The project will consist of:
- a study on existing automated speech recognition solutions and the availability of training data (speech and related transcription);
- the creation of an open-source basic speech recognition prototype solution;
- a new collection of training data (speech and related transcription).
The study will among others provide answers to these questions:
- What types of automated speech recognition solutions are currently on the market?
- How are these solutions developed? Which expert knowledge, data, budget, infrastructure, computing power do they need?
- According to what criteria can the solutions be assessed and compared with each other?
- What is the state of the art and the results achieved by these solutions today,
- What are the predictions for future technology developments?
- What markets exist today for automated speech recognition solutions?
- Who are the suppliers, brands/ products, buyers/ users and applications and what are their market shares in Europe and globally?
- What are the predictions about the future of this market?
The open-source basic speech recognition prototype solution will cover low-resource EU official languages with at least one language from Northern, Eastern and Southern Europe supported by the data collection activities linked to the procurement.
The project is financed under the 2022 Language Technology Solutions call for tenders implementing the DIGITAL Europe Programme in the field of language technologies.