RobERt: the Dreaming Detective for "Habitable" Exoplanets

The Earth got itself another innovative exoplanets detective, and that is the RobERT or the Robotic Exoplanet Recognition.

RobERT is the latest technology launched by men to search for potentially "habitable" planets. It employs machine-learning techniques that are capable of copying human recognition and dreaming processes in its search for exoplanets.

Astronomers created RobERT to look for habitable planets by looking at light spectra from distant systems. By retrieving spectral information about the gasses present in an exoplanet's atmosphere, RobERT might be able to identify the habitable ones.

RobERT will debut at the National Astronomy Meeting 2016 in Nottingham.

"Different types of molecules absorb and emit light at specific wavelengths, embedding a unique pattern of lines within the electromagnetic spectrum," Dr. Ingo Waldmann, lead of the development team said in a statement. "We can take light that has been filtered through an exoplanet's atmosphere or reflected from its cloud-tops, split it like a rainbow and then pick out the 'fingerprint' of features associated with the different molecules or gasses," Waldmann added.

RobERT will try to mimic how good the human brains are when it comes to finding patterns in spectra and in associating the findings to previous experiences or information. The developers think that with an enormous amount of data, the human mind won't be enough to sift through and sort out patterns, that is why RobERT was born. 

RobERT is designed to learn from examples and is capable of building its own experiences, this way RobERT will be like a "seasoned astronomer" that can sort out tons and tons of data fed to it for analysis. This process may take weeks for a human astronomer, but will only take RobERT seconds to analyze.

Deep belief neural networks (DBNs) were created more than a decade ago; it is used for speech recognition, Internet search and tracking customer behavior. But RobERT's DBN is more complex with three laters of processors or neurons. The information fed to it will pass all three layers of neurons before the final identification of gasses. For training purposes, the developers created 85,750 simulated spectra with five different exoplanets to test RobERT's learning process. After the training, RobERT garnered 99.7 percent accuracy.

"RobERt has learned to take into account factors such as noise, restricted wavelength ranges and mixtures of gasses," Waldmann said in a statement published by Phys.Org.  "He can pick out components such as water and methane in a mixed atmosphere with a high probability, even when the input comes from the limited wavebands that most space instruments provide and when it contains overlapping features," Waldmann added.

But another amazing feature of the latest exoplanet detective is its ability to dream. RobERT is capable of entering a "dreaming state" where it can create full spectra based on its experiences. This new feature will help researchers in creating a spectra based on incomplete information. The astronomical community is anticipating the "breathtaking" data RobERT will gather in the near future.