How NASA Is Using Deep Learning To Find Landing Sites On Moon

Intel organised an eight-week summer program, the NASA Frontier Development Lab (FDL) event at Santa Clara campus California facility. It was conducted to find methods for using artificial intelligence for the variety of problems. Intel is the major partner in the Frontier Development Lab (FDL), provided support to the research program. Intel announced it support in the research of the use of artificial intelligence(AI) to solve challenges in various fields like weather, space resources as a partner in FDL. Intel is trying to fill the knowledge gaps by using AI which will help in solving several problems. Around 200 TB of data was crunched using AI from the 3D imagery of moon which was gathered from a satellite. Using AI, researchers created maps of the moon’s poles. 

Also Read: 4 Revolutionary Aspects of Artificial Intelligence

However, the major work of NASA is to collect data and then extracting information from that data. The space agency has started enlisting Intel to crunch some of those numbers. It includes Nervana, a deep learning startup to translate numbers into information. Intel is supporting all the researchers who used Intel Nervana deep learning technology for creating lunar poles maps. The major challenge is detecting craters and features within the dramatic shadow of the polar regions. The task takes many weeks to complete even the small section of a moon by hand. So with the deep learning technology, the researchers took less time to achieve the same results as an expert. They even suggested that solar maps of all the rocky objects could be automated using deep learning. 

Although the General Manager of Artificial Intelligence products Group Naveen Rao believe that they have very less access to the latest tools than other countries still they always supported the efforts. It took two and half weeks to transform the data by researchers. On the other hand, the Nervana system took only a minute finish the task with almost 98 percent accuracy. It lets NASA determine the landing points on the surface for lunar rovers more accurately.