in an objective manner. Therefore, an additional novelty in this work concerns the utilisation of artificial raters that guide the system through the space of musical features, allowing the exploration of its convergence characteristics: does the system converge to optimal melodies, is this convergence fast enough for potential human listeners and is the trajectory to convergence "interesting' and "creative" enough? The experimental results reveal that the proposed methodological framework represents a fruitful and robust, novel approach to interactive music composition.

A chemical energy approach of avascular tumor growth: multiscale modeling and qualitative results.

In the present manuscript we propose a lattice free multiscale model for avascular tumor growth that takes into account the biochemical environment, mitosis, necrosis, cellular signaling and cellular mechanics. This model extends analogous approaches by assuming a function that incorporates the biochemical energy level of the tumor cells and a mechanism that simulates the behavior of cancer stem cells. Numerical simulations of the model are used to investigate the morphology of the tumor at the avascular phase. The obtained results show similar characteristics with those observed in clinical data in the case of the Ductal Carcinoma In Situ (DCIS) of the breast.

An integrated genetic and cytogenetic map for the Mediterranean fruit fly, Ceratitis capitata, based on microsatellite and morphological markers.

A genetic map based on microsatellite polymorphisms and visible mutations of the Mediterranean fruit fly (medfly), Ceratitis capitata is presented. Genotyping was performed on single flies from several backcross families. The map is composed of 67 microsatellites and 16 visible markers distributed over four linkage groups. Fluorescence in situ hybridization of selected microsatellite markers on salivary gland polytene chromosomes allowed the alignment of these groups to the second, fourth, fifth and sixth chromosome. None of the markers tested showed segregation either with the X or the third chromosome. However, this map constitutes a substantial starting point for a detailed genetic map of C. capitata. The construction of an integrated map covering the whole genome should greatly facilitate genetic studies and future genome sequence projects of the species.

Cell-nuclear data reduction and prognostic model selection in bladder tumor recurrence.

OBJECTIVE: The paper aims at improving the prediction of superficial bladder recurrence. To this end, feedforward neural networks (FNNs) and a feature selection method based on unsupervised clustering, were employed. MATERIAL AND METHODS: A retrospective prognostic study of 127 patients diagnosed with superficial urinary bladder cancer was performed. Images from biopsies were digitized and cell nuclei features were extracted. To design FNN classifiers, different training methods and architectures were investigated. The unsupervised k-windows (UKW) and the fuzzy c-means clustering algorithms were applied on the feature set to identify the most informative feature subsets. RESULTS: UKW managed to reduce the dimensionality of the feature space significantly, and yielded prediction rates 87.95% and 91.41%, for non-recurrent and recurrent cases, respectively. The prediction rates achieved with the reduced feature set were marginally lower compared to the ones attained with the complete feature set. The training algorithm that exhibited the best performance in all cases was the adaptive on-line backpropagation algorithm. CONCLUSIONS: FNNs can contribute to the accurate prognosis of bladder cancer recurrence. The proposed feature selection method can remove redundant information without a significant loss in predictive accuracy, and thereby render the prognostic model less complex, more robust, and hence suitable for clinical use.

Effective Backpropagation Training with Variable Stepsize.

The issue of variable stepsize in the backpropagation training algorithm has been widely investigated and several techniques employing heuristic factors have been suggested to improve training time and reduce convergence to local minima. In this contribution, backpropagation training is based on a modified steepest descent method which allows variable stepsize. It is computationally efficient and posseses interesting convergence properties utilizing estimates of the Lipschitz constant without any additional computational cost. The algorithm has been implemented and tested on several problems and the results have been very satisfactory. Numerical evidence shows that the method is robust with good average performance on many classes of problems. Copyright 1996 Elsevier Science Ltd.