11/04/2023
Taljard,Prinsloo,Goosen-Creating electronic resources for African languages:challenges&opportunities
The availability of electronic resources for African languages is a burning issue, despite many (successful) efforts of African language linguists and computational linguists to extend the pool of resources needed for Human Language Technology applications. HLT applications are dependent on large quantities of machine-readable data, and one strategy for the collection of such data is digitization of existing resources. Digitization is the conversion of analogue text, video and audio material into a digital format which would allow downstream processing. The purpose of our presentation is to provide an overview of the tools, procedures, standards and best practices which we have utilized in our digitization effort. We contextualize our work within a specific environment, which is capacity and time constrained, and also less resourced with regard to technical skills. With regard to the digitization of text, we pay particular attention to our selection of OCR software, with the specific orthographies of the South African languages in mind. We furthermore demonstrate that cleaning of OCRed material is a scalable process, depending on the intended use of such material. We also briefly describe the processes followed for the digitization of video and audio material, which are to a certain extent less challenging than digitizing textual material. One of the main challenges for making electronic resources available on open access platforms, which would maximize their usability for potential users, is copyright, therefore a brief discussion on salient copyright issues is included in the presentation. Lastly, we explore the possibilities of collaboration with interested parties. shared via The South African Centre for Digital Language Resources (SADiLaR)
The availability of electronic resources for African languages is a burning issue, despite many (successful) efforts of African language linguists and comput...
