Abstract:
Depending on the social features of the speaker and the social setting in which they are speaking, the relationship between meaning and context might alter. In order to interpret the meaning from dysarthric speech, this paper proposes a theoretical framework for employing speech-event representations, also known as situational projections. The multi-layered approach has been broken down
into four main components: a few-shot learner that builds up speaker familiarity; a situational projection component that marshals natural sentences and the built-up familiarity markers into a vector triple; a contextualizer that builds up ontological concepts of the input triple; and finally, a transducer that assumes the function of a logical listener.