Artificial intelligence has catapulted the world into the future with platforms that can simulate how humans talk and even process information — but the tech can also help solve history’s great mysteries.
Researchers with Yamagata University Institute of Nasca and IBM Japan used a deep learning AI model to uncover Peruvian geoglyphs etched into the Nazca desert that date back to between 500 BC and 500 AD.
Geoglyphs are depressions made to the earth to create various shapes and lines, with Peru having what are considered the world’s most famous geoglyphs known as the Nazca Lines.
Geoglyphs are often massive, with previously discovered Nazca Lines reaching up to 1,200 feet long, which makes them virtually impossible to detect when on the ground. Archeologists first uncovered the Peruvian geoglyphs nearly 100 years ago after the advent of planes, when pilots spotted the shapes from the air.
Using the AI system, researchers were able to uncover four new geoglyphs that depict a “humanoid” figure holding what appears to be a club, one that depicts a fish, another showing a bird and one that appears to show a pair of legs.
There is debate among academics as to why people made the geoglyphs, with some speculating they wanted to honor deities who they believed could see the shapes from above, while some argue extraterrestrials played a role and the lines are remnants of an airfield for alien spacecraft.
Up until recently, archeologists and researchers would examine aerial photos of the area with just their naked eyes to try to find new geoglyphs, which “requires a substantial amount of time, posing a challenge in efficiency and scalability,” according to the Yamagata researchers.
The scientists turned to artificial intelligence in their quest, training a deep learning system to identify potential Nazca Lines based on previous geoglyphs found in the area.
“Due to the requirement of detecting unconfirmed geoglyph candidates, careful consideration and ingenuity were needed in order to train a deep learning object detection model using training data of very limited quality and quantity,” researchers said.
Utilizing AI paid off for the Yamagata team, as the tech was able to work 21 times faster compared to when just humans analyzed such photos.
“We could identify new geoglyph’s candidates approximately 21 times faster than with the naked eye alone,” the researchers said in the study, which was published in the Journal of Archaeological Science. “The approach would be beneficial for the future of archaeology in a new paradigm of combining field survey and AI.”
Following the success of integrating AI into archeological research, the Yamagata researchers will now team up with the IBM T. J. Watson Research Center, based out of New York, to expand their research to the entire region where the lines were discovered.
“In addition, we plan to work with the Ministry of Culture of Peru to implement activities aimed at protecting the geoglyphs discovered using AI,” the researchers said.
Archeologists have previously used artificial intelligence to uncover other mysteries of the world, with the computers taking what is often the most difficult job for scientists and explorers: physically searching land for artifacts, lost cities and burial grounds.
AI systems trained to detect patterns on land using satellite and sonar images have already proven fruitful for other archaeologists, with AI detecting a Mesopotmian burial site in 2021 based on satellite images and another AI system detecting shipwrecks with 92% accuracy.
AI technology has also helped scientists translate ancient texts, with researchers at the University of Chicago’s Oriental Institute and Department of Computer Science recently training a system on thousands of images and ancient characters that can translate ancient inscriptions with an 80% accuracy.
For the scientists at Yamagata University, they highlighted that archaeologists will likely see a boom in use of AI in the future.
“Recent advances in automated sensing enabled by the proliferation of drones, robotics and Light Detection and Ranging (LiDAR), Big Data, and artificial intelligence may fuel the next wave of archeological discovery,” they said.