Humans have long been fascinated by what lies beyond our solar system. Unfortunately, it took a lot of effort and time to reach the stage where scientists are able to explore the far reaches of the universe without any issues. The first ever planet located beyond the solar system was found during the early 1990s. In the following years, researchers have been able to find more than 3400 new planets. But determining whether these planets can support life requires extensive research. Scientists spend days or even weeks at a time analyzing any and all available data.
With the launch of the James Webb Space Telescope, the amount of data that is gathered may become overwhelming. Due to this backlog, there is a chance that new discoveries are put on hold for the foreseeable future. In order to combat this, researchers at the University College London have created an artificial intelligence called RobERt. RobERt is able to perform detailed scans of new planets a lot quicker than any person ever could.
The AI accomplishes this by examining various gasses that become visible as light passes through them. Depending on their structure, certain gasses either let the light pass through or they absorb it. Using the info gathered this way scientists can determine what a planet’s atmosphere consists of and if there is any possibility for it to support life. This is not only related to alien life forms, but to the possibility that those planets can support human life as well.
The basics behind the AI are similar to the way a human brain functions. Not unlike the way humans think, the AI’s deep-belief neural network, or DBN for short, filters all the data it receives through multiple layers of neurons. Each time the data passes through the silicon neurons of the AI, it gets processed until the “brain” reaches an answer it considers to be correct.
Similar to a human brain, a deep-belief neural network learns by the process of trial and error. In order to teach RobERt, the researchers gave it access to 85,000 simulations with varying gas spectra. In time, the AI was able to determine the exact mix of gases present on the planet an astonishing 99.7 percent of the time. Researchers even did everything in their power to make the calculations more difficult, by providing incomplete data and samples. In spite of this, the AI’s success rate was higher than the researchers expected.
But this is only a part of what RobERt is able to accomplish. With the data collected by the AI,scientists may be able to better understand how our solar system, and others we come across,have been created. But this is only possible if enough data about other solar systems is collected. And this is where the true value of RobERt comes in. Over time scientists might be lucky enough to find a planet similar to Earth that is able to support human settlers.