Spotting exoplanets is not as simple as pointing a telescope at the sky and searching for planetary objects. The worlds orbiting distant stars are too dim and small for that, but we can track them down using satellites on planetary orbits such as TESS and the far-flung Kepler. These missions produce a lot of data that someone needs to evaluate, and Warwick University researchers think they can do it faster with AI. To illustrate this, the team developed a machine-learning algorithm that confirmed exactly 50 exoplanets in observational data.
Astronomers have two methods at their disposal to detect exoplanets. There is the approach of radial velocity, which controls stars for small counter-motions caused by the gravitational pull of planets. A more sensitive technique, and that used by TESS and Kepler, relies on luminosity in the host star. When the planes of a solar system are well tuned, their planets move away from the star from our perspective. By controlling those dips in brightness, we can determine the presence of exoplanets with a high degree of certainty.
The problem with the transit method is that it produces a mountain of light data for stars, many of which have no visible exoplanets. It takes a combination of computer analysis and human oversight to identify candidates and confirm their existence. The system developed at the University of Warwick is the first that candidate exoplanets can take and perform all necessary analyzes to confirm or rule out their planetary status. Previous attempts to use AI, such as Google’s TensorFlow-based algorithm, can only rank candidates by the chance that they are real planets.
The researchers did not just turn a switch around and have an AI that can dig through data to spot planets. They had to train the neural network with data from confirmed exoplanets and false positives so that it could identify those telltale signs in new data. The 50 exoplanets confirmed by the University of Warwick extend the spectrum of gas giants in Neptune to rocky worlds smaller than Earth. It is particularly difficult to attach smaller planets with the transit method, so that speaks to the accuracy of the AI.
According to the new study, about a third of all confirmed exoplanets have been identified using one analytical method, which is not ideal. Even if existing techniques spot all observing exoplanets, we should have more options just for proper validation, scientists say. They hope to see the new machine learning system evolve as it detects more planets, becoming an important part of the planet exploration process.
Read now:
