Neural Networks, AI, Phone-based VR, Machine Learning, Computer Vision and the CUNAT Automated Translation App – not your father’s archaeological toolkit
Donald H. Sanders · 2018
Hundreds of thousands of ancient Mesopotamian legal treatises, beer recipes, religious rituals, international treaties, medical records, astronomical observations, and mathematical problems-texts that tell us exactly what happened in the words of the very people whose actions formed our shared cultural history-languish untranslated, because there are too many of them, translation takes too long, and there are too few linguistic experts. And thousands more tablets are uncovered every year. This paper describes our CUNAT (the CUNeiform Automated Translator) app and how it sets out to use unconventional (for archaeology) approaches to solving the translation problem. CUNAT will allow anyone to create virtual reality models of complex inscribed artifacts and automatically extract meaning from them more effectively and quickly than existing tools. The app uses photogrammetry, computer vision, AI, and neural networks to complete the tasks.