Reproducible Automatic Speech Recognition System
The use case aims at developing a reproducible Automatic Speech Recognition (ASR) system. The use-case proposed is specific for the rich community of Human Language Technologies users in South Africa and the system will be built as a Science Gateway and the data it uses will be stored on an Open Access Repository (from Northwest University.
For storing and retrieving its data, the ASR Science Gateway will use a clone of the the Sci-GaIA Open Access Repository (OAR). DataCite-issued Digital Object Identifiers (DOIs) will automatically be assigned by the OAR to each ASR record to improve its findability, discoverability and reusability.
For analysing use case’s data, the ASR is connected to a FutureGateway API Server – developed in the context of the INDIGO-DataCloud project – which seamlessly executes jobs on local, Grid and Cloud resources belonging to the Africa & Arabia Regional Operation Centre.
Link to the ASR Science Gateway (to be added)