Exoplanets Discovered through AI

National Aeronautics and Space Administration (NASA) announced the discovery of two new exoplanets, Kepler-90i and Kepler-80g using a deep learning neural network - an artificial intelligence (AI) tool that mimics the workings of a human brain. This was first time AI was used for such a discovery in the universe.

Exoplanets Discovered through AI
  • The discovery demonstrates that good science not only solves problems but also can also improve itself

What Are Exoplanets?

  • Exoplanets are the planets outside our solar system that orbits another star.
  • First evidence of an exoplanet was noted as early as 1917 since then they are routinely being discovered
  • The number of exoplanets already found now stands at 3,567.
  • The nearest exoplanet is Proxima Centauri b, located 4.2 light-years (1.3 parsecs) from Earth and orbiting Proxima Centauri, the closest star to the Sun

Significance of Discovery of Kepler 90i

  • With the discovery of the planet Kepler 90i, orbiting the star Kepler 90, there is another star besides the Sun with eight planets orbiting it.
  • Christopher Shallue from Google and Andrew Vanderburg from University of Texas, Austin discovered it using a deep learning neural network

How Does AI Work? What Are Neural Networks?

  • Artificial intelligence (AI) overs mechanism which bless machines with “intelligence,” - at least in limited terms - emulating human being՚s unique reasoning faculties.
  • Neural network is a system of hardware or software patterned after the operation of neurons in the human brain, that is, they model the neurons with their interconnections.
  • The neural network develops an input output relation based on correct input and corresponding output presented to it. This process is called training, once trained the network is then expected to produce the correct output for a new input even not in training sample.
Neural Networks
  • For example, the neural network can use examples to automatically infer rules for recognizing handwritten digits. These examples are called training patterns. Furthermore, by increasing the number of training examples, the network can learn more about handwriting, and so improve its accuracy.
  • Once the network is trained, the idea is that it would recognize any handwritten digit pattern.
  • In this case scientist “trained” their computer to analyse light readings made by NASA՚s Kepler Space Telescope, which are archived and made available for anyone to use. During its mission from 2009 to 2013, the Kepler Space Telescope surveyed nearly 200,000 stars, with 35,000 possible planet signals. The network was made to learn to identify true signals using 15,000 previously vetted signals.
  • The network then studied the weaker signals recorded from 670 star systems that had multiple known planets orbiting them, finally coming up with this discovery.
  • The network also identified another Earth-sized exoplanet, Kepler 80g, orbiting the star Kepler 80. This is a very stable system in which Kepler 80g and four of its neighbors are locked together in a so-called resonant chain.

Applications of Deep Learning and Neural Networks

Some of the classical applications include:

  • Colorization of Black and White Images.
  • Adding Sounds To Silent Movies.
  • Automatic Machine Translation.
  • Object Classification in Photographs.
  • Automatic Handwriting Generation.
  • Character Text Generation.
  • Image Caption Generation.
  • Automatic Game Playing.

Deep learning and neural networks have many applications successfully

Robotvetter Program

  • After the initial discoveries of exoplanets as their numbers grew, there was need for automating the initial vetting process.
  • Robotvetter program was the first attempt at automating the process of rejecting false positives in the signal.

Examrace Team at Aug 21, 2021