**Abstract**

The research of planets from outside our Solar System, termed exoplanets, has opened a wide range of new possibilities. Some of the current interests in exoplanet research are related to their discovery and the characterization of their atmospheres. Finding these planets is important because it may lead to answering several questions; such as the formation of planets and stellar systems, and possibly finding life outside planet Earth. There are several works that propose using artificial intelligence to ease the processes involved in exoplanet research. Many studies have focused on the detection of such celestial bodies, as well as reducing the number of false detections. Recently, the study of exoplanet atmospheres has also received considerable attention, due to its potential for finding life on these planets. In this work, we describe an artificial intelligence approach for reducing the number of spurious detections of exoplanets using the transit technique. This approach is based on using spectral multiresolution analysis techniques, which allow the artificial intelligence algorithms to better identify the exoplanet signals.

**Keywords:** artificial intelligence, deep learning, exoplanets, light curves, machine learning, multiresolution analysis, neural networks
