Voice of WordNet
December 1, 2004 - December 1, 2005
This is part hacking and part art. Construct a multimedia, interactive visualization of the wordnet ontology. The system would display a part of the wordnet graph centered on a particular lexical entry. It would vocalize part of the entries definition via a standard text to speech application. After some time (e.g., 10 seconds), a random neighbor in the graph would be selected and the process repeated. Occasionally, the entire process would start over focused on a random entry to avoid being stuck in poorly connected parts of the graph). The neighbor selection process could be biased by input from the user, perhaps in the form of audio input. There would be ambient music playing in the background, e.g., new age or dub.
Contact Tim FInin, finin@Umbc.edu is interested