A Survey of Natural Language Processing Task for Amazigh: Achievement, Expectation and Perspective
摘要
The Amazigh language, spoken in Morocco, Algeria, Tunisia, Libya, Siwa (Egypt), Niger, and Mali, is primarily used for oral communication and has only recently been officially integrated into Moroccan mass media and formal education. Like most languages that have only recently been subject to Natural Language Processing (NLP), Amazigh lacks sufficient resources and tools in this respect due to a certain number of difficulties. These are dialectic variations, unusual spelling, a lack of annotated corpora, and a lack of linguistic tools and resources. It is on the basis of these problems that investments in the collection and annotation of text corpora, development of linguistic resources like dictionaries, part of speech markers, and modification of translation processes have to be made. The writing system, morphology—which depends on a special method of word formation derived from roots and patterns—and the dearth of linguistic corpora make computational approaches to the Amazigh language very demanding. In the present study, we discuss the current status of the Amazigh NLP knowledge in Morocco and suggest additional technologies and processes to support the language's development in the digital world, as well as promoting the sharing of resources in scientific research to encourage the development of NPL technologies adapted to Amazigh languages, enabling the Amazigh community to be autonomous by offering them equitable access to modern communication and information technologies.