Translating lost languages using machine learning

Description

Summary

• Lost languages are more than a mere academic curiosity; Unfortunately, most of them have such minimal records that scientists can’t decipher them by using machine-translation algorithms like Google Translate.
• The algorithm learns to embed language sounds into a multidimensional space where differences in pronunciation are reflected in the distance between corresponding vectors.
• This design enables them to capture pertinent patterns of language change and express them as computational constraints.