2008ApJ...687..542P

Achieving maximum scientific results from the overwhelming volume of astronomical data to be acquired over the next few decades demands novel, fully automatic methods of data analysis. Here we concentrate on eclipsing binary (EB) stars, a prime source of astrophysical information, of which only some hundreds have been rigorously analyzed, but whose numbers will reach millions in a decade. We describe the artificial neural network (ANN) approach which is able to surmount the human bottleneck and permit EB-based scientific yield to keep pace with future data rates. The ANN, following training on a sample of 33,235 model light curves, outputs a set of approximate model parameters [T₂/T₁, (R₁+R₂)/a, esinω, ecosω, and sini] for each input light curve data set. The obtained parameters can then be readily passed to sophisticated modeling engines. We also describe a novel method \typ polyfit for preprocessing observational light curves before inputting their data to the ANN and present the results and analysis of testing the approach on synthetic data and on real data including 50 binaries from the Catalog and Atlas of Eclipsing Binaries (CALEB) database and 2580 light curves from OGLE survey data. The success rate, defined by less than a 10% error in the network output parameter values, is approximately 90% for the OGLE sample and close to 100% for the CALEB sample–sufficient for a reliable statistical analysis. The code is made available to the public. Our approach is applicable to EB light curves of all classes; this first paper in the eclipsing binaries via artificial intelligence (EBAI) series focuses on detached EBs, which is the class most challenging for this approach.