Project organization

1. Speech Input/Output . This involves robust speech recognition for the transformation of speech corrupted by background and channel noise to text, and speech output, producing natural sounding and highly intelligible speech from computer generated text. This work area is mainly contained in Part I.
2. Syntactic and semantic analysis. This involves the grammatical analysis of the text proposals generated by the speech recognizer and the semantic interpretation thereof. The semantic analysis is included in Part I whereas syntactic analysis mainly comes in Part II.
3. Dialog management and system integration. This defines the structure of the human-machine dialog, and the interaction between the system components. A basal component of dialog management, based upon results from the TABOR project (see below) will be included in Part I. Part II contains the development of a full-fledged dialog management system.
4. Language resources. Speech and text corpora for the specific task of public transport information will be collected and annotated in Part I. In Part II a more comprehensive corpus will be collected, covering both a wide range of inter-speaker variation, including a wider dialect variety.