Source author record

Marc Parizeau

Marc Parizeau appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

astro-ph Machine Learning

Catalog footprint

What is connected

2works

2topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2006arXiv

Genetic Programming, Validation Sets, and Parsimony Pressure

Fitness functions based on test cases are very common in Genetic Programming (GP). This process can be assimilated to a learning task, with the inference of models from a limited number of samples. This paper is an investigation on two methods to improve generalization in GP-based learning: 1) the selection of the best-of-run individuals using a three data sets methodology, and 2) the application of parsimony pressure in order to reduce the complexity of the solutions. Results using GP in a binary classification setup show that while the accuracy on the test sets is preserved, with less variances compared to baseline results, the mean tree size obtained with the tested methods is significantly reduced.

preprint2003arXiv

Automatic Detection of Expanding HI Shells Using Artificial Neural Networks

The identification of expanding HI shells is difficult because of their variable morphological characteristics. The detection of HI bubbles on a global scale therefore never has been attempted. In this paper, an automatic detector for expanding HI shells is presented. The detection is based on the more stable dynamical characteristics of expanding shells and is performed in two stages. The first one is the recognition of the dynamical signature of an expanding bubble in the velocity spectra, based on the classification of an artificial neural network. The pixels associated with these recognized spectra are identified on each velocity channel. The second stage consists in looking for concentrations of those pixels that were firstly pointed out, and to decide if they are potential detections by morphological and 21-cm emission variation considerations. Two test bubbles are correctly detected and a potentially new case of shell that is visually very convincing is discovered. About 0.6% of the surveyed pixels are identified as part of a bubble. These may be false detections, but still constitute regions of space with high probability of finding an expanding shell. The subsequent search field is thus significantly reduced. We intend to conduct in the near future a large scale HI shells detection over the Perseus Arm using our detector.