Developing Natural Language Processing (NLP) for Low Resource Languages (LRLs) is an obstacle in the way of linguists’ journey toward machine translation of all natural languages. Basically, NLP is the collection of data and machine understanding of languages that we speak. Processing natural languages is what allows us to create machine translation tools such as google translate. However, as many probably know, the accuracy of these types of tools varies widely, especially for lesser-known languages. These kinds of languages, ones that have limited available data (dictionaries, books, news/media, etc.), are often referred to as Low Resource Languages (LRLs). Creating accurate translations alone is difficult enough without sufficient references, but designing a system to translate between languages becomes exponentially more difficult. Today, most of the NLP research available is for just about 20 of the thousands of languages spoken in the world, which is why it is so important to study, record, and understand as many languages as possible, before they have the chance to become dormant.

Now that communication across the globe has become so much easier and more accessible, translation is beginning to hold us back. When natural disasters strike, we should be able to effortlessly communicate with places that are far away and need help. When earthquakes happen in places with few resources that are available after the disaster, they should be able to rely on other countries and places to be able to communicate effectively and provide assistance. However, time is running out to save all the languages we have today since a language is lost to dormancy more than once every two weeks. Using this approximation, nearly half of Earth’s 7000 languages are expected to become dormant within the next 100 years.