Luis Simoins, founder of ML Analytics, won first place in the Ariel Machine Learning Data Challenge, dedicated to developing highly accurate artificial intelligence and machine learning methods to describe planets orbiting other stars, such as those to be studied by the Ariel mission, which the European Space Agency ( ESA) plans to launch in 2029.
“The winning decision estimates the size of the planets as they pass in front of their stars,” said a statement sent to the editorial office. These estimates reached an average error of 0.00007, which is significantly lower than other teams, which already puts the method very close to Ariel’s mission accuracy targets.
The Portuguese researcher, who has been working on using AI to solve space problems since 2008, is very happy about this achievement. “It is a great pride for me and ML Analytics, the company I founded, to bring this award to Portugal,” he stresses in a statement.
“It is the culmination of decades of research to support and develop artificial intelligence. It is my hope that this small step for science will lead, through Ariel’s mission, to a large expansion of human knowledge of the universe.”Adds Louis Simoins.
The European Space Agency mission will study the atmosphere and chemistry of 1,000 exoplanets, and this becomes especially challenging given the need to observe planets passing over their star, sometimes hundreds of light years away.
At this distance, one of the main questions that arises is the difference between planet, star and instrument. It is for this reason that the machine learning community was tasked with developing a solution that could contribute to the success of the mission.
Pedro Machado, Ariel Mission Representative in Portugal, explains in the video the components of ESA’s future launch.
O Ariel Machine Learning Data Challenge The main goal was to develop the ability to eliminate noise introduced into the Ariel satellite signal by sunspots, as well as noise from instruments.
