Abstract
In this paper we describe a speech recognition system for teaching reading skills to young children, and various research issues and activities necessary to make the system work. A full-scale application is described, which embodies the research. The application interacts with a child to help him or her read, and provides opportunities for the child to show off what he or she has read, and to annotate the reading with audio and video comments. Special focus is given to creating an acoustic model specifically for children, and to designing an interface to deal with complexities of a speech recognition application.