On the application of embedded training to connected letter recognition for directory listing retrieval.

Automatic speech recognition has advanced to the stage where it is now practical to recognize connected strings of words (e.g. digits, letters, city names, airline terms) from a word reference set of isolated tokens of each of the words in the vocabulary. Recently an improved training technique, called embedded word training, was proposed in which reference word patterns were extracted from within connected word sequences themselves. In this manner the word reference patterns had properties more closely related to those of the words in a connected environment. The idea of using such embedded word training has previously been tested in only a single task, namely connected digit recognition. In this investigation we extend the embedded word training procedure to handle letters of the alphabet for use in a directory listing retrieval task.