Supported projects

ColAnTal project lies at the intersection of Natural Language Processing (NLP) and anthropology. It is driven by two key observations. First, despite recent advances in NLP, most of the world’s linguistic heritage remains excluded: of the approximately 7,000 documented languages, only between 200 and 500 benefit from the latest tools (Costa-jussà et al. 2022; Xue et al. 2021; Zeman et al. 2024). This is because these technologies rely on written data, often harvested from the internet. As a result, any language with limited digital presence is effectively left out. Second, anthropological recordings and archives are treasure troves of data, yet they are often only partially digitized or documented, which limits their usability, preservation, and valorization. This project therefore explores potential collaborations between these two disciplines with the goal of improving data and tool sharing. The concrete outcomes will be twofold: enhancing the visibility, reuse and longevity of anthropological data on the one hand, and expanding language coverage in NLP on the other.