The internet today is made up of a rich, diverse tapestry of ideas, cultures, perspectives, opinions, and resources. Leveraging popular search engines, people access these ideas and perspectives through their internet resources (‘resources’ here is an umbrella term to refer to the various formats of information we find on the internet – encompassing Wikipedia pages, blog posts, academic papers, AI datasets, models, applications, and many more). The ability to find (or not find) a specific resource profoundly affects the representation of that resource on the internet. Therefore, the discoverability of resources (and perspectives, cultures) greatly affects the usage, and, to a large extent, the influence of that resource (and its origin community). In today’s fast-paced digital world, if you can not find something, it practically does not exist for you (for example, before Lanfrica Records, this resource about the Ibibio language was invisible on the internet because it was just on a Google sheet).

Africa is one of the most linguistically and culturally diverse regions in the world – a continent rich in languages and traditional knowledge. Despite this, African languages are barely represented on the internet, and consequently, in digital technologies which make use of information from the internet. In today’s internet, there is a large dearth of African and indigenous perspectives, and current digital technologies do not support many African languages and viewpoints. As an example, Google Translate, a popular and widely used translation service, supports only a minuscule fraction of the more than 2000 African languages. Additionally, current voice technologies (Siri, Alexa, and Google Home) do not support a single African language! The African languages are not supported, not necessarily because of the lack of data for these languages, but primarily because the available data can not be found.

The few existing indigenous African resources, encompassing articles, papers, and language data, remain elusive within our current internet landscape. This obscurity stems from the under-representation of African languages, perspectives, and viewpoints on the internet, which renders these resources dispersed, siloed, or hidden. Such a lack of visibility severely impedes the digital representation of African language resources, as well as the advancement of African language technologies.

Lanfrica Records

Lanfrica Records aims to mitigate the difficulty encountered in the discovery of African language resources by creating a centralized, language-first catalog. For instance, if you’re looking for resources (linguistic datasets or research papers) in a particular African language, Lanfrica will point you to the different sources on the web that have such datasets in the desired language.

The Lanfrica Records platform adopts a participatory approach by allowing the general community to contribute resources. At Lanfrica, we link resources: Linking means using just the key metadata (title, short description, and url to where the resource is hosted). We then pre-process the url (find the African languages through our algorithms, if already not given, map it to the best resource type for it, connect it with other related resources on Lanfrica Records) and put it on the Lanfrica Records. Most importantly, Lanfrica does not only offer access to African language data sets for various applications (machine translation, natural language processing, automatic speech recognition, sentiment analysis, text-to-speech, named entity recognition, question answering, media, etc.), but it also provides access to relevant scientific papers, existing software, projects, and language technologies. As such, it takes a truly holistic approach to covering resources for African languages.

Our overarching goals with this project are to expand the content of the platform, improve the user experience, and support community contributions. f you are interested in working with our team, have a look at the opportunities below. All the positions are remote.