Spoken language resources (SLR), for example annotated speech databases, pronunciation lexica and language models, are necessary both for present day speech driven applications, and for the development of a technology capable of handling real, spontaneous speech. SLR are costly to produce because high levels of skill and manpower are needed for each language. The goal of the project is to minimize these costs by creating an efficient infrastructure capable of ensuring the production of SLR for both research and product development, carrying out the distribution of such SLR taking into account quality control and reusability issues, and defining SLR for future research purposes. The above infrastructure is to be tested by launching, in collaboration with other relevant projects, an initiative aimed to bring production centres together and handle the distribution of SLR. The consortium is strongly positioned to contribute to this infrastructure and exploit its results in that it comprises leading European players in the field: telecommunications companies, speech technology providers and research centres.
Funding SchemeCSC - Cost-sharing contracts
CB2 5LD Cambridge