A three-person team named “Black Bean” won the “OpenAI to Z Challenge” by discovering 67 potential ancient sites in the Amazon, each about a square mile. This was announced recently, with notable judges including Egyptologist Sarah Parcak. The team utilized advanced deep learning models, analyzing remote sensing LiDAR data, and satellite images from platforms like Google Earth Engine and NASA’s digital elevation models. By employing OpenAI’s GPT-4o model, they identified patterns of known archaeological sites across unexplored regions in Brazil, focusing on areas near water sources—common locations for ancient civilizations. Yao Zhao, a member of “Black Bean,” highlighted that their method allowed for rapid analysis of vast geographic data, facilitating archaeological discoveries without extensive groundwork. The team received a $250,000 cash prize and premium OpenAI credits. This achievement underscores the potential for AI applications in archaeology.
Source link
