The multimodal potential of linguistic atlases

The subproject 'Multimodal potential of linguistic atlases' builds on projects like the Bavarian Linguistic Atlas (Bayerischer Sprachatlas). In the designated timeframe and using the funding resources available, the object is to develop a computer-assisted method for the phonetic transcription of dialect recordings as the first building block for the comprehensive AI-based integration and analysis of multimodal language data.

The long-term goal is to round up the collected data in a multimodal complex of cartographic representation, audio presentation and phonetic transcription. The idea is to employ suitable algorithms to substantially reduce the workload required for phonetic transcriptions, tap extensive audio recordings for further analysis and do the spadework for more comprehensive methods ensuring the KI-based evaluation of available language data.

The subproject 'Multimodal potential of linguistic atlases' draws on groundwork done in separate, previously completed language geography projects (the Linguistic Atlas of Lower Bavaria [SNiB]). Following the inclusion of Teuthonista, a phonetic transcription system for the transcription of dialects, in Unicode, a move that was staunchly advocated by Passau dialectologists, this is another important step in ensuring the AI-based integration and analysis of multimodal language data.