Third International Conference on Spoken Language Processing (ICSLP 94)
In this paper we present a suitable architecture for the unification of many aspects of natural language processing research. The architecture - called Pantome - has been developed from non-linear phonology, but applies more generally to other linguistic activity. The approach taken is cognitively motivated, inherently parallel, and requires revision of the notion of segment. We outline how we can demonstrate the validity of our claims through use of the architecture in a text-to-speech conversion application. Related papers [9, 16] present other aspects of the implementation of the architecture in this speech and language processing application.
Bibliographic reference. Edmondson, William / Iles, Jon (1994): "A non-linear architecture for speech and natural language processing", In ICSLP-1994, 29-32.