AI finds previously invisible traces of fire being used 800,000 years ago

AI finds previously invisible traces of fire being used 800,000 years ago

With the help of an algorithm, a research group from Israel has found traces of pre-human use of fire where traditional methods have failed. There is now a sixth place on earth where there is evidence that pre-humans used fire more than half a million years ago. So far, however, visual traces – i.e. primarily the color change of burnt objects – have been necessary for this, but there was no such help with the objects found now analyzed. Instead, an algorithm recognized patterns at the molecular level in found flints that go back to heating and was even able to determine the temperature. The method could provide archeology with further insights.

View into the Evron quarry during the excavations in 1976/77

(Image: Evron Quarry Excavation Archive)

As the team led by Zane Stepka from the Weizmann Institute of Science in Rechovot in Israel explains, current research assumes that pre-humans began to use fire in a controlled manner at least a million years ago. That would be at the transition from Homo Habilis to Homo Erectus, the consequences for human development were definitely drastic: Cooked food made it possible for the brain to grow. However, it is difficult to confirm this chronology in view of the archaeological circumstances, because up to now visible traces of the use of fire have been relied on. Evidence for the widespread use of fire therefore only goes back about 200,000 years, older ones are extremely rare.

The team suggests that this could now change with their method. As they explain in the journal Proceedings of the National Academy of Sciences, they have dedicated themselves to a site known as Evron Quarry. Traces of humans found there are up to a million years old. This includes 26 flint tools. To find out if they were once exposed to fire, the team collected similar rocks from the area, heated them in fire, and used the molecular structure measurements to train an algorithm.

Some of the flints

(Image: Zane Stepka/Weizmann Institute of Science)

From the comparison to the measurement data of unheated flints, the AI ​​should then find differences that are invisible to humans and learn to detect traces of fire. In fact, they succeeded and the deep learning algorithms could even determine what temperature was reached. The prehistoric tools were therefore heated to different degrees, sometimes up to 600 degrees Celsius. The archaeologists who have been involved in the investigation so far have been certain beforehand that nothing can be found on the objects. “We should have bet,” says Filipe Natalio.

Overall, the project is a “resounding success,” adds Natalio. He not only refers to the evidence itself, but also to its importance for research as a whole. “If we apply this method to sites that are one or even two million years old, we could find something completely new,” explains Stepka. It is not the first time that AI techniques have been used in archaeology. Among other things, this involved taking on particularly time-consuming and tedious work, in particular sorting broken glass. Algorithms can also provide valuable help in the analysis of historical texts.


To home page

#finds #previously #invisible #traces #fire #years

Leave a Comment

Your email address will not be published.