🤖 AI Summary
Researchers at the Weizmann Institute applied advanced spectroscopy paired with deep learning to detect molecular signatures of past heating in artifacts from Evron Quarry (Western Galilee), a site dated between 0.8 and 1.0 million years ago. Their AI models — selected after testing classical and deep-learning approaches — learned subtle, nonvisual chemical patterns that correlate with heating and produced temperature estimates for individual items. Analysis of 26 flint tools showed a range of heat exposure (some exceeding 600 °C), and a separate spectroscopic assessment of 87 faunal remains found heat-induced structural changes in an extinct elephant tusk. These signals are invisible to traditional visual inspection, which limits reliable fire evidence to much younger deposits.
The work is significant because it offers one of the oldest lines of evidence suggesting controlled use of fire and provides a scalable, data-driven method for archaeology. By moving beyond color-change and charcoal detection, the new pipeline enables molecular-level temperature inference that can be applied to other Lower Paleolithic sites, potentially expanding the sparse global record of early pyrotechnology and informing hypotheses about cooking, brain evolution, and technological innovation. The study highlights the value of interdisciplinary methods — AI, spectroscopy and paleohuman expertise — to recover hidden behavioral signatures in deep-time contexts.
Loading comments...
login to comment
loading comments...
no comments yet